Long non-coding RNAs and accelerated aging in bipolar disorder

Abstract

Bipolar disorder (BD) has been associated with accelerated biological aging, potentially driven by genetic and environmental factors. Long non-coding RNAs (lncRNAs), key regulators of epigenetic, telomere attrition, and cellular aging processes, may play a role in accelerated aging in BD. This review summarizes current evidence on aging-associated lncRNAs and their relevance to BD. We searched PubMed for English-language original studies published up to June 2, 2025, using keywords related to lncRNAs, aging-related mechanisms, and aging-associated neuropsychiatric disorders. A total of 112 articles reported 163 lncRNAs, of which 19 were common to aging-related mechanisms and aging-associated neuropsychiatric disorders. Among these, ANRIL, HOTAIR, TUG1, MALAT1, NEAT1, and GAS5 have been reported in both aging-related contexts and BD; however, their relevance to BD requires further confirmation in independent and well-characterized cohorts. The remaining overlapping lncRNAs may represent additional candidates of interest for future investigation rather than established contributors to BD pathophysiology.

Abbreviation list

AMPK

AMP-activated protein kinase

BDNF

Brain-Derived Neurotrophic Factor

CDKN2A/B

Cyclin-Dependent Kinase Inhibitor 2A/B

EZH2

Enhancer of Zeste Homolog 2

GR

Glucocorticoid Receptor

HOX

Homeobox

IGF2

Insulin-like Growth Factor 2

MAPK

Mitogen-Activated Protein Kinase

mTOR

Mechanistic Target of Rapamycin

NF-κB

Nuclear Factor kappa-light-chain-enhancer of activated B cells

OXPHOS

Oxidative Phosphorylation

PBMCs

Peripheral Blood Mononuclear Cells

PGC-1α

Peroxisome proliferator-activated receptor-gamma coactivator 1-alpha

PRC

Polycomb Repressive Complex

PRC1/2

Polycomb Repressive Complex 1/2

PUMILIO proteins

RNA-binding proteins involved in post-transcriptional regulation

SIRT1

Sirtuin 1

SNP

Single Nucleotide Polymorphism

SRP

Signal Recognition Particle

1. Introduction

Bipolar disorder (BD) arises from complex genetic, neurobiological, psychological, and environmental interactions and is associated with accelerated aging, increased physical burden, multimorbidity, and mortality (Goes, 2023; McIntyre et al., 2020). A recent review proposed twelve hallmarks of accelerated aging: genomic instability, telomere attrition, epigenetic alterations, loss of proteostasis, deregulated nutrient sensing, mitochondrial dysfunction, cellular senescence, stem cell exhaustion, and altered intercellular communication, disabled macroautophagy, chronic inflammation, and dysbiosis (López-Otín et al., 2023). Biological indicators of accelerated aging in BD have been demonstrated in telomere, epigenetic, chronic inflammation and mitochondrial studies (Spano et al., 2024; Lima et al., 2022; Lam et al., 2023; Miwa et al., 2022; Fries et al., 2020, 2025a, 2025b; Salarda et al., 2021; Yang et al., 2024a; Tessema et al., 2024). A possible genetic vulnerability to aging, together with predisposing and perpetuating factors such as unhealthy lifestyle habits, medication side effects, and inadequate medical care, may contribute to accelerated aging in BD through overlapping molecular mechanisms, including mitochondrial dysfunction and chronic inflammation (Chan et al., 2022). On the other hand, evidence suggests that long-term lithium treatment may counteract telomere attrition, as patients with BD receiving lithium show longer leukocyte telomere length (LTL) than never-treated patients and LTL comparable to that of healthy controls, supporting a potential telomere-protective effect of lithium (Salarda et al., 2021; Pisanu et al., 2020).

During aging, epigenetic and post-transcriptional changes occur, including altered expression of long non-coding RNAs (lncRNAs) (Policarpo et al., 2021). LncRNAs are RNA molecules longer than 200 nucleotides that do not code for proteins. They regulate gene expression and affect many different mechanisms involved in the aging process (Nappi, 2024; He et al., 2018; Li et al., 2023a). Certain lncRNAs are directly involved in cellular aging and age-related disorders and may link BD pathology with premature aging (Nappi, 2024; Ni et al., 2022; Mattick and Makunin, 2006). Potential roles of lncRNAs in aging hallmarks relevant to BD are presented in Fig. 1.

Fig. 1.

Potential Roles of lncRNAs in Aging Hallmarks Relevant to Bipolar Disorder (BD)..

(Figure created with BioRender.)

In this narrative review, we examined lncRNAs identified through a screening of studies related to aging processes and aging-associated neuropsychiatric disorders, including BD. Our aim was to explore the role of these lncRNAs in the aging-related mechanisms observed in BD, describe the direction of their alterations, and contextualize these findings within a broader biological framework.

2. Method

A literature search was conducted using PubMed, incorporating a combination of keywords related to long non-coding RNAs (“long non-coding RNA∗” OR “lncRNA∗” OR “aging-associated lncRNA” OR “long noncoding RNA∗” OR “lnc RNA”) and biological or pathological processes linked to aging-related (“genomic instability” OR “deregulated nutrient-sensing” OR “loss of proteostasis” OR “mitochondrial dysfunction” OR “stem cell exhaustion” OR “altered intercellular communication” OR “chronic inflammation” OR “dysbiosis” OR “disabled macroautophagy” OR “cellular senescence” OR “epigenetic alterations” OR “telomere attrition” OR “premature aging” OR “accelerated aging” OR “biological aging” OR “cellular aging” OR “inflammageing”) as well as aging-associated neuropsychiatric disorders (“severe mental∗” OR “bipolar dis∗” OR “schizo∗” OR “Alzheimer∗” OR “Parkinson∗”).” Articles published up to June 2, 2025, were considered. This work represents a structured narrative review rather than a formal systematic review, as study selection and synthesis were qualitative and hypothesis-generating in nature.

2.1. Eligibility criteria

2.1.1. Inclusion criteria

Studies were eligible if they:

• 1.Reported associations between specific lncRNAs and aging mechanisms, accelerated aging phenotypes, or age-related/severe neuropsychiatric disorder
• 2.Included human or animal subjects
• 3.Provided experimental evidence from in vitro and/or in vivo models
• 4.Demonstrated functional relevance via lncRNA knockdown or overexpression OR strong mechanistic evidence from aging or neuropsychiatric models.
• 5.Included validation analyses in patient-derived tissues

Were supported by evidence from at least one independent study at the functional and/or clinical level.

2.1.2. Exclusion criteria

Studies were excluded if they:

• 1.Were not primarily focused on lncRNAs
• 2.Addressed aging or neuropsychiatric disorders without molecular or functional lncRNA data
• 3.Relied exclusively on in silico analyses without experimental validation
• 4.Were reviews, editorials, conference abstracts, or case reports
• 5.Were not published in English.

2.1.3. Study selection process

The database search identified 1526 records. After title and abstract screening, 1342 records were excluded, and 184 articles were assessed in full text. Following full-text evaluation, 72 articles were excluded for not meeting the predefined eligibility criteria. Ultimately, 112 original research articles were included in the qualitative synthesis and data extraction. The study selection process is illustrated in Fig. 2. Overlapping lncRNAs between the two groups were identified and narratively reviewed in two parts: first, their relevance to BD, and second, their relevance to other neuropsychiatric disorders. Titles, abstracts, and full texts were screened by a single reviewer (SE), consistent with the narrative and hypothesis-generating nature of the review.

Fig. 2.

PRISMA 2020 flow diagram of literature search and study selection.

2.1.4. Data extraction

From each eligible study, the following data were extracted:

• 1.lncRNA identity and biological function
• 2.Experimental model (human tissue, animal model, in vitro/in vivo)
• 3.Aging-related mechanism(s) investigated
• 4.Method of lncRNA modulation (knockdown/overexpression)
• 5.Validation in patient-derived samples
• 6.Associated neuropsychiatric phenotype, where applicable

Data extraction was performed manually and cross-checked to ensure consistency and accuracy.

3. Results

Among 163 lncRNAs identifed, 83 were associated with aging-related mechanisms and 97 with aging-associated neuropsychiatric disorders, with 19 overlapping between the two groups (Supplement 1). Key lncRNAs identified included ANRIL, GAS5, H19, HOTAIR, MALAT1, MEG3, NEAT1, TUG1, BDNF-AS, lincRNA-p21, PANDA, XIST, Lethe, NORAD, KCNQ1OT1, 7SL, AIRN, TERC, and TERRA. A comprehensive overview of lncRNAs associated with bipolar disorder is presented in Table 1. The following section first discusses lncRNAs that may be implicated in accelerated aging in BD, followed by those that may be involved through shared molecular mechanisms.

Table 1.

Comprehensive Overview of lncRNAs in Bipolar Disorder.

LncRNA	Number of Supporting Studies	Direction of Change in BD	Tissue Source	Clinical State	Findings Summary (BD and Aging Studies)	References
ANRIL	Multiple (BD: 2, Aging: Multiple)	Downregulated in BD	Peripheral blood/Various tissues (aging studies)	Not specified	Multiple studies in both BD and aging contexts. BD:downregulation; SNP rs1333048 associated with BD risk. Aging: Implicated in cellular senescence, epigenetic alterations, inflammation, mitochondrial dysfunction	(Zhang et al., 2024; Teshnizi et al., 2022a; Namvar et al., 2020)
HOTAIR	Multiple (BD: 2, Aging: Multiple)	No significant difference/SNP associations in BD	Serum/Various tissues (aging studies)	Drug-naive/drug-free and post-treatment	Multiple studies in BD and aging. BD: no difference; SNP associations with BD risk. Aging: Involved in epigenetic regulation, cellular senescence, protein ubiquitination	(Sargazi et al., 2022; Bella et al., 2023; Trotman et al., 2021)
NEAT1	Multiple (BD: 1, Aging: Multiple)	No significant difference in BD	Peripheral blood/Brain tissue (aging studies)	Not specified	Single BD study Multiple aging studies: Increases in aged brain regions, regulates mitochondrial function, involved in stress responses and memory impairment. Studies in Alzheimer's and Parkinson's disease	(Barry et al., 2015; Sayad et al., 2019a; Li et al., 2023b)
TUG1	Multiple (BD: 2, Aging: Multiple)	Upregulated/No difference (conflicting)	Peripheral blood/Various tissues (aging studies)	Not specified	Conflicting BD studies Aging: Decreases with aging in humans and mice; upregulated in age-related cataracts; regulates mitochondrial bioenergetics. Studies in Parkinson's disease	(Sayad et al., 2019a; Zamani et al., 2023; Xu et al., 2022)
MALAT1	Multiple (BD: 3, Aging: Multiple)	Downregulated/Upregulated/Increased post-treatment (conflicting)	Peripheral blood/Serum/Brain tissue (aging studies)	Various states including post-treatment	Multiple conflicting BD studies Aging: Decreases with aging due to ROS; regulates mitochondrial bioenergetics. Studies in Parkinson's and Alzheimer's disease	(Bella et al., 2023; Shirvani Farsani et al., 2020; Eghtedarian et al., 2022)
GAS5	Multiple (BD: 2, Aging: Multiple)	Downregulated/Upregulated post-treatment	Peripheral blood mononuclear cells/Various tissues (aging studies)	Various including post-treatment	Two BD studies Aging: Regulates glucocorticoid receptor, NF-κB, mTOR pathways; involved in mitophagy. Studies in neurodegeneration and Parkinson's disease	Bella et al., 2023; Zamani et al., 2023; Chen et al., 2017a)
TERC	Multiple (Aging studies only)	Not studied in BD	N/A for BD/Neural stem cells (aging)	N/A	Multiple aging studies: Critical for telomere maintenance and elongation; regulates telomerase activity; essential for neural stem cell survival during aging. Deficiency leads to short telomeres and chromosomal instability	(Grammatikakis et al., 2014; Ghanam et al., 2017)
TERRA	Multiple (Aging studies only)	Not studied in BD	N/A for BD/Telomeric regions (aging)	N/A	Multiple aging studies: Regulates telomere length and stability; forms RNA:DNA hybrids; suppresses telomere elongation; essential for neural stem cell viability; may support repair at critically short telomeres	(Grammatikakis et al., 2014; Canale et al., 2023)
PANDA	Multiple (BD: 1, Aging: Multiple)	Upregulated in BD	Peripheral blood	Not specified	Single BD study showing elevation. Multiple aging studies: Regulated by p53 in response to DNA damage; interacts with PRC2; regulates cell cycle and stress responses; modulates protein trafficking	(Sayad et al., 2019a; Grammatikakis et al., 2014; Peng et al., 2017; Ghanam et al., 2017; Talepoor and Doroudchi, 2024)
MEG3	Multiple (BD: 2, Aging: Multiple)	Downregulated/Upregulated (conflicting)	Peripheral blood/Brain tissue (aging studies)	Not specified	Conflicting BD studies. Multiple aging studies: Regulates p53, inhibits telomerase; modulates PI3K/AKT and AMPK pathways. Studies in Alzheimer's disease and epilepsy	(Xu et al., 2022; Grammatikakis et al., 2014; Kim et al., 2015b; Statello et al., 2020; Maloum et al., 2022; Hosseini and Mokhtari, 2024; Jiang et al., 2020, Zhang et al., 2020; Yi et al., 2019; Deng et al., 2020, Hamilton et al., 2019, Li et al., 2023e)
BDNF-AS	Multiple (BD: 1, Aging: Multiple)	SNP association (rs1519480 A allele increases BD risk)	Genetic study/Brain tissue (aging)	N/A	Single BD genetic study. Multiple aging studies: Negatively regulates BDNF expression; interacts with PRC2; involved in neurodevelopment and synaptic plasticity	(Policarpo et al., 2021; Ghafouri-Fard et al., 2021; Shkundin and Halaris, 2023)
XIST	Multiple (BD: 1, Aging: Multiple)	Upregulated in females with BD	Postmortem prefrontal cortex/Various tissues (aging)	Postmortem	Single BD postmortem study in females. Multiple aging studies: Central role in X chromosome inactivation; regulates DNA methylation in aging; involved in aging-induced osteoporosis; sex-specific effects	(Grammatikakis et al., 2014; Li et al., 2022; Ji et al., 2015; Chen et al., 2020, Gayen et al., 2016, Gordon et al., 2014, Zhang et al., 2021b)
lincRNA-p21	Multiple (BD: 1, Aging: Multiple)	Downregulated in BD	Peripheral blood	Not specified	Single BD study Multiple aging studies: Critical transcriptional repressor in p53 pathway; regulates cell cycle arrest, apoptosis, DNA damage response; interacts with hnRNP-K	(Chen et al., 2017a; Maloum et al., 2022)
H19	Multiple (Aging studies only)	Not studied in BD	N/A for BD/Muscle, vascular tissue (aging)	N/A	Multiple aging studies: Regulates DNA methylation; involved in cellular senescence; upregulated in aged muscles; declines with aging; promotes atherosclerosis; regulated by estrogen (sex-specific); interacts with AMPK and IGF2 pathways	(Grammatikakis et al., 2014; Kim et al., 2015b; Li et al., 2021, 2024b; Ren et al., 2018; Pan, 2017; Özeş et al., 2016; Wang et al., 2020; Basak et al., 2018)
lncRNA 7SL	Multiple (Aging studies only)	Not studied in BD	N/A for BD/Various tissues (aging)	N/A	Multiple aging studies: Modulates autophagy; involved in cell cycle regulation and cellular senescence; regulates p53 through interaction with TP53 mRNA affecting translation	(Grammatikakis et al., 2014; Ghanam et al., 2017; Talepoor and Doroudchi, 2024)
NORAD	Multiple (Aging studies only)	Not studied in BD	N/A for BD/Cytoplasmic (aging)	N/A	Multiple aging studies: Acts as reservoir for PUMILIO proteins; critical for genome stability; deletion leads to chromosomal instability and early aging phenotype in mice; forms topoisomerase complex for DNA damage response	(Yao et al., 2019; Ghanam et al., 2017; Statello et al., 2020)
KCNQ1OT1	Multiple (Aging studies only)	Not studied in BD	N/A for BD/Various tissues (aging)	N/A	Multiple aging studies: Involved in imprinted gene silencing via PRC2 interaction; regulates DNA methylation in aging; histone modifications change with aging	(Grammatikakis et al., 2014; Guo et al., 2021)
AIRN	Multiple (Aging studies only)	Not studied in BD	N/A for BD/Various tissues (aging)	N/A	Multiple aging studies: Necessary for paternal Igf2r silencing; regulates epigenetic modifications through PRC2 interaction; involved in DNA methylation regulation during aging	(Grammatikakis et al., 2014; Yao et al., 2019; Guo et al., 2021)
LncRNA Lethe	Multiple (Aging studies only)	Not studied in BD	N/A for BD/Immune cells (aging)	N/A	Multiple aging studies: Regulates immune response; binds to NF-κB subunit RelA; inhibits inflammatory protein production; involved in age-related inflammation	(Talepoor and Doroudchi, 2024; Kim et al., 2015b)

PRC1/2: Polycomb Repressive Complex 1/2 NF-κB: Nuclear Factor kappa B CDKN2A/B: Cyclin-Dependent Kinase Inhibitor 2A/B BDNF: Brain-Derived Neurotrophic Factor EZH2: Enhancer of Zeste Homolog 2 HOX: Homeobox H3K27me3 Histone H3 Lysine 27 trimethylation LRKK2: Leucine-rich repeat kinase 2 PGC-1alfa: Peroxisome proliferator-activated receptor-gamma coactivator 1-alpha AMPK/SIRT1: AMPK/Sirtuin 1 OXPHOS: Oxidative Phosphorylation GR: Glucocorticoid Receptor mTOR: Mechanistic Target of Rapamycin PBMCs: Peripheral Blood mononuclear Cells SNP: Single Nucleotide Polymorphism.

3.1. Associations between aging-related LncRNAs and BD

Key lncRNAs, including ANRIL, HOTAIR, NEAT1, TUG1, MALAT1, and GAS5, may link aging-related processes with bipolar disorder through their involvement in fundamental hallmarks of aging. Their putative roles in accelerated aging and the expression changes reported in bipolar disorder are summarized in Table 2.

Table 2.

Accelerated aging in bipolar disorder and the roles of LncRNAs.

LncRNA	Hallmarks of Aging	Evidence in BD	Findings in BD	Key References
ANRIL	Epigenetic alterations, inflammation	Multiple human studies	Downregulated in blood; rs1333048 SNP linked to higher BD I risk; NF-κB upregulated in BD brains	(Zhang et al., 2024; Teshnizi et al., 2022a; Namvar et al., 2020)
HOTAIR	Epigenetic alterations, senescence	Genetic association studies	No serum expression change; SNPs (rs920778, rs1899663) linked to BD II risk	(Sargazi et al., 2022; Bella et al., 2023; Trotman et al., 2021)
NEAT1	Inflammation, mitochondrial dysfunction	Indirect/preclinical	No blood expression change; animal data: stress adaptation and synaptic effects	(Barry et al., 2015; Sayad et al., 2019a; Li et al., 2023b)
TUG1	Mitochondrial dysfunction, nutrient sensing	Conflicting human studies	Elevated in BD or no difference	(Sayad et al., 2019a; Zamani et al., 2023; Xu et al., 2022)
MALAT1	Mitochondrial dysfunction, senescence	Multiple human studies	Downregulated in BD or upregulated in BD or upregulated in BD, post-treatment	(Bella et al., 2023; Shirvani Farsani et al., 2020; Eghtedarian et al., 2022)
GAS5	Inflammation, mitochondrial dysfunction, nutrient sensing	Treatment-dependent	Downregulated in BD PBMCs or upregulated in BD, post-treatment	(Bella et al., 2023; Zamani et al., 2023; Chen et al., 2017a)

NF-κB: Nuclear Factor kappa B PBMCs: Peripheral Blood mononuclear Cells SNP: Single Nucleotide Polymorphism.

ANRIL (Antisense Non-coding RNA in the INK4 Locus): ANRIL (CDKN2B-AS1) is a lncRNA located at the 9p21 chromosomal region and is thought to influence aging through multiple hallmarks, including cellular senescence, epigenetic alterations, deregulated nutrient sensing, chronic inflammation, impaired macroautophagy, and mitochondrial dysfunction. It limits the progression of senescence by repressing the expression of the p15 gene, which functions as a key inducer of cellular senescence through inhibition of CDK4/6-mediated cell cycle progression (Zhang et al., 2024). Additionally, ANRIL epigenetically modulates the expression of distal genes by recruiting the PRC1 and PRC2 epigenetic regulatory complexes to target loci through Alu sequences (Grammatikakis et al., 2014). Moreover, by regulating miR-146a expression via the SASP, ANRIL influences endothelial senescence and angiogenesis in the context of age-related vascular dysfunction (Lin et al., 2025).

It mediates the transcriptional repression of BDNF by recruiting EZH2 to the BDNF promoter region, thereby regulating the expression of proteins involved in endothelial function and mitochondrial dynamics (Su et al., 2022). Furthermore, by modulating the expression of neighboring genes such as CDKN2A and CDKN2B, ANRIL may influence key cellular processes, including proliferation, apoptosis, and inflammation (Ding et al., 2025). Through modulation of the NF-κB pathway, ANRIL plays a critical role in immune homeostasis and has been implicated in various chronic inflammatory conditions (Zhou et al., 2015). In addition, ANRIL has also been shown to exert regulatory effects on deregulated nutrient sensing (Li et al., 2023c), impaired macroautophagy (Yang et al., 2024b; Zeng et al., 2019), and mitochondrial dysfunction (Bure et al., 2022). In Alzheimer's disease, ANRIL silencing promotes neurite outgrowth through microRNA-125 signaling (Zhou et al., 2020).

Studies of postmortem brain tissue from individuals with BD have shown NF-κB upregulation, suggesting a potential association with disease progression (Rao et al., 2010; Sun et al., 2001). In a recent study by Teshnizi et al. (2022b), the expression of ANRIL, a NF-κB-related lncRNA, was found to be significantly downregulated in peripheral blood samples from individuals with BD compared to HCs. Moreover, another study reported that single nucleotide polymorphisms (SNPs) within the ANRIL gene, particularly rs1333048, are associated with an increased risk of BD type I. Individuals carrying the A/A genotype at this locus had a significantly higher risk of developing BD than those with the C/C genotype (Namvar et al., 2020).

HOTAIR (HOX antisense intergenic RNA): HOTAIR is a well-characterized lncRNA that regulates the expression of the HOX gene family through its interaction with the PRC1 and PRC2 complexes (Trotman et al., 2021). It plays a central role in chromatin remodeling and transcriptional repression, primarily via its association with polycomb repressive complexes (Hajjari and Salavaty, 2015). HOTAIR has been shown to directly bind to LRKK2 mRNA, forming an RNA duplex that stabilizes its expression and subsequently induces neuronal apoptosis (Policarpo et al., 2021). Furthermore, it has been implicated in the facilitation of protein ubiquitination and degradation processes, thereby potentially triggering cellular senescence (Grammatikakis et al., 2014), and it is known to contribute to cancer cell proliferation, apoptosis, and aging (Kim et al., 2015a).

Given its prominent role in epigenetic regulation, HOTAIR has been hypothesized to play a part in the molecular pathophysiology of neuropsychiatric disorders, including BD. In a pilot study by Bella et al. (2023), the expression of HOTAIR was analyzed in serum samples from drug-naive or drug-free individuals with BD, both at the time of diagnosis and after five months of pharmacological treatment. Despite its established regulatory functions, the study found no statistically significant differences in HOTAIR expression between individuals with BD and HCs at any time point. In a case-control study involving 194 BD patients and 163 healthy controls, four single nucleotide polymorphisms (rs1899663, rs12826786, rs4759314, and rs920778) within the HOTAIR locus were significantly associated with BD risk under multiple genetic models. Notably, the CT genotype of rs920778, GT genotype of rs1899663, and CT genotype of rs12826786 were found to increase the risk of BD type II, whereas the GG genotype of rs4759314 was associated with an 83% reduction in BD type II risk. Haplotype analysis further revealed that CTTA and CTCG combinations across these loci were positively correlated with BD susceptibility (Sargazi et al., 2022). Findings from other studies suggest that HOTAIR may still serve as a prognostic indicator in BD (Roy et al., 2024).

NEAT1 (Men ε = nuclear paraspeckle assembly transcript 1-nuclear enriched abundant transcript 1): It is a lncRNA essential for the formation of paraspeckles - nuclear bodies involved in gene regulation under cellular stress conditions. NEAT1 plays a critical role in cellular energy metabolism by regulating mitochondrial gene expression. During aging, its expression has been shown to increase in specific brain regions, particularly in the subependymal zone (Barry et al., 2015). Notably, in aged bone marrow mesenchymal stem cells (BMSCs), downregulation of NEAT1 significantly alleviates mitochondrial dysfunction, highlighting its regulatory influence on mitochondrial homeostasis (Zhang et al., 2022). In addition to its role in mitochondrial function, NEAT1 is involved in the regulation of immune responses, modulating signaling pathways such as NF-κB and controlling IL8 gene expression following immune stimulation (Kim et al., 2015a). Furthermore, NEAT1 has been implicated in the p53-mediated tumor suppressor pathway, suggesting its involvement in cellular stress responses, cancer-related processes, and the regulation of immune response genes such as IL-8 (Imamura et al., 2014; Yao et al., 2019). Reactive oxygen species induced defects in TFAM signaling lead to mitochondrial damage, alter NEAT1 expression, and impair paraspeckle formation (Zhang et al., 2021a).

Animal studies have demonstrated that NEAT1 deficiency leads to altered stress responses, social interaction deficits, and changes in neuronal excitability, highlighting its role in stress reactivity, mood regulation, and stress adaptation (Barry et al., 2017; Kukharsky et al., 2020). NEAT1 is also associated with neuronal histone methylation and age-related memory impairment (Li et al., 2023b; Zhao et al., 2020; Wang et al., 2019; Butler et al., 2019; Katsel et al., 2019). It regulates Aβ clearance in neuroglial cells, affects microtubule stabilization, and decreased levels in schizophrenia may modulate oligodendrocyte transcription. However, in aged mice, NEAT1 induces epigenetic alterations in the hippocampus that impair long-term memory formation (Butler et al., 2019). In Parkinson's disease, NEAT1 has been shown to reduce inflammation (Sivagurunathan et al., 2023). In Alzheimer disease mouse models, NEAT1 improved cognition, decreased inflammatory cytokine levels, and suppressed oxidative stress (Li et al., 2023d).

Clinical studies have also explored NEAT1 expression in individuals with BD. For instance, a study by Sayad et al. (2019) assessed the expression levels of NEAT1 in the peripheral blood of individuals with BD and healthy controls. NEAT1 did not show significant differential expression in this study (Sayad et al., 2019a).

TUG1 (taurine upregulated gene 1): TUG1 was initially identified as a lncRNA upregulated in retinal cells following taurine treatment. Since then, it has been implicated in a variety of biological processes, including neuronal development, apoptosis regulation, and cancer-related pathways (Fernandes et al., 2018; Zhou et al., 2019). At the cellular level, TUG1 interacts with PRC1 and PRC2 complexes, contributing to mechanisms involved in cellular differentiation and survival (Policarpo et al., 2021). Furthermore, it has been shown to enhance SIRT1 expression and support AMPK signaling by targeting miR-200a-3p (Li et al., 2024a). It activates the AMPK/SIRT1 pathway, enhancing neuronal mitophagy and reducing oxidative stress (Yao et al., 2025; xing Xue et al., 2022). By interacting with PRC1/PRC2, TUG1 regulates cell survival and neurogenesis (xing Xue et al., 2022). It regulates mitochondrial bioenergetics (Jianyin et al., 2016) and epigenetically controls PGC-1 expression (Ageeli Hakami, 2024). Reduced TUG1 expression leads to decreased PGC-1 levels (Jianyin et al., 2016). By interacting with PRC1/PRC2, TUG1 regulates cell survival and neurogenesis (xing Xue et al., 2022). In both humans and mice, This pattern suggests that TUG1 may decrease during physiological aging but become upregulated in pathological or stress-related aging states (Gimbel et al., 2022; Li et al., 2017). In renal aging, TUG1 directly targets PGC-1α, thereby regulating mitochondrial homeostasis (Zhu et al., 2024). TUG1 levels increase in neurodegenerative processes (e.g., Parkinson's disease) and can exacerbate neuroinflammation (Zhai et al., 2020). In Parkinson's disease cell and mouse models, TUG1 expression is upregulated, whereas its silencing protects against in vitro apoptosis, oxidative stress, and neuroinflammation, as well as in vivo pathological damage and neuroinflammation (Zhai et al., 2020).

Indeed, a study by Sayad et al. (2019) reported significantly elevated TUG1 expression levels in individuals with BD compared to healthy controls, suggesting a potential role for this lncRNA in the molecular mechanisms underlying BD (Sayad et al., 2019a). Elevated expression of TUG1 in BD has been associated with apoptosis imbalance (Roy et al., 2024; Sayad et al., 2019b). However, these findings were not replicated in a subsequent study by Zamani et al. (2023), which found no significant difference in TUG1 expression between individuals with BD and healthy individuals (Zamani et al., 2023).

MALAT1 (metastasis-associated lung adenocarcinoma transcript 1, NEAT2): It is a highly conserved and abundantly expressed lncRNA, primarily localized in nuclear speckles, where it may serve as a scaffold for protein-protein, protein-RNA, and protein-DNA interactions, potentially facilitating and enhancing these molecular interactions. It plays a crucial role in regulating gene expression at both transcriptional and post-transcriptional levels, influencing processes such as alternative splicing, cell cycle progression, synaptogenesis, and cellular senescence (Grammatikakis et al., 2014; Arun et al., 2020). During aging, MALAT1 levels decrease as a secondary response to elevated ROS (Ruan et al., 2022). It also regulates mitochondrial bioenergetics by maintaining the glycolysis/OXPHOS balance, making it responsive to oxidative stress and disruptions in proteostasis.

