MLL, TERRA, and Telomere Chromatin Dynamics in Replicative Senescence and Aging

1. Introduction

The biology of aging has long been debated between two opposing paradigms: one viewing aging as a stochastic accumulation of cellular and molecular damage involving oxidative stress, DNA mutations, telomere shortening, and genomic instability, and another proposing that aging is a programmed, adaptive process encoded in the genome and epigenetically regulated through chromatin-based transcriptional switches [1,2]. Within the programmed framework, aging can be described as a form of phenoptosis, the programmed death of the individual, a process detrimental at the organismal level yet potentially advantageous in terms of supra-individual selection (species or population) [3]. In addition, an alternative view has emerged in which aging is not strictly programmed but arises from regulated biological processes that, while not evolved to cause aging, progressively drive it through their continued or altered activity, consistent with a quasi-programmed model [4,5]. Although this model has been predominantly developed in the context of persistent mTOR/TOR signaling and cellular hyperfunction, the same logic can be extended to telomere-driven replicative senescence as an adaptive/tumor suppressive mechanism whose cumulative effect becomes deleterious over time.

The discovery of replicative senescence, first described by Hayflick and Moorhead in 1961, provided a pivotal foundation for this debate [6]. Their demonstration that normal human fibroblasts undergo a finite number of divisions before entering senescence, the Hayflick limit, implied the existence of an intrinsic cellular “clock” regulating proliferative capacity. A mechanistic explanation was later provided by Olovnikov’s description of the end-replication problem, which established that DNA polymerase cannot fully replicate the 3′ termini of linear chromosomes, resulting in progressive telomere shortening with each cell division [7]. The subsequent identification of telomerase by Blackburn and colleagues revealed a ribonucleoprotein reverse transcriptase capable of elongating telomeric DNA in germline, stem, and cancer cells, thereby counteracting this attrition [8]. Together, these findings established telomere shortening as a molecular timer that limits somatic cell proliferation, and replicative senescence as a tumor-suppressive response that links progressive telomere shortening to enforced limitation of proliