Non-canonical regulated cell death and epigenetic mechanisms in the nervous system

Introduction

Cell death is a widespread phenomenon in living organisms, and to date, more than 20 forms of cell death have been identified, each defined by unique molecular pathways. Cell death was initially believed to be simply a consequence of various types of injury, leading to the premature, unregulated death of living cells, a process known as necrosis. Later, apoptosis, also referred to as programmed cell death, a regulated form of cell death, was discovered, in which cells are responsible for their own demise. Besides apoptosis, anoikis is another form of apoptosis-regulated cell death. Anoikis, resulting after detachment from the extracellular matrix (ECM), activates both the intrinsic (mitochondrial) and extrinsic (death receptor) apoptotic pathways (Lossi, 2022), ultimately converging on caspases. The loss of cell-matrix adhesion due to impaired integrin signaling triggers anoikis, altering the B-cell lymphoma 2 (Bcl-2) family proteins, inducing cytochrome c release, and activating caspases, thereby inhibiting ECM anchorage-independent growth (Taddei et al., 2012). Principal participants include focal adhesion kinase (FAK), SRC proto-oncogene (SRC), phosphatidylinositol 3-kinase (PI3K), extracellular signal-regulated kinase (ERK), and regulatory proteins such as Bcl-2 interacting mediator of cell death (Bim), p53 upregulated modulator of apoptosis (Puma), Bcl-2-associated agonist of cell death (Bad), and the anti-apoptotic Bcl-2, with initiation varying by pathway yet culminating in the activation of executioner caspase-3 (CASP3) (Taddei et al., 2012).

Neuronal vulnerability in the pathological or aging brain cannot be fully accounted for by canonical apoptosis and/or anoikis. Instead, neurons and glial cells increasingly rely on non-canonical regulated cell death (RCD) pathways whose activation thresholds are shaped by long-term epigenetic remodeling, metabolic decline, and chronic stress exposure. In this integrated view, cell death is not a terminal binary event but the outcome of progressive shifts in transcriptional permissiveness, chromatin accessibility, and metabolic resilience.

Non-canonical RCD refers to several regulated mechanisms of cellular demise that fall outside the classical apoptosis–necrosis dichotomy and, in many cases, do not strictly conform to the molecular hallmarks originally used to define these two primary forms of cell death (Table 1; Figure 1). These types of death include necroptosis; several forms of cell death with cytoplasm vacuolization (autophagy, entosis, paraptosis, methuosis); other forms of cell death associated with mitochondrial dysfunction (mitochondrial permeability transition–regulated necrosis, mitoptosis, parthanatos, oxeiptosis); metal accumulation (ferroptosis, cuproptosis); or immunoreactive processes (pyroptosis, NETosis) – for official nomenclature see (Galluzzi et al., 2018). In neural tissues—where post-mitotic status, high metabolic demand, and complex neuron–glial interactions impose unique constraints—non-canonical cell death mechanisms are particularly relevant in development, aging, and disease. In addition, in neurons, there are context-dependent or hybrid forms of death that combine elements of multiple pathways or deviate from canonical definitions (Yuan et al., 2003; Fricker et al., 2018). Nonapoptotic RCDs are signal-dependent, inhibitable, and mechanistically distinct from accidental necrosis. They are characterized by caspase independence, although there may be crosstalk with caspase-mediated death. Depending on cell type, metabolic state, redox balance, and the inflammatory milieu, RCD can follow a different pathway. Importantly, numerous nonapoptotic RCDs are inherently pro-inflammatory, influencing neuroimmune responses (Duan et al., 2022; Yan et al., 2025). Several types of nonapoptotic RCDs are characterized by cell membrane rupture, allowing cellular contents to leak into the surrounding tissue. They are now classified as forms of lytic RCD (Wang et al., 2025). Pyroptosis, necroptosis, ferroptosis, and PANoptosis are the main types of lytic death. Broadly speaking, they arise when infections, cellular injury, or immune system signaling occur.

TABLE 1. Main features of the different types of non-canonical RCDs in neurons and glia. For references, see the main text. Abbreviations are found in Supplementary Material.

Without classical apoptotic features.

Epigenetic drift associated with aging—marked by DNA methylation erosion, altered histone modification landscapes, and alterations in chromatin architecture—regulates the expression of critical modulators of necroptosis, ferroptosis, parthanatos, and other inflammatory death pathways (Liu et al., 2025). These alterations diminish the threshold for nonapoptotic cell death, especially in post-mitotic neurons exhibiting restricted regeneration potential. With age, genes that are typically silenced can be reactivated by histone acetylation and the removal of repressive chromatin marks. Therefore, RCD pathways can be initiated/restarted. The aging process involves changes in several types of epigenetic modifications (Lossi et al., 2024), including DNA methylation, histone modifications, chromatin remodeling, noncoding RNAs, and RNA modifications in neurons. These changes govern the progression of all the main non-canonical RCD pathways (Liu et al., 2025). In parallel, stable epigenetic reprogramming toward a pro-inflammatory phenotype occurs in microglial aging (Márquez-Ropero et al., 2020). Thus, epigenetic control of glial death pathways indirectly dictates neuronal survival.

In the following, we will discuss the different types of nonapoptotic RCDs, including the molecular and cellular pathways involved, the most relevant structural features that distinguish each type of death from the others, and their epigenetic regulation. For epigenetic regulation, a hierarchical framework will be used by distinguishing: (i) direct epigenetic regulation of RCD genes (e.g., promoter DNA methylation or defined histone marks), (ii) epigenetic permissiveness or priming through long-term chromatin remodeling, and (iii) indirect or associative links involving metabolic or stress-responsive pathways that converge on epigenetic modifiers.

Main types of nonapoptotic RCDs in neural cells and their epigenetic regulation

Necroptosis

According to the recommendations of the Nomenclature Committee on Cell Death 2018, necroptosis (also referred to as programmed necrosis) is a form of RCD triggered by perturbations of extracellular or intracellular homeostasis that critically depends on mixed lineage kinase domain-like protein (MLKL), receptor-interacting protein kinase 3 (RIPK3), and (at least in some settings) on the kinase activity of receptor-interacting protein kinase 1 (RIPK1) (Galluzzi et al., 2018). Necroptosis (Figure 2) is activated when death receptor signaling, e.g., through tumor necrosis factor receptor 1 (TNFR1) or toll-like receptors (TLRs), occurs in the context of caspase-8 (CASP8) inhibition (Vercammen et al., 1998; Denecker et al., 2001). After oligomerization, MLKL initiates necroptosis by binding to phosphatidylinositol lipids and cardiolipin. Specifically, MLKL translocates from the cytosol to the plasma membrane and intracellular membranes, where it directly disrupts membrane integrity, leading to plasma membrane permeabilization and necrotic cell death (Wang H. et al., 2014). Necroptosis is often suppressed under physiological conditions by CASP8, making its activation context-specific in neurons, oligodendrocytes, and microglia (Fritsch et al., 2019). Neuronal necroptosis usually displays delayed or partial execution, consistent with neurons’ energetic constraints. In the nervous system, necroptosis has a role in ischemia–reperfusion injury, amyotrophic lateral sclerosis (ALS), multiple sclerosis (MS), Parkinson’s disease (PD), and Alzheimer’s disease (AD). A family of RIPK1 inhibitors, the necrostatins, can inhibit necroptosis (Smith and Yellon, 2011).

Necroptosis is tightly controlled at the transcriptional level under physiological conditions. A seminal study in mutant mouse embryonic fibroblasts has demonstrated that 2-hydroxyglutarate inhibits necroptosis via DNA methyltransferase 1 (DNMT1)-dependent hypermethylation of the RIPK3 promoter, providing a clear example of metabolite-driven, promoter-specific epigenetic gating of a core component of necroptosis (Yang et al., 2017). In young neural tissue, expression of RIPK1, RIPK3, and MLKL is restrained, while CASP8 activity suppresses necroptotic signaling. On the contrary, in brain aging (Thadathil et al., 2021), AD (Caccamo et al., 2017; Koper et al., 2020), PD (Iannielli et al., 2018; Oñate et al., 2020), ALS (Ito et al., 2016), and retinal degeneration (Huang et al., 2018), there is an increase of RIPK1, RIPK3, and MLKL. Data on the link between necroptosis and direct epigenetic regulation of its regulating genes in neurons are summarized in Table 2. Changes in DNA methylation and hydroxylation affect both mitochondrial DNA (mtDNA) replication and gene expression levels (Coppedè and Stoccoro, 2019). The mtDNA copy number was increased in SOD1- or C9orf72-mutated ALS patients, suggesting that an exacerbated motoneuronal death occurs through a necroptotic mechanism (Giallongo et al., 2022). At the cell level, the epigenetic modulation of RIPK3 by transglutaminase 2 (TG2)-dependent serotonylation of trimethylated lysine 4 of histone H3 (H3K4me3) positively affects necroptosis in cultured fibroblasts (Vecchio et al., 2025). Another very recent paper has shown that the absence of CASP8 promotes indirect epigenetic reprogramming through metabolic-responsive pathways that converge on epigenetic modifiers, leading to a neural stem cell-like state in neuroendocrine-derived human small cell lung cancer cells (Androulidaki et al., 2025).

TABLE 2. Summary of the epigenetic regulation of non-canonical RCDs in neurons and glia. For references, see the main text. Abbreviations are found in Supplementary Material.

Mitoptosis also acts as a regulator ofthe epigenetic landscape.

Ferroptosis

According to the recommendations of the Nomenclature Committee on Cell Death 2018, ferroptosis is a form of RCD initiated by oxidative perturbations of the intracellular microenvironment, under constitutive control of glutathione peroxidase 4 (GPX4) and inhibited by iron chelators and lipophilic antioxidants (Galluzzi et al., 2018). As the name suggests, ferroptosis is iron-dependent. It leads to an excessive, harmful accumulation of phospholipid hydroperoxides, as a failure of GPX4-mediated detoxification. Besides GPX4, other key regulators include System Xc, a cystine/glutamate antiporter also known as solute carrier family 7 member 11 (SLC7A11), and acyl-CoA synthetase long-chain family member 4 (ACSL4) (Figure 2).

Key features of ferroptosis include the inactivation of GPX4 and the accumulation of oxidized phospholipids inside the cell. There is no caspase activation or DNA fragmentation. Neurons are highly susceptible to ferroptosis because they contain high levels of polyunsaturated fatty acids (PUFA), display elevated mitochondrial activity, limited capacity for lipid turnover, and exhibit elevated oxidative metabolism. Ferroptosis intersects with aging-related redox imbalance and may occur in PD, AD, ALS, Huntington’s disease, and brain aging. Importantly, ferroptosis does not produce classical apoptotic hallmarks and may coexist with mitochondrial dysfunction and synaptic failure long before neuronal soma loss becomes evident. Several substances act as ferroptosis inhibitors, including iron chelators, ferrostatin-1, methyltransferase-like 3 (METTL3), the key enzyme for N6-methyladenosine (m6A) RNA modification, and liproxstatin-1.

Ferroptosis represents a paradigmatic example of how metabolic decline and epigenetic regulation intersect (Table 2), as neurons require continuous expression of antioxidant and lipid-regulating genes, such as GPX4, SLC7A11, and iron-buffering proteins. Therefore, the mechanisms of epigenetic regulation of ferroptosis include direct DNA methylation and histone modifications, as well as epigenetic permissiveness or priming through long-term chromatin remodeling. Age-related changes in DNA methylation and chromatin accessibility at GPX4 and SLC7A11 loci reduce transcriptional robustness, rendering neurons vulnerable to iron-dependent lipid peroxidation. Ferroptosis suppressor protein 1 (FSP1) is a central regulating molecule of ferroptosis. FSP1 operates via the FSP1-coenzyme Q10 (CoQ10)-NAD(P)H axis and the vitamin K redox cycle. FSP1 is governed by upstream elements, such as transcription factors and noncoding RNA (ncRNA), and is influenced by direct epigenetic alterations that impact the advancement of FSP1-associated disorders (Li et al., 2023). These alterations, in turn, affect the expression of, e.g., solute carrier family 2 member 1 (SLC2A1), which is vital for transporting glucose across cell membranes, especially into the brain (Barros et al., 2017), and 3-hydroxy-3-methylglutaryl-CoA lyase (HMGCL), a crucial mitochondrial enzyme in ketogenesis (Liśkiewicz et al., 2021) through direct epigenetic regulation. Importantly, ferroptosis does not require active transcription at the time of execution; instead, it reflects a failure of long-term epigenetic maintenance programs that usually preserve redox homeostasis (Table 2). Therefore, ferroptotic degeneration frequently manifests as a gradual, prodromal process marked by the progressive accumulation of cellular damage, especially iron-dependent lipid peroxidation, before the emergence of overt features. Ferroptotic neurodegeneration involves nuanced iron dysregulation, diminished antioxidant defenses (e.g., GPX4), and increased neuronal stress, ultimately culminating in synaptic dysfunction and axonal loss before definitive neuronal death, as observed in conditions such as AD and PD (Abdukarimov et al., 2025). In the autoimmune encephalitis mouse model of MS, the histone methyltransferase (HMT) G9a facilitates the formation of the repressive epigenetic marker H3K9me2 (dimethylated lysine 9 of histone 3), resulting in the suppression of anti-ferroptotic gene expression, particularly of GPX4, the catalytic subunit of glutamate-cysteine ligase (GCLC), and cystathionine-β-synthase (CBS). As a result, G9a depletes glutathione (GSH) and induces lipid peroxidation, ultimately leading to ferroptosis (Rothammer et al., 2022). H3K14 lactylation exacerbates neuronal ferroptosis by inhibiting calcium efflux following intracerebral hemorrhagic stroke (Sun et al., 2025). ncRNA epigenetic modifications, including miR-214, miR-124, miR-106b-5p, miR-122-5p, and miR-23a-3p also regulate ferroptosis (Candelaria et al., 2021; Chen et al., 2021). Other noncoding RNA modifications appear to regulate ferroptosis in PD. In the 6-hydroxydopamine (6-OHDA)-induced PD disease model, miR-335 binds the 3′-untranslated region of ferritin heavy chain 1 (FTH1) mRNA, leading to reduced FTH1 protein expression. FTH1 can inhibit the Fenton reaction involving Fe³⁺ and Fe²⁺, thereby reducing lipid reactive oxygen species (ROS) production. Furthermore, downregulation of FTH1 protein expression leads to intracellular free iron accumulation and mitochondrial dysfunction, ultimately resulting in ferroptosis of dopaminergic neurons (Tian et al., 2020). The YTH N6-methyladenosine RNA binding protein 1 (YTHDF1) recognizes and binds m6A-containing mRNAs, and regulates their stability (Wang X. et al., 2014). It promotes p53 translation and induces ferroptosis during acute cerebral ischemia/reperfusion through m6A-dependent binding (Chang et al., 2025). Very recently, METTL3, a known inhibitor of ferroptosis, was demonstrated to promote recovery from traumatic brain injury (TBI) by regulating m6A modification and RNA stability of GPX4 (Qin et al., 2025). In all cases above, epigenetic regulation occurs indirectly through metabolic or stress-responsive pathways that converge on epigenetic modifiers.

Parthanatos

According to the recommendations of the Nomenclature Committee on Cell Death 2018, parthanatos is a form of RCD initiated by poly (ADP-ribose) polymerase 1 (PARP1) hyperactivation and precipitated by the consequent bioenergetic catastrophe, coupled with apoptosis-inducing factor (AIF) and macrophage migration inhibitory factor (MIF)-dependent DNA degradation (Galluzzi et al., 2018) (Figure 2). It is today well known that PARP1 has a dual function in cell death and survival. As a protective factor, PARP1 is activated to recruit DNA repair proteins and nucleases to sites of DNA damage, thereby facilitating DNA repair (Beck et al., 2014). Conversely, hyperactivation of PARP1 facilitates parthanatos in stroke (Liu et al., 2022). In parthanatos, severe DNA damage causes PARP1 to become hyperactive, depleting NAD⁺ and ATP, and translocating AIF from the nucleus to the mitochondria (Wang et al., 2016). Parthanatos is particularly significant in neurons, given their substantial dependence on oxidative metabolism, and it contributes to excitotoxicity, PD, and ischemia-reperfusion injury, characterized by severe oxidative DNA damage. Given the age-related decline in NAD⁺ availability, parthanatos may represent a key convergence point between aging, metabolic failure, and neuronal vulnerability. Under this perspective, one has to consider the role of sirtuins, a family of NAD⁺-dependent proteins crucial for metabolism, stress response, DNA repair, and cell survival, often called “longevity genes” due to their role in the aging process, as they appear to be essential in the suppression of the more common diseases of aging, including cardiovascular and neurodegenerative diseases, obesity, cancer, etc. (Morris, 2013).

Parthanatos also plays a key role in linking genetic instability to metabolic fatigue. Therefore, epigenetic regulation of parthanatos may occur at different hierarchical levels. The aging brain is characterized by the accumulation of oxidative DNA damage, which is aggravated by age-related chromatin relaxation and the reduced expression of DNA repair mechanisms (Xu et al., 2023; Moura et al., 2024). In keeping with these observations, an indirect epigenetic regulation of RCD genes occurs after hyperphosphorylation of histone H2Ax in a P53-knockout, poorly differentiated colon carcinoma cell line, which activates the parthanatos pathway (Dinhof et al., 2020). Excessive PARP1 activation under these conditions leads to catastrophic NAD⁺ depletion and mitochondrial failure (Xu et al., 2023). In a broad sense, PARP1 is emerging as a significant player in chromatin regulation, leading to epigenetic permissiveness/priming through long-term chromatin remodeling. PARP1 has been shown to link to histone acetylation and methylation, with PARP1 targets including both histone deacetylases (HDACs) and histone proteins (Ciccarone et al., 2017). From an epigenetic standpoint, decreasing NAD⁺ levels hinder the function of sirtuins and other chromatin-modifying enzymes, exacerbating transcriptional instability and increasing susceptibility to parthanatos (Huang et al., 2014; Liang et al., 2023; Moshtaghion et al., 2024). This feed-forward loop, an indirect link involving metabolic or stress-responsive pathways that converge on epigenetic modifiers, is critical in situations where neurons are already low on energy, such as in excitotoxic and ischemic conditions. The epigenetic significance of PARP1 and PARylation (the addition of ADP-ribose chains to proteins, often after DNA damage) is under investigation. PARP1 and PARylation may indirectly activate epigenetic modifiers via distinct metabolic pathways. PARylation seems to occur on the DNA molecule itself (Talhaoui et al., 2016) and to interact with DNA methylation, potentially influencing its mechanisms (Ciccarone et al., 2017). Sirtuin 6 (SIRT6) interacts physically with PARP1 and mono-ADP-ribosylates PARP1 at lysine 521 following oxidative stress or AKT inhibition (Mao et al., 2011). This change in PARP1 greatly increases its ability to poly-ADP-ribosylate (PARylate), leading to AIF nuclear translocation and tumor growth suppression (Zhang et al., 2022). PARylation of deteriorating photoreceptors in the mouse retinal degeneration 1 (rd1) model exhibited considerable overlap with deacetylated photoreceptor nuclei, suggesting increased HDAC activity potentially associated with PARylation (Sancho-Pelluz et al., 2010). Poly (ADP-ribose) glycohydrolase (PARG), in contrast, removes ADP-ribose chains, and its silencing after DNA methylation leads to an increase in parthanatos (Pascal and Ellenberger, 2015), as poly (ADP ribose) signals to mitochondrial AIF to trigger this type of RCD (Wang et al., 2009). The lysine demethylase 6B (KDM6B) gene can also be epigenetically regulated to control parthanatos. KDM6B encodes a lysine-specific demethylase that removes methyl groups from di- or tri-methylated lysine 27 of histone H3 (H3K27me2 or H3K27me3). H3K27me3 is a repressive epigenetic mark that controls how chromatin is organized and how genes are turned off (Jones et al., 2018). KDM6B can inhibit O6-methylguanine-DNA methyltransferase (MGMT), a crucial DNA repair protein, resulting in elevated levels of damaged DNA within the cell and, consequently, hyperactivation of PARP1 (Yang et al., 2022). Thus, in this case, parthanatos is primed through long-term chromatin remodelling and a metabolic pathway involving indirect epigenetic mechanisms.

Autophagy-dependent cell death (ADCD)

According to the recommendations of the Nomenclature Committee on Cell Death 2018, ADCD is a form of RCD that mechanistically depends on the autophagic machinery (Galluzzi et al., 2018) (Figure 2). True ADCD remains controversial in neurons, where it may be linked to developmental neuronal pruning, neurodegenerative disorders with impaired protein aggregation, and lysosomal storage disorders. Notably, ADCD must be distinguished from autophagy, which is not causally linked to death. ADCD depends on autophagy-related (ATG) proteins, a conserved family essential for autophagy, which is vital for maintaining neuronal and glial function; consequently, disrupted autophagy is associated with various neurologic diseases (Fleming et al., 2025). Besides ATG proteins, ADCD relies on Unc-51-like autophagy activating kinase 1 (ULK1), PI3K, and microtubule-associated protein 1 light chain 3 (LC3) (Yu et al., 2018). This kind of RCD is highly dependent on cellular conditions, as autophagy usually protects neurons by maintaining proteostasis and organelle quality control. But a pathological shift to ADCD occurs when there is long-term stress, lysosomal malfunction, and protein aggregation (Malik et al., 2024).

There are three basic forms of ADCD based on molecular and structural characteristics: bulk, organelle-specific, and autosis (Eickhorst et al., 2020). Excessive bulk autophagy is marked by the substantial buildup of autophagosomes and autolysosomes, resulting in the excessive consumption of cytoplasmic constituents and organelles. Excessive organelle-specific autophagy involves selective, hyperactivated degradation of specific organelles, such as mitochondria (mitophagy) (Basak and Holzbaur, 2025) or the endoplasmic reticulum (ER) (Cai et al., 2016; Merighi and Lossi, 2022), which are essential for cellular energy or protein synthesis. Mitophagy appears to play a role in aging and neurodegeneration, as evidenced by alterations in mtDNA epigenetic markers and an elevation in mtDNA deletions in the brains of aged mice (Tanhauser and Laipis, 1995; Dzitoyeva et al., 2012). In various organisms, mitophagy correlates with longevity and lifespan, as oxidative damage to mtDNA in the brain connects with decreased lifespan in birds and mammals (Herrero and Barja, 1999). Finally, autosis is a specialized form of ADCD, characterized by nuclear morphological changes and focal perinuclear swelling. It relies exclusively on the Na⁺/K⁺-ATPase pump (Depierre et al., 2024).

Direct epigenetic processes strictly control autophagy. But if lysosomal and autophagy-related genes are chronically epigenetically dysregulated, autophagy can change from being prosurvival to prodeath (Malik et al., 2024). In aged neurons and in proteopathies, incomplete or excessive autophagic flux may be necessary for cell death, blurring the distinction between adaptive and maladaptive responses. This transition is not abrupt but reflects a gradual erosion of transcriptional coordination among proteostasis, lysosomal function, and mitochondrial quality control pathways. Several histone modifiers can repress autophagy in various neurodegenerative diseases. These include mutations in enhancer of zeste 2 polycomb repressive complex 2 subunit (EZH2), a critical regulator of numerous developmental genes that increases H3K27me3 levels, a repressive histone mark that inhibits transcription of autophagy-related genes (Cao et al., 2002). Another repressive histone mark associated with silenced autophagy-related gene expression is H3K9me (Nakayama et al., 2001; Barski et al., 2007). Other epigenetic modifications, instead, led to increased autophagy gene expression and activity. Trimethylation of H3K36, sustained by an HMT encoded by SET domain containing 2 histone lysine methyltransferase (SETD2), promotes transcriptional elongation and RNA splicing (Kim et al., 2011). The loss of SETD2 leads to alternative splicing of ATG12, a process essential for autophagosome formation (González-Rodríguez et al., 2020). Lysine demethylase 6A (KDM6A) and KDM6B are two additional histone lysine demethylases linked to neurodegenerative diseases and shown to be involved in autophagy. These epigenetic modifiers particularly demethylate the repressive histone marks H3K27me2 and H3K27me3, which activate transcription of genes associated with autophagy (Agger et al., 2007). A lack of lysine (K)-specific demethylase 1A (KDM1A), the first histone lysine-specific demethylase discovered, caused zebrafish larvae to have moving, learning, and memory defects, as well as a buildup of autophagy and apoptosis (Zou et al., 2025). Histone modifiers also play a role in negative feedback loops that prevent autophagy from persisting for too long. Among these, three neurodegenerative disease-related HMTs are involved in the increase in H3K4me3 levels, and hence have a role in autophagy-related gene repression: SET domain containing 1A histone lysine methyltransferase (SETD1A), KDM1A, and lysine methyltransferase 2A (KMT2A) (Lewerissa et al., 2024). Other classes of epigenetic modifiers control autophagy-related gene expression related to neurodegeneration indirectly and include ATP-dependent chromatin remodelers, the transcription factors Yin Yang 1 (YY1) and YY1-associated protein 1 (YY1AP1), and DNA methyltransferases, including DNA methyltransferase 3 alpha (DNMT3A) and methyl-CpG-binding protein 2 (MECP2) (Lewerissa et al., 2024).

Pyroptosis (non-canonical inflammasome signaling)

According to the recommendations of the Nomenclature Committee on Cell Death 2018, pyroptosis is a type of RCD that critically depends on the formation of plasma membrane pores by members of the gasdermin protein family, often (but not always) as a consequence of inflammatory caspase activation (Galluzzi et al., 2018) (Figure 2). Pyroptosis depends on inflammasomes that contain nucleotide-binding oligomerization domain, leucine-rich repeat, and pyrin domain (NLRP). These are a group of cytosolic multiprotein complexes that are essential innate immune sensors. NLRP-containing inflammasomes trigger the activation of caspase-1 (CASP1), which causes the production of interleukin 1β (IL-1β) and interleukin 18 (IL-18). This process also makes the cell membrane permeable through pores formed by gasdermin E (GSDME) (Xu et al., 2024). Pyroptosis mostly occurs in microglia and, to a lesser extent, in astrocytes, contributing to neuroinflammation rather than direct neuronal loss and, by extension, indirectly aggravating neuronal degeneration (Cai et al., 2025). Mature neurons never undergo complete pyroptosis; however, inflammasome signaling in glial cells exacerbates neuroinflammation and indirectly facilitates neuronal degeneration. Pyroptosis thus facilitates disease progression rather than directly causing neuronal death. Several direct epigenetic mechanisms regulate pyroptosis. Epigenetic regulation of pyroptosis in hippocampal neurons was shown to be promoted by histone deacetylase 1 (HDAC1) through the miR-15a-5p/CASP1 axis (Lv et al., 2025). Within the hippocampus, overexpression of the SET domain-containing 1B histone lysine methyltransferase (SETD1B) increased H3K4me3 levels, reduced neuronal pyroptosis, and mitigated cognitive impairment (Wang et al., 2024). Similarly, miRNA-140-3p safeguarded hippocampus neurons from pyroptosis (Wu et al., 2022). Other research has shown that histone acetylation has a function in controlling pyroptosis. Acetylation of H3 and H4 at the nucleotide-binding oligomerization domain, leucine-rich repeat, and pyrin domain 3 (NLRP3) promoter, regulated by signal transducer and activator of transcription-3 (STAT3), can upregulate NLRP3. Upregulation of NLRP3 could lead to painful neuropathy and pyroptosis caused by bortezomib (Liu et al., 2018). Inhibition of histone deacetylase 2 (HDAC2) guards against oxidative stress by maintaining growth/differentiation factor 15 (GDF15) levels via the acetylation at lysine 27 of histone H3 (H3K27ac), whereas histone deacetylase 6 (HDAC6)-mediated deacetylation of HSP90 activates microglial NLRP3, which induces neuroinflammation (Ri et al., 2024; Zhou et al., 2024), and contributes to cognitive decline in mouse hippocampus (Takada et al., 2021). Irregular methylation patterns in inflammasome components, such as NLRP3 and apoptosis-associated speck-like protein containing a CARD (ASC), have also been linked to amplified pyroptotic activity (Huang et al., 2021). METTL3 is frequently activated in pathological conditions, leading to elevated m6A levels on target mRNAs, such as NLRP3, thereby facilitating pyroptosis (Guan et al., 2025). Moreover, the m6A methylation of si-peroxisome proliferator-activated receptor γ (PPARγ) RNA by methyltransferase-like 14 (METTL14) promotes activation of the absent in melanoma 2 (AIM2) inflammasome, thereby contributing to neuronal pyroptosis (Wu et al., 2024).

Other emerging pathways of non-canonical nonapoptotic RCDs

Several other forms of RCDs may occur in neurons or glia. Most of these forms have been described initially in, e.g., neoplastic cells, remain poorly characterized in neural cells, and their relevance to neuronal physiology and pathology is still not fully appreciated. Moreover, neurons frequently exhibit hybrid death phenotypes that combine features of apoptosis, necrosis, and metabolic collapse—compartmentalized degeneration—such as axonal or synaptic loss preceding soma death, further challenging traditional classifications.

Cuproptosis

Cuproptosis is a well-described metal-dependent death program (Xue et al., 2023) (Figure 2). Solute carrier family 31 member 1 (SLC31A1) is the human gene that codes for high-affinity copper uptake protein 1 (CTR1), a critical cell membrane protein responsible for importing copper into cells (Petris et al., 2003). Cuproptosis key features are the Cu-dependent aggregation of lipoylated mitochondrial proteins and the consequent proteotoxic stress (Zhang L. et al., 2024). Key players include ferredoxin 1 (FDX1), which reduces Cu²⁺ to the toxic Cu⁺; lipoic acid synthase (LIAS) and lipoyltransferase 1 (LIPT1), which activate the lipoylation machinery; and mitochondrial enzyme dihydrolipoamide S-acetyltransferase (DLAT), which disrupts cellular energy production and protein homeostasis. Copper causes DLAT to aggregate, disrupting the tricarboxylic acid (TCA) cycle and Fe-S cluster proteins, leading to proteotoxic stress and cell death (Liu et al., 2025). It is unclear whether cuproptosis is relevant to the nervous system. Recent investigations have associated cuproptosis with AD (Chen G. et al., 2024), PD (Huang et al., 2024), and cognitive impairment after lead exposure, facilitated by the administration of DLAT (Zhao Y. et al., 2025). Epigenetic regulation of cuproptosis mainly occurs through indirect mechanisms that modulate cellular copper levels. Among histone modifications, an increase in H3K27ac increases ATPase copper-transporting β (ATP7B) expression, enabling cells to expel copper into the extracellular environment more effectively. ATP7B is, in fact, primarily involved in the pathogenesis of Wilson’s disease, which is characterized by hepatic and brain copper accumulation (Sarode et al., 2021). Epigenetic regulators of cuproptosis in non-neural cells comprehend the long intergenic noncoding RNA 1614 (LINC01614), the miRNAs miR-204-5p and miR-338-3p (Takahashi et al., 2017; Wu et al., 2023). These regulators act indirectly on the epigenetic modifiers. LINC01614 acts as a competitive endogenous RNA for miR-204-5p. It binds to miR-204-5p, thereby blocking its targeting and degradation of mRNA targets, notably SLC31A1, a negative regulator of cuproptosis. The miR-204-5p block increases SLC31A1 levels, enhancing copper internalization by cells and promoting cuproptosis. The role of miR-338-3p in cuproptosis, in contrast, remains poorly understood. However, this microRNA has been implicated in the regulation of the apoptosis-associated tyrosine kinase (AATK) mRNA. This kinase plays roles in differentiation, apoptosis, and possibly neuronal degeneration (Kos et al., 2012). Other indirect epigenetic regulators of cuproptosis include EZH2 and metal-regulatory transcription factor 1 (MTF1). EZH2 participates in histone methylation and, ultimately, transcriptional repression by catalyzing the addition of methyl groups to histone H3 at lysine 27, thereby acting as a positive regulator by silencing protective factors such as nuclear factor erythroid 2-related factor 2 (Nrf2) or SLC7A11 (Yang et al., 2023). MTF1, instead, is a negative regulator (Westin and Schaffner, 1988). Recent observations have implicated histone methylation in reducing neuronal cuproptosis in sepsis-associated encephalopathy (Zhang Y. et al., 2025).