In Parkinson Disease mouse models, reduced MALAT1 expression exerts neuroprotective effects (Zhang et al., 2016). MALAT1 contributes to Parkinson's disease pathology by modulating multiple mechanisms, including α-synuclein proteostasis, neuroinflammation, autophagy, and neuronal apoptosis (Ni et al., 2022). In Alzheimer's disease cell lines, animal models, and human brain tissue, MALAT1 expression is downregulated (Chanda et al., 2022).

In the context of BD, studies have reported conflicting findings regarding MALAT1 expression levels. Shirvani Farsani et al. (2020) observed a significant downregulation of MALAT1 in the peripheral blood of individuals with BD compared to healthy controls, suggesting its potential as a diagnostic biomarker (Shirvani Farsani et al., 2020). Conversely, Reyhane Eghtedarian et al. (2022) reported an upregulation of MALAT1 in individuals with BD, indicating a possible association with vitamin D receptor signaling pathways (Eghtedarian et al., 2022). Furthermore, Fabrizio Bella et al. (2023) found that MALAT1 expression was significantly increased in individuals with BD following pharmacological treatment, highlighting its potential role as a dynamic biomarker responsive to therapeutic interventions (Bella et al., 2023).

GAS5 (Growth Arrest-Specific 5): It is a lncRNA that accumulates in growth-arrested cells and functions as a decoy for the glucocorticoid receptor (GR), thereby inhibiting GR-mediated gene transcription and modulating stress responses. It promotes apoptosis by activating the p53 and caspase pathways and induces cell cycle arrest via p21 (Chen et al., 2017b). GAS5 also regulates both glucocorticoid receptor and NF-κB signaling pathways (Xu et al., 2020). By suppressing mTOR, GAS5 activates mitophagy. Moreover, mitochondria-localized GAS5 modulates the TCA cycle flux under energy stress, thereby regulating cellular metabolism and growth (Sang et al., 2021).

Beyond its role in cell cycle regulation and apoptosis, GAS5 has been implicated in various physiological and pathological processes, including cancer progression and immune responses (Kaur et al., 2022). Downregulation of GAS5 in hippocampal cell lines has been shown to contribute to neurodegeneration (Patel et al., 2023). In rats, GAS5 inhibition suppresses mitochondrial apoptosis, increases Akt and GSK3β phosphorylation, and reduces neuronal injury (Wang et al., 2021). Additionally, maternal separation has been shown to increase GAS5 expression in the hippocampus, which is associated with heightened anxiety and spatial memory deficits (Banerjee et al., 2024). In Parkinson's disease mouse models, GAS5 regulates microglial inflammatory responses, improves neuronal insulin signaling in aged mice, and reduces neuroinflammation (Xu et al., 2020). Low GAS5 expression is linked to impaired neuroendocrine regulation and synaptic integrity (Patel et al., 2023).

Decreased GAS5 expression has also been observed in BD, which may contribute to GR overactivation and chronic inflammation (Zhou and Chen, 2020). However, studies have reported contrasting findings regarding GAS5 expression levels in BD. For instance, Zamani et al. (2023) observed a significant downregulation of GAS5 in the peripheral blood mononuclear cells of individuals with BD compared to healthy controls. Conversely, Bella et al. (2023) reported an upregulation of GAS5 expression in individuals with BD following treatment.

3.2. Other aging-related LncRNAs and their potential implications in BD

For several lncRNAs, inclusion was based on robust mechanistic evidence in aging or neuropsychiatric models rather than direct BD-specific validation. These include PANDA, MEG3, BDNF-AS, XIST, lincRNA-p21, H19, NORAD, KCNQ1OT1, AIRN, Lethe, TERC, and TERRA, summarized in (Table 3). While direct evidence in BD is limited or absent for many of these lncRNAs, their mechanistic involvement in aging and stress-response pathways highlights them as candidates for future investigation rather than established contributors to BD pathophysiology.

Table 3.

Other LncRNAs and their roles in aging.

LncRNA	Aging-Related Effects	Evidence Base	Relevance to BD	Interpretation for BD	References
PANDA	Nutrient sensing, cellular senescence, genomic instability	Preclinical/indirect	No direct BD data	Stress- and p53-related aging pathways; hypothesis-generating for BD	(Sayad et al., 2019a; Grammatikakis et al., 2014; Peng et al., 2017; Ghanam et al., 2017; Talepoor and Doroudchi, 2024)
MEG3	Senescence, telomere attrition, neurodegeneration	Human & experimental	Limited BD evidence	Stress- and apoptosis-related; possible link to depressive phenotypes	(Xu et al., 2022; Grammatikakis et al., 2014; Kim et al., 2015b; Statello et al., 2020; Maloum et al., 2022; Hosseini and Mokhtari, 2024; Jiang et al., 2020; Zhang et al., 2020; Yi et al., 2019)
BDNF-AS	Neurodegeneration, reduced plasticity	Human & experimental	Indirect BD relevance	Regulates BDNF; potential relevance to cognition and mood regulation	(Policarpo et al., 2021; Ghafouri-Fard et al., 2021; Shkundin and Halaris, 2023)
XIST	Epigenetic alterations, sex-specific aging	Human & experimental	Indirect	May contribute to sex differences in BD; no direct expression data	(Grammatikakis et al., 2014; Li et al., 2022; Ji et al., 2015; Chen et al., 2020)
lincRNA-p21	Senescence, genomic instability	Experimental	No direct BD data	p53-mediated stress response; hypothesis-generating	(Chen et al., 2017a; Maloum et al., 2022)
H19	Senescence, nutrient sensing, epigenetic aging	Human & experimental	Indirect	Metabolic and sex-specific aging pathways; possible modifier	(Grammatikakis et al., 2014; Kim et al., 2015b; Li et al., 2021, 2024b; Ren et al., 2018; Pan, 2017; Özeş et al., 2016; Wang et al., 2020; Basak et al., 2018)
lncRNA 7SL	Proteostasis loss, impaired autophagy	Experimental	No direct BD data	Cellular stress regulation; exploratory relevance	(Grammatikakis et al., 2014; Ghanam et al., 2017; Talepoor and Doroudchi, 2024)
NORAD	Genomic instability, proteostasis	Human & animal	Indirect	Aging burden and genome maintenance; plausible BD relevance	(Yao et al., 2019; Ghanam et al., 2017; Statello et al., 2020)
KCNQ1OT1	Epigenetic alterations, neurodegeneration	Experimental	No direct BD data	Imprinting-related epigenetic aging; exploratory	(Grammatikakis et al., 2014; Guo et al., 2021)
AIRN	Epigenetic aging, metabolic dysregulation	Experimental	No direct BD data	Metabolic–epigenetic interface; hypothesis-generating	(Grammatikakis et al., 2014; Yao et al., 2019; Guo et al., 2021)
LncRNA Lethe	Chronic inflammation	Experimental	No direct BD data	NF-κB–linked inflammaging; indirect relevance	(Talepoor and Doroudchi, 2024; Kim et al., 2015b)
TERC	Telomere attrition, stem cell exhaustion	Human & experimental	Limited BD data	Telomere biology linking aging and BD	(Grammatikakis et al., 2014; Ghanam et al., 2017)
TERRA	Telomere instability, inflammation	Experimental	Very limited BD data	Telomere regulation; preliminary relevance	(Grammatikakis et al., 2014; Canale et al., 2023)

BDNF: Brain-Derived Neurotrophic Factor NF-κB: Nuclear Factor kappa-light-chain-enhancer of activated B cells.

4. Discussion

This review consolidates emerging evidence underscoring the central role of lncRNAs in the interface between BD and accelerated biological aging. LncRNAs may function as candidate biomarkers and may also act as mechanistic modulators of cellular processes fundamental to aging, including senescence, inflammation, apoptosis, mitochondrial homeostasis, and telomere dynamics. Collectively, these pathways contribute to the systemic and neurobiological deterioration frequently observed in BD, supporting the hypothesis that accelerated aging is a critical component of its pathophysiology.

Conflicting findings have been reported for TUG1 expression in bipolar disorder, with Sayad et al. (2019) reporting upregulation and Zamani et al. (2023) observing no significant difference, despite both studies analyzing peripheral blood samples. This heterogeneity likely reflects differences in clinical state, treatment exposure, and sample composition, rather than true biological inconsistency. TUG1 expression may be elevated during acute mood episodes but normalize during euthymia, and pharmacological treatment may further modulate its expression. Notably, aging-related studies indicate that TUG1 decreases during physiological aging but increases in pathological aging conditions such as cataracts and Parkinson's disease. Taken together, these findings support the interpretation of TUG1 as a state-dependent biomarker, highlighting the need for longitudinal studies with precise clinical characterization. Three independent studies report divergent expression patterns for MALAT1, underscoring its highly dynamic regulation. Shirvani Farsani et al. (2020) reported downregulation at baseline, potentially reflecting acute oxidative stress or heightened inflammatory burden. In contrast, Eghtedarian et al. (2022) observed upregulation, which was associated with vitamin D receptor signaling, suggesting sensitivity to nutritional and metabolic status. More recently, Bella et al. (2023) demonstrated increased MALAT1 expression following treatment, indicating a treatment-responsive profile. Differences in biospecimen type (whole blood vs serum) may further contribute to variability. Collectively, these findings indicate that MALAT1 is not a stable trait marker but rather a highly context-dependent, state-sensitive lncRNA, reflecting dynamic metabolic and redox states as well as treatment effects. For GAS5, the apparent discrepancy between studies reflects a temporal and treatment-dependent pattern rather than conflicting evidence. Zamani et al. (2023) reported downregulation at baseline in untreated patients, whereas Bella et al. (2023) observed upregulation following pharmacological treatment. Reduced baseline GAS5 expression may indicate glucocorticoid receptor overactivation and impaired stress-response regulation in untreated bipolar disorder. Treatment-associated changes in GAS5 expression are consistent with partial restoration of glucocorticoid receptor–mediated transcriptional control. This interpretation is consistent with aging literature linking GAS5 to mitophagy regulation and inflammatory signaling. Together, these findings support GAS5 as a treatment-response biomarker rather than a static disease indicator.

Overall, these examples illustrate that heterogeneity in lncRNA findings should be interpreted within a biological and clinical context, rather than as irreconcilable contradictions. lncRNAs such as TUG1, MALAT1, and GAS5 appear to function as dynamic regulators, sensitive to illness phase, metabolic state, and treatment exposure, features that align closely with models of accelerated and pathological aging in bipolar disorder.

Despite these inconsistencies, the biological relevance of these lncRNAs remains compelling. NEAT1, for instance, is well established as a regulator of stress response and synaptic homeostasis, suggesting its involvement in BD pathophysiology even though direct human evidence remains limited. MEG3, as a maternally imprinted gene, may represent a transgenerational vulnerability factor, linking maternal stress and mitochondrial regulation to offspring risk for BD. TERRA and TERC play key roles in telomere maintenance. Telomere shortening has been consistently observed in BD (Spano et al., 2024; Barbé-Tuana et al., 2016), and dysregulation of TERRA may contribute to genomic instability and premature cellular aging. Moreover, XIST introduces an additional layer of complexity by mediating sex-specific epigenetic regulation through X-chromosome inactivation (Chen et al., 2020). This process may contribute to gender differences in stress response, mitochondrial function, and immune activity domains increasingly recognized as critical to both aging and BD susceptibility.

Emerging evidence suggests that several lncRNAs implicated in accelerated aging and cellular stress pathways may also represent potential therapeutic targets of lithium in BD. For instance, the pro-apoptotic lncRNA GAS5, which shows reduced expression in BD (Zamani et al., 2023), may represent a compensatory or context-dependent response associated with neuronal stress regulation. This aligns with lithium's known anti-apoptotic action through GSK-3β inhibition (Snitow et al., 2021) indicating a possible convergent mechanism. Similarly, ANRIL, which exerts pro-inflammatory and pro-apoptotic effects via the CDKN2A/B locus and NF-κB signaling (Ding et al., 2025), may be indirectly modulated by lithium through its inhibition of GSK-3β and suppression of the NLRP3 inflammasome (Wang et al., 2024), both of which reduce pro-inflammatory cytokine production. HOTAIR, a lncRNA that promotes senescence and neuroinflammation through epigenetic regulation (Policarpo et al., 2021; Trotman et al., 2021), may also intersect with lithium's anti-inflammatory and epigenetic modulatory effects (Marie-Claire et al., 2021). Given lithium's complex mechanism of action and the restricted activity of telomerase in human tissues, it is likely that additional molecular pathways contribute to lithium-induced telomere protection (Squassina et al., 2017). In this context, lncRNAs may represent promising targets for future research on telomere dynamics and the pathophysiology of BD. lncRNAs involved in telomere maintenance, such as TERRA and TERC (Ghanam et al., 2017; Talepoor and Doroudchi, 2024; Canale et al., 2023) could contribute to lithium's telomere-stabilizing and neuroprotective effects (Cardillo et al., 2018; Themoteo et al., 2022; Martinsson et al., 2013). In addition, the observation that MALAT1 expression normalizes or increases following pharmacological intervention supports its potential use as a dynamic biomarker for monitoring therapeutic efficacy and biological recovery trajectories.