Cell death associated with mitochondrial dysfunction

Mitochondrial permeability transition (MPT)–regulated necrosis

The mitochondrial permeability transition (MPT) is a rapid, irreversible increase in the permeability of the inner mitochondrial membrane to small solutes up to 1.5 kDa, leading to the gradual dissipation of mitochondrial transmembrane potential (Δψm) (Bernardi and Pavlov, 2022). Elevated intracellular calcium concentrations and ROS induce MPT. Pore formation renders the inner mitochondrial membrane accessible to tiny solutes through the mitochondrial permeability transition pore (mPTP). The unregulated influx of water into the mitochondrial matrix leads to the osmotic disintegration of the organelle, hence suspending mitochondrial functions, including ATP synthesis (Robichaux et al., 2023). In the 1990s, the process that finally triggers a type of RCD in the MPT was demonstrated. This type of RCD often (but not always) has necrotic morphological traits (Crompton and Costi, 1990) (Figure 2). The induction of necrosis or apoptosis upon MPT onset is linked to the cell’s ATP levels: elevated ATP levels facilitate the initiation and progression of apoptosis, whereas diminished energy levels favor the necrotic route (Eguchi et al., 1997). Sustained mitochondrial Ca²⁺ overload and oxidative stress trigger MPT–regulated necrosis that may occur in acute brain injury (ABI) and excitotoxic neuronal death (Robichaux et al., 2023). Its key component is the cyclophilin D (CyPD)–dependent MPT pore. Consequently, the cell suffers bioenergetic collapse and adopts a necrotic shape. MPT-regulated necrosis plays a significant role in various neurodegenerative disorders, such as AD, PD, ALS, and frontotemporal dementia (Morganti et al., 2018; Baev et al., 2024). The link, if any, between MPT-regulated necrosis and the epigenetic landscape remains to be thoroughly investigated. It seems reasonable that epigenetic regulation of MPT–regulated necrosis occurs through indirect metabolic mechanisms. DNA methylation and histone acetylation of the peptidyl-prolyl cis-trans isomerase F, mitochondrial (PPIF) gene, which encodes CyPD, regulate the threshold for mPTP opening (Yu et al., 2023). The HDAC4/5 inhibitor LMK-235 restores voluntary movement in MPTP-induced PD mice, suggesting that histone deacetylase inhibitors act indirectly to increase vesicular monoamine transporter 2 (VMAT2) expression and thus ameliorate PD symptoms (Lee et al., 2025).

Mitoptosis

Mitoptosis is a distinct mechanism for mitochondrial removal that differs from mitophagy (Figure 2). Mitochondria undergo fission, perinuclear aggregation, encapsulation into mitoptotic bodies, and ensuing expulsion from the cell, rather than lysosomal degradation (Tinari et al., 2007). Key molecules in mitoptosis include the Bcl-2-associated X protein (Bax) and Bcl-2 homologous antagonist/killer (Bak), which facilitate pore formation in the mitochondrial membrane (Brunelle and Letai, 2009); the translocase of the inner mitochondrial membrane 8 (TIMM8a), which is involved in mitochondrial protein import and potentially stress response (Roesch et al., 2004); and dynamin-related protein 1 (Drp1) that drives mitochondrial fission (Wang et al., 2015; Otera et al., 2016). The two main types of mitoptosis are the inner membrane type (loss of matrix/cristae) and the outer membrane type (rupture of the outer membrane, releasing cristae debris). Nonetheless, they are fundamentally independent of the conventional autophagy machinery, although hybrid forms may occasionally occur. Mitoptosis has a role during neural development, as nerve growth factor (NGF) -induced differentiation of the neural tube is linked to mitoptosis mediated by activating beclin1-regulated autophagy (Ambra-1) (Yazdankhah et al., 2014).

Despite being contentious, data indicate that methylation of mtDNA (e.g., 5-methylcytosine or N6-methyladenine) directly influences mitochondrial gene expression and organelle function. Enzymes such as mitochondrial DNA methyltransferase 1 (mtDNMT1) are found in the mitochondrial matrix, underscoring mitochondria’s key position as an energy and epigenetic center controlling nuclear function, development, and pathology (Xu et al., 2026).

Oxeiptosis

Oxeiptosis is a death signaling pathway that couples a ROS sensor to a non-canonical cell-death execution pathway (Figure 2), comprising Kelch-like ECH-associated protein 1 (KEAP1), phosphoglycerate mutase family member 5 (PGAM5), and apoptosis-inducing factor mitochondrion-associated 1 (AIFM1) signaling (Scaturro and Pichlmair, 2018). The appearance of oxeiptotic cells resembles that of cells undergoing apoptosis, characterized by cellular shrinkage and extensive DNA damage and fragmentation. Recent findings have associated oxeiptosis with AD (Zhong et al., 2024) and PD (Rong et al., 2026). The link between oxeiptosis and the epigenetic landscape is that ROS may alter DNA and/or histones, thereby influencing the expression of genes involved in oxeiptosis. Regulation of epigenetic modifiers is indirect and involves multiple metabolic pathways under both normal and pathological conditions. In oxeiptosis, high levels of ROS could reduce S-adenosylmethionine (SAM), a significant methyl donor. Reducing SAM levels could cause widespread DNA hypomethylation and histone modifications, altering gene expression (Kietzmann et al., 2017). Nrf2 is a key player in the antioxidant response that protects cells from oxidative damage. It has two roles in how cells respond to oxidative stress. DNA hypermethylation inhibits Nrf2 and induces oxeiptosis (Holze et al., 2018). Conversely, inhibition or downregulation of HDACs (e.g., with pan-HDAC inhibitors such as vorinostat or trichostatin A) typically results in Nrf2 upregulation. The latter occurs because HDAC blockade hinders KEAP1 production, thereby facilitating Nrf2 accumulation and nuclear entry to perform a protective role (Wang et al., 2012). Nonetheless, excessive Nrf2 overexpression may disrupt the KEAP1 complex, thereby increasing cellular susceptibility to oxeiptosis when oxidative thresholds are exceeded (Scaturro and Pichlmair, 2018). Other indirect epigenetic regulators of oxeiptosis may be circular RNAs. Circular RNA forkhead box O3 (circFOXO3) has been associated with numerous diseases, including cancer, cardiovascular conditions, neurological disorders, aging, and inflammation (Zhao M. et al., 2025). It has been suggested that circFOXO3 plays a role in neurodegeneration, as it protects mouse hippocampus HT22 cells from glutamate-induced oxidative damage by regulating the mitochondrial apoptotic pathway (Lin et al., 2020). CircFOXO3 has also been demonstrated to promote oxeiptosis in intervertebral disk degeneration (Chen X. et al., 2024).

NETosis

NETosis-like processes involve immune cells, primarily neutrophils, releasing their DNA and antimicrobial proteins to trap pathogens (neutrophil extracellular traps–NETs) (Figure 2). NETs are extracellular strands of decondensed DNA, complexed with histones and neutrophil granule proteins (Sørensen and Borregaard, 2016). NETosis can differ from regular cell death (apoptosis/necrosis) in that the cell survives (vital NETosis), dies (suicidal NETosis), or releases mitochondrial DNA (mitoNETosis). NETosis-like processes are essential in infection, inflammation, and autoimmune diseases. Key molecules involved in NETosis are NADPH oxidase 4 (NOX4) (Diaconescu et al., 2025) and protein arginine deiminase 4 (PAD4) (Thiam et al., 2020). Cell alterations in NETosis include chromatin decondensation, breakdown of the nuclear envelope, and rupture of the plasma membrane (in suicidal NETosis) or release of vesicles (in vital NETosis) to eliminate DNA. In cerebral ischemia/stroke, NETs contribute to neuroinflammation and thrombosis (Kim et al., 2019; Lou et al., 2024). In AD (Kays et al. 2025), they accumulate around amyloid β (Aβ) plaques, resulting in persistent inflammation, compromised blood-brain barrier (BBB), and neuronal injury, thereby facilitating disease progression (Pietronigro et al., 2017; Chavoshinezhad et al., 2025). Epigenetic regulation of NETosis occurs via direct mechanisms. DNA demethylation triggers NETosis in neutrophil-like cells derived from HL-60 cells, concomitant with an elevation in citrullinated histone H3 (H3Cit) (Yasuda et al., 2020). Remarkably, different epigenetic modifications associated with NETosis have been observed in systemic lupus erythematosus and ANCA-associated vasculitides (Frangou et al., 2019).

Cell death with cytoplasm vacuolization

Paraptosis

Paraptosis is characterized by cytoplasmic vacuolation resulting from swelling of the ER and mitochondria, driven by impaired protein homeostasis, calcium overload, and ROS oxidative stress (Figure 2). It involves pathways including the insulin-like growth factor 1 (IGF-1) receptor (IGF-1R), mitogen-activated protein kinase (MAPK), and requires protein synthesis. Important cellular events are the expansion and vacuolization of the ER, the buildup of calcium in the mitochondria, the activation of MAPK and c-Jun N-terminal kinase (JNK), and the malfunctioning of ion channels, such as large-conductance, calcium-activated potassium channels and big potassium (BK) channels. Malfunction of ion channels leads to osmotic lysis and ATP depletion, independent of caspases. These key mechanisms are not exclusive, as they are at least partly shared with autophagy, cuproptosis, and apoptosis (Wu et al., 2020; Kist and Vucic, 2021). In vitro, intracellular acidification induced by the inhibition of the Na⁺/H⁺-exchanger leads to caspase-independent death of cerebellar granule neurons, which biochemically but not morphologically resembles paraptosis (Schneider et al., 2004). Later, it was shown that paraptosis was implicated in the neurodegeneration observed in mice homozygous for a transgene encoding Δ40p53 (p44), a short, naturally occurring isoform of the tumor suppressor p53 gene (Pehar et al., 2010). When overexpressed, p44 accelerates aging and shortens lifespan. In keeping with p44 biological activity, p44 transgenic mice displayed cognitive decline and synaptic impairment early in life. The synaptic deficits were attributed to hyperactivation of IGF-1R signaling and altered metabolism of the microtubule-binding protein tau (Pehar et al., 2010). However, it remains unclear whether specific neuronal cells, such as e.g., retinal ganglion cells, may undergo paraptosis under definite physiological or pathological conditions (Kunst et al., 2024). Growing evidence suggests that paraptosis is not simply a passive degenerative response to stress, but an epigenetically regulated process that requires permissive chromatin configurations and continuous gene expression (Sperandio et al., 2000). Epigenetic modifications regulate the expression of genes involved in a series of processes related to paraptosis, such as ER stress signaling, MAPK activation, calcium homeostasis, and proteostasis (Sen et al., 2016). Therefore, these modifications are indirect, as they converge on epigenetic modifiers via distinct metabolic or stress-responsive pathways. Recently, changes in histone code regulation at different hierarchical levels have been reported during the initiation of paraptosis-like death of HEp-2 tumor cells (Solovieva et al., 2024). In particular, these changes included the ubiquitination of histone H2B, histone cluster 1 H2B family member k (H2BC12), and different HMTs responsible for transcription and the repair of damaged DNA, as well as acetylating and ubiquitin-conjugating proteins. No data on similar histone modifications are available for neural cells. Paraptosis can be inhibited in vitro by epigenetic silencing or downregulation of programmed cell death 6-interacting protein (AIP1/Alix) (Sperandio et al., 2004).

Entosis

By definition, entosis is a type of RCD that originates from actomyosin-dependent cell-in-cell internalization and is executed by lysosomes (Galluzzi et al., 2018) (Figure 2). Entosis is a process in which live cells are engulfed by other cells. It differs from phagocytosis, as the Rho family of GTPases (Ras homolog family member A- RhoA; Ras-related C3 botulinum toxin substrate 1 - Rac1; Cell division cycle 42 - Cdc42) execute it and the downstream effector Rho-associated coiled-coil containing protein kinase (ROCK) generates actomyosin tension for invasion. The internalized cell is segregated into a vacuole, which often fuses with lysosomes to break down the cell in an atypical yet autophagy-associated manner. The cell can survive, divide, or be expelled (Kim et al., 2024). ECM detachment is one of the primary triggers. Cadherins and calcium help cells adhere to one another, and specific channels, such as ORAI1 (calcium release-activated calcium modulator 1), regulate calcium entry into cells (Overholtzer et al., 2007). Entosis is a pathogenic mechanism that causes microcephaly by activating p53 (Sterling et al., 2024). Protein associated with LIN7 1 (Pals1) mutants display microcephaly with entosis, which activates p53 and induces apoptosis (Sterling et al., 2023). This previously unknown pathogenic mechanism is a unique cellular dynamic in dividing cortical progenitors that causes cell death. Microencephaly is associated with reduced brain cortex due to massive cell death, with nuclei enveloped by plasma membranes within dividing cells. These structures represent a dynamic process characterized by prolonged mitosis and abnormalities in cytokinesis. Several recent models have shown that p53 activation causes cell death and microcephaly. In these animal models, p53 inactivation suppresses apoptotic cell death caused by mitotic delay or DNA damage (Sterling et al., 2024). Entosis has also been reported in gliomas (Han et al., 2024). Epigenetically, the methylation of the cadherin 1 (CDH1) gene indirectly interferes with entosis by regulating ATGs (Tedaldi et al., 2021).

Methuosis

Methuosis is a form of nonapoptotic cell death characterized by hyperactivated macropinocytosis, resulting in significant cytoplasmic vacuolization that compromises the cell membrane (Maltese and Overmeyer, 2014) (Figure 2). The underlying mechanisms include disruption of rat sarcoma virus (Ras)/Rac1/ADP-ribosylation factor 6 (Arf6) signaling, failure of macropinosome-lysosome fusion, and extensive cytoskeletal reorganization (specifically, disruption of actin and tubulin), ultimately leading to cellular collapse and the release of danger-associated molecular patterns (DAMPs), such as ATP. The principal molecular components comprise Rac1, Arf6, G-protein-coupled receptor kinase-interacting protein 1 (GIT1), RhoA, and ROCK1. The endosomal sorting complex required for transport-III (ESCRT-III) attempts to rectify the compromised membrane (and fails). Methuosis was described in differentiated cultures of SH-SY5Y neuroblastoma after exposure to methamphetamine (Nara et al., 2010), or after activation of the TrkA NFG receptor (Lavoie et al., 2005), and in ventral midbrain dopaminergic neuron cultures maintained under continuous excitatory amino acid blockade (Cubells et al., 1994). Epigenetic regulation of methuosis depends on Arf6, which leads to epigenetic permissiveness or priming via long-term chromatin remodeling. Certain chromatin accessibility patterns regulate Arf6 expression. Assays for transposase-accessible chromatin with sequencing (ATAC-seq) peaks reveal genome-wide patterns of open chromatin, identifying regulatory regions, such as enhancers and promoters, where DNA is accessible to transcription factors and thus indicative of active gene expression. ATAC-seq peaks are markedly enriched in the promoter and intragenic regions of the Arf6 gene, especially in prostate and lung malignancies. These regions have much open chromatin, which is linked to Arf6 overexpression and, consequently, methuosis (Qiu et al., 2022).

PANoptosis

PANoptosis is a form of interrelated RCD modalities first proposed by Malireddi and colleagues in 2019 (Malireddi et al., 2019). PANoptosis combines features of pyroptosis, apoptosis, and necroptosis and is driven by multiprotein complexes called PANoptosomes (Figure 3). The PANoptosome may consist of Z-DNA-binding protein 1 (ZBP1), an immune sensor; AIM2; RIPK1; or NLR family pyrin domain-containing 12 (NLRP12), a regulator of inflammation. Therefore, four distinct PANoptosomes have been described: the ZBP1-PANoptosome, the AIM2-PANoptosome, the RIPK1-PANoptosome, and the NLRP12-PANoptosome (Li and Qu, 2025). Of these, the ZBP1-PANoptosome and the AIM2-PANoptosome also contain NLRP3. The formation of PANoptosomes simultaneously activates multiple RCD pathways. Mitochondrial dysfunction signals the PANoptosome to trigger cell death.

PANoptosis has been implicated in several neurological disorders (Hou et al., 2025). For example, the AIM2-PANoptosome is involved in AD, as the buildup of Aβ oligomers intensifies neuronal cell death, RIPK1 protein aggregation, and mitochondrial autophagy failure via AIM2 inflammasome activation (Cao et al., 2021), knocking out AIM2 suppresses Aβ deposition and microglial activation (Wu et al., 2017), and ZBP1-PANoptosome activation in microglia causes neuronal mitochondrial ROS bursts and tau hyperphosphorylation, accelerating synapse loss (Meng et al., 2024). Recent data suggest that AIM2-mediated neuroinflammation plays a role in PD etiology beyond traditional RCD pathways. In MPTP-induced PD models, AIM2 demonstrates anti-inflammatory effects by inhibiting cyclic GMP-AMP synthase (cGAS)-dependent antiviral responses. Inhibition occurs by suppressing AKT-regulatory factor 3 (IRF3) phosphorylation, indicating an inflammasome-independent regulatory mechanism (Rui et al., 2022). Complementary pharmacological research suggests that crocin mitigates lipopolysaccharide (LPS)-induced neuroinflammation and dopaminergic neuron degeneration in PD models by downregulating the expression of both AIM2 and NLRP1 (Alizadehmoghaddam et al., 2024). In addition, α-synuclein fibrils activate the NLRP3 inflammasome to trigger PANoptosis in dopaminergic neurons, inducing a cascade that promotes substantia nigra pars compacta neuron degeneration (Nguyen et al., 2022). Other neurological conditions in which PANoptosis has been implicated are Huntington’s disease, ALS, and several forms of acute CNS injury (Hou et al., 2025).

The execution phase of lytic RCD

The execution phase of lytic RCD, often viewed as a passive process, is in fact an active mechanism that facilitates plasma membrane rupture by polymerizing the protein Ninjurin-1 (NINJ1) (Figure 3). In response to upstream signals (i.e., pyroptosis, necroptosis, ferroptosis, and PANoptosis), NINJ1 oligomerizes, forming linear amphipathic filaments (Filament/Hinge Model) or rings (Cookie-Cutter Model) that compromise the plasma membrane (Yang et al., 2025). Filaments formed by NINJ1 have a hydrophobic side that interacts with the cell membrane bilayer, disrupting its structure, creating or enlarging holes, and leading to extensive cellular lysis and the release of DAMPs. The process is two-step: pore-forming proteins (e.g., gasdermin D, GSDMD) initiate membrane permeability, and NINJ1 acts as the ultimate executioner of terminal lysis. Increased cytosolic calcium can act as a trigger or stabilizer, predisposing the membrane to NINJ1-mediated rupture (Wang et al., 2025). Still, the role of NINJ1 in the nervous system remains partly unclear. Peripherally, NINJ1 is mainly expressed in dorsal root ganglion (DRG) neurons and Schwann cells (Araki and Milbrandt, 1996), where it plays a role in neural regeneration after traumatic or nontraumatic peripheral nerve injury (Liu et al., 2021). In the CNS, NINJ1 has been detected in the meninges, the choroid plexus, and the parenchymal perivascular space of normal rat brains. It mediates endothelial adhesion in the brains of rats with experimental autoimmune encephalomyelitis (Ahn et al., 2009), and intervenes in ischemic stroke (Ifergan et al., 2011; Ahn et al., 2014).

Sialic acid-binding Ig-like lectins (SIGLECs) are predominantly found in immune cells and generate activating or inhibitory signals; they are also expressed on the surface of cells in the nervous system and have several functions in the normal and pathological brain (Siddiqui et al., 2019). Different SIGLECs perform different tasks in the nervous system. Sialic acid-binding Ig-like lectin 11 (SIGLEC11) is an inhibitory immune receptor expressed on human microglia. At the same time, the polysialylated neuronal cell adhesion molecule (PSA-NCAM) is a glycan molecular ligand found on neurons that binds to SIGLEC11 to promote neuroprotection (Wang and Neumann, 2010) and repair and regeneration (Saini et al., 2020). Recent studies have identified sialic acid-binding Ig-like lectin 12 (SIGLEC12) as a specific mediator of plasma membrane rupture downstream of MLKL (Noh et al., 2026) and NINJ1 as a shared execution factor required for terminal membrane rupture across multiple lytic RCD modalities (Kayagaki et al., 2021; Ramos et al., 2024), including necroptosis, pyroptosis, and ferroptosis. These molecules represent convergence points at the level of physical membrane disruption, conceptually downstream of pathway-specific signaling cascades. According to current data, SIGLEC12 is not expressed in the nervous system (https://www.proteinatlas.org/ENSG00000254521-SIGLEC12/tissue). Still, other SIGLECs, including SIGLEC1, 2, 3, 4, 11, F, and H, have been implicated in aging and numerous neurological conditions, including late-onset AD, multiple sclerosis, ALS, and ceroid lipofuscinosis (Siddiqui et al., 2019). It is therefore tempting to speculate that NINJ1 and SIGLEC12 may play a role in neurological conditions such as stroke, neurodegeneration, and neuroinflammation (Zhu and Xu, 2025).

Aspects of RCDs specific to neurons

One of the most characteristic features of RCDs in neurons is that the cell compartmentalization into different domains, i.e., the axon and its terminals, the dendrites, and the perikaryon, may lead to process or synaptic degeneration without an immediate soma death (Conforti et al., 2007). Hybrid and compartmentalized death phenotypes in neurons have been primarily studied in neurodegenerative conditions such as ALS, ALS/motor neuron disease (MND), AD, and PD. In these conditions, the first known site of cellular degeneration is most often at synapses and/or the distal axon (Schmalbruch et al., 1991; Cifuentes-Diaz et al., 2002; Pun et al., 2006), but in AD, it is at presynaptic endings and dendritic spines (Jacobsen et al., 2006). Axonal and synaptic degeneration frequently precedes soma loss. Neurite degeneration may involve ferroptotic (Zhang Y. et al., 2024), necroptotic (Arrázola et al., 2019), or other metabolic mechanisms (Zuo et al., 2023) independently of nuclear execution programs. Epigenetic changes in the neuronal soma may therefore prime vulnerability, while local metabolic insults determine the site and modality of degeneration. The presence of hybrid phenotypes highlights the inadequacies of strict pathway classifications. It endorses a systems-level viewpoint wherein epigenetic states dictate susceptibility, and localized stressors affect the prevailing death program.

Another characteristic of the non-canonical RCDs in neurons is that their execution may be postponed, allowing these cells to survive for extended periods after the initiation of cell death (Cahill et al., 2017; Barrio-Alonso et al., 2018).

The presence of energy constraints can divert apoptosis, an energy-consuming process (Zamaraeva et al., 2005), toward necrotic or hybrid forms, as in the case of ATP depletion (Kushnareva and Newmeyer, 2010). Finally, in neurons, it may be possible to observe mixed RCD phenotypes with partial caspase activation without full apoptotic morphology, as is the case in necroptosis and/or MPT–regulated necrosis (Tait and Green, 2010). Also, as described above, PANoptosis is implicated in several forms of CNS injury or disease, highlighting the coexistence of multiple RCD pathways in neurons and/or glia.

Discussion: an integrative epigenetic perspective on neuronal death

Epigenetic aging is a common cause of increased vulnerability to several neurodegenerative diseases

The epigenetic changes that occur with age lay the molecular groundwork for neurons’ increased susceptibility to RCD pathways. Changes that occur with age alter the transcriptional and chromatin contexts in neurons, lowering the threshold for cell death rather than triggering a specific execution program. Aging neurons exhibit global DNA hypomethylation, focal hypermethylation at regulatory elements, altered histone modification patterns characterized by decreased H3K27me3 and H3K9me3 (Sen et al., 2016), and H4K20me3 and H4K12ac (Sbriz et al., 2025) loss of chromatin fidelity, increased transcriptional noise (Pal and Tyler, 2016), and stochastic gene expression changes (Pal and Tyler, 2016; Bartz et al., 2023), are all characteristics of aging neurons (López-Otín et al., 2013; Bartz et al., 2023). These modifications lead to dysregulated or random gene expression of genes involved in stress response, inflammation, and apoptosis, increasing transcriptional noise and making it harder for neurons to keep their identity. As a result, neurons can move to different RCD pathways. With age, more and more chromatin regions linked to innate immune signaling, such as NF-κB, interferon-stimulated genes, and inflammasome components, become more readily accessible. This epigenetic priming raises the baseline inflammatory level (Franceschi et al., 2007). It sensitizes neurons and glial cells to death pathways with inflammatory outputs, such as necroptosis, pyroptosis, and PANoptosis (Rajesh and Kanneganti, 2022). Priming does not determine pathway selection but supports convergence after stress thresholds are surpassed. Epigenetic aging alters the expression of nuclear-encoded mitochondrial genes essential for oxidative phosphorylation, mitophagy, and redox homeostasis. Lower levels of mitochondrial quality-control genes and higher levels of pro-oxidant pathways make it easier for mitochondria to stop working, build up ROS, and disrupt calcium homeostasis (Sun et al., 2016; Picca et al., 2017; Picca et al., 2019). Mitochondrial failure acts as a common upstream signal for multiple pathways, rendering epigenetic mitochondrial priming a vital link between aging and various RCD processes (Tait and Green, 2010).

The age-associated epigenetic suppression of adaptive stress-response pathways, such as NRF2, FOXO, and autophagy-related genes, reduces neuronal resilience to metabolic, oxidative, and proteotoxic stress (Pajares et al., 2017). Consequently, energy-dependent apoptosis becomes ineffective or partial, leading to a transition towards necrotic, hybrid, or lytic forms of cell death, including necroptosis and PANoptosis (Kushnareva and Newmeyer, 2010). Consequently, epigenetic aging influences execution results without definitively determining them. In addition, epigenetic aging in neurons creates compartment-specific vulnerabilities by differentially regulating gene expression in the soma, axon, terminals, synapses, and dendrites (Conforti et al., 2007; Conforti et al., 2014). Consequently, localized degeneration (e.g., synaptic or axonal loss) may occur independently of imminent somatic death, resulting in hybrid or delayed death phenotypes. This compartmentalization highlights that aged neurons frequently exhibit overlapping molecular characteristics of many RCD pathways instead of traditional textbook morphologies.

Interconnection and stratification of RCD circuits in neurons

In neurons, several RCD pathways seldom function as independent linear cascades. Instead, they create a highly interlinked network wherein apoptosis and non-canonical RCD share upstream stress sensors and downstream execution modules (Galluzzi et al., 2018). Essential molecules, including p53, RIPK1, CASP8, PARP1, BAX/BAK, and mitochondrial bioenergetic state, function as integrators rather than just route switches. Consequently, neurons frequently live in a pre-committed, metastable state, in which many apoptotic pathways are partially activated but not fully engaged. The seemingly “dominant” pathway at the ultrastructural or biochemical level often indicates a late execution bias rather than the initial damage (Tang et al., 2019). Pathway dominance arises from the interplay of four primary determinants: energy availability and mitochondrial efficacy; integrity of checkpoint regulators; chromatin accessibility and transcriptional latency; age-related heterochromatinization; and subcellular compartmentalization.

Energy availability is a significant constraint, as ATP-driven neurons are prone to apoptosis, because caspase activation, apoptosome formation, and chromatin condensation are energy-dependent processes. The bioenergetic collapse, typically caused by mitochondrial dysfunction or oxidative stress, leads to a prevalence of non-canonical, ATP-independent cell death pathways that bypass ATP-dependent checkpoints (Green and Kroemer, 2004). Additionally, age-related decline in mitochondrial epigenetic regulation (including mtDNA methylation and nuclear–mitochondrial transcriptional uncoupling) diminishes the apoptotic threshold (Coppedè and Stoccoro, 2019). The efficacy of checkpoint regulators predominantly depends on the availability of CASP8, which functions as a pivotal hierarchical node: active CASP8 inhibits necroptosis by cleaving RIPK1/RIPK3, whereas its epigenetic repression by promoter methylation or HDAC recruitment promotes necroptosis (Oberst et al., 2011; Fritsch et al., 2019; Newton et al., 2019). Likewise, epigenetic silencing of GPX4 or SLC7A11 promotes ferroptosis rather than apoptosis during oxidative stress (Zhang H. et al., 2025). Chromatin accessibility and transcriptional delay are significant, given that neurons are transcriptionally limited cells. Pathways necessitating fast de novo transcription, such as inflammatory pyroptosis or necroptosis, are impeded when chromatin condensates due to age-related heterochromatinization, depletion of H3K27ac and H4K16ac, or the redistribution of repressive marks H3K9me3 and H3K27me3 (Maze et al., 2015). Consequently, in this situation, death processes that depend on post-translational mechanisms prevail (Benito and Barco, 2015). As previously mentioned, in neurons, RCD can be localized inside distinct cellular domains (Terenzio et al., 2017). Different neuronal compartments display pathway-specific sensitivity (Donato et al., 2019). For example, synapses and axons mainly trigger local mitochondrial and ferroptotic pathways (Qin et al., 2021), whereas the cell body mainly triggers nuclear-mediated apoptosis or parthanatos (Andrabi et al., 2008; Wang et al., 2016). Significantly, epigenetic dysregulation of compartment-specific gene expression intensifies this spatial hierarchy.

Epigenetic modulation of RCD determination

Instead of functioning as binary on/off switches, epigenetic modifications create probabilistic tendencies toward particular RCD modalities by establishing expression thresholds for executioner proteins, determining responsiveness to stress sensors, and controlling the timing of transcription-dependent amplification loops (López-Otín et al., 2013). DNA methylation, histone changes, and chromatin remodelers jointly dictate the most kinetically available route during stress. Moreover, in the aging brain and in neurodegenerative disorders, epigenetic drift results in diminished pathway redundancy, a decreased ability to reverse sublethal stress, and an early commitment to non-apoptotic, inflammatory forms of cell death (Sen et al., 2016). This erosion converts pathway overlap from a protective attribute into a detriment.

Neuroepigenetic vulnerability

Neuroepigenetic fragility is a state in which epigenetic changes caused by aging, disease, or stress make it more complicated for a neuron to handle conflicting death signals, maintain cell death pathways, and choose the least harmful execution mode (Nativio et al., 2018; Nativio et al., 2020). In this condition, neurons resort to energy-conserving mechanisms, inflammation, and irreversible cell death, even in response to modest stimuli. Neuroepigenetic fragility is characterized by a reduced apoptotic threshold, incomplete apoptosis, disinhibition of necroptosis, ferroptosis, and PANoptosis-like convergence, heightened cell-to-cell heterogeneity in death phenotypes, and a robust association between cell death and neuroinflammation.

Clinical significance

Neuroprotection strategies have focused on blocking specific RCD pathways, but this approach has often been limited in effectiveness due to pathway redundancy and compensatory mechanisms. Addressing neuroepigenetic fragility could provide a superior strategy through epigenetic restabilization via HDAC inhibitors (Gräff and Tsai, 2013), bromodomain and extra-terminal (BET) modulators (Martella et al., 2026), and DNMT modulation (Asada et al., 2020) to reinstate hierarchical control; interventions on the mitochondria–chromatin axis to reactivate energy-dependent apoptosis; targeted re-establishment of apoptotic capability to prevent inflammatory cell death; and precise epigenetic editing of key regulators, such as CASP8, GPX4, and PARP1 (Zhao et al., 2026). This strategy seeks to modify the processes of neuronal death, so reducing collateral damage and chronic inflammation, rather than outright stopping death itself.

Conclusion

The evidence together indicates that epigenetic aging functions as a permissive, systems-level catalyst for neuronal vulnerability, rather than as a deterministic initiator of any specific RCD mechanism. Age-related changes in DNA methylation, histone modifications, chromatin structure, and transcriptional accuracy gradually diminish the regulatory ability that typically preserves hierarchical dominance over competing cell death pathways. Consequently, neurons enter a metastable state marked by reduced death thresholds, compromised stress adaptation, mitochondrial dysfunction, and inflammatory priming, which promotes convergence towards non-apoptotic and lytic forms of cell death when stress exceeds tolerance limits. In this context, pathway overlap indicates a lack of epigenetic regulation rather than redundancy, leading to hybrid, compartment-specific, and inflammatory death phenotypes typical of neurodegenerative diseases. The concept of neuroepigenetic fragility embodies this change. It provides a conceptual and therapeutic focal point: targeting the epigenetic framework that controls death pathway selection may be more effective than inhibiting individual execution mechanisms. Restoring epigenetic stability and hierarchical control has become a realistic approach to govern neuronal death, reduce neuroinflammation, and ultimately slow the progression of neurodegenerative disorders.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fceld.2026.1793102/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Jianke Zhang, Thomas Jefferson University, United States

Reviewed by: Peng He, Sanford Burnham Prebys Medical Discovery Institute, United States

Leila Abkhooie, Lorestan University of Medical Sciences, Iran