Taken together, these observations suggest that lncRNA alterations in BD are highly context-dependent, shaped by treatment exposure, illness phase, and tissue specificity. Lithium's therapeutic effects may partly involve modulation of lncRNAs regulating apoptosis, inflammation, epigenetic remodeling, and telomere-related pathways; however, direct mechanistic evidence remains limited and warrants experimental validation.

Although most of these findings derive from aging and neurodegenerative models, the convergence of mitochondrial, inflammatory, and stress-response pathways suggests potential relevance to BD-related accelerated aging rather than direct disease specificity.

Beyond their diagnostic and mechanistic relevance, lncRNAs may also represent potential components of molecular aging clocks. In contrast to conventional DNA methylation–based clocks, which largely reflect cumulative molecular damage, an lncRNA-based approach could capture regulatory adaptability, reflecting how gene-expression and stress-response networks adjust to environmental and metabolic pressures over time.

The dysregulation of aging-associated lncRNAs in BD has potential implications for understanding the disorder's high comorbidity burden. BD is associated with premature onset of cardiovascular disease, metabolic syndrome, and cognitive decline, conditions linked to accelerated biological aging. LncRNAs regulating mitochondrial function, inflammation and telomere maintenance may represent molecular links between psychiatric pathology and systemic aging. For instance, ANRIL is implicated in cardiovascular disease, and H19 promotes atherosclerosis, both prevalent in BD. MEG3 regulates metabolic pathways relevant to BD's metabolic comorbidities. However, whether lncRNA dysregulation causally contributes to comorbidity or merely reflects shared pathophysiology remains unresolved.

From a translational perspective, treatment-responsive lncRNAs may serve as dynamic biomarkers for monitoring therapeutic response and biological recovery, complementing clinical symptom scales. However, their utility as diagnostic markers is limited by heterogeneity across studies and the need for standardization of measurement protocols, clinical characterization, and tissue sources. Furthermore, the use of peripheral blood, while practical, may not reflect CNS-specific changes, as evidenced by NEAT1 and XIST findings in brain tissue.

In summary, lncRNAs can represent an emerging molecular bridge between the biology of aging and the neuropsychiatric mechanisms of BD. By elucidating their roles in cellular resilience, mitochondrial function, immune regulation, and epigenetic control, we may advance toward the development of lncRNA-guided biomarkers and therapeutics capable of mitigating both the psychiatric and systemic consequences of accelerated aging in BD.

5. Future directions

Current evidence is limited by small sample sizes, cross-sectional designs, heterogeneous clinical populations, and reliance on peripheral blood. Many lncRNAs have been studied in only one or two BD cohorts, and replication in independent samples with detailed clinical characterization is essential. Conflicting findings for TUG1, MALAT1, and MEG3 highlight the need for longitudinal studies assessing lncRNA expression across illness phases and treatment trajectories. The lack of brain tissue studies for most lncRNAs limits mechanistic interpretation, as peripheral expression may not reflect CNS pathology.

Methodological standardization is critical. Differences in sample type, RNA extraction methods, and expression quantification may contribute to variability. Integration with clinical data, including detailed medication histories, illness duration, number of episodes, and comorbidity profiles is necessary to disentangle state vs. trait effects. Sex-stratified analyses are essential given the sex-specific roles of XIST and MEG3 and the female predominance of BD type II.

Mechanistic studies are needed to establish causality. Functional experiments in cellular and animal models can determine whether lncRNA alterations drive BD-related pathology or represent downstream consequences. Studies in lithium-treated cohorts can test whether lithium modulates specific lncRNAs and whether these changes correlate with therapeutic response or telomere protection. Finally, investigation of understudied lncRNAs (TERC, TERRA, NORAD, Lethe) with strong mechanistic rationale but no BD-specific evidence may reveal novel therapeutic targets.

One particularly promising avenue is the development of an lncRNA-based biological aging clock. Because many lncRNAs respond rapidly and sensitively to cellular stress, inflammation, and metabolic fluctuations, such a clock could reflect acute physiological stressors and subtle regulatory shifts more precisely than traditional transcriptomic or epigenetic clocks. Such tools may offer a more nuanced proxy for biological resilience and vulnerability, with potential implications for early identification of accelerated aging and the development of more individualized treatment strategies.

In conclusion, current evidence supports a role for lncRNAs in the molecular mechanisms linking BD and accelerated aging. The dynamic, context-dependent nature of lncRNA expression underscores the importance of integrating clinical phenotypes, treatment status, and illness phase into future research. By elucidating lncRNA roles in mitochondrial function, inflammation, and telomere maintenance, we may identify molecular links between psychiatric symptoms and systemic comorbidities, ultimately advancing precision medicine approaches for BD.

Contributions

SE led the research, conducted the primary literature review, and drafted the manuscript.

DC supervised the project and provided guidance on study design and overall direction.

HC contributed to manuscript writing, critically reviewed and revised the manuscript.

IF contributed to manuscript writing, critically reviewed.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of the final (and subsequently accepted) version of the manuscript, the authors used GPT-4o (OpenAI) and Gemini to enhance fluency and readability. All content generated using these tools was carefully reviewed and edited by the authors, who take full responsibility for the final manuscript.

Funding

This research was not funded.

Declaration of competing interest

The authors declare that they have no competing financial interests or personal relationships that could have appeared to influence reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article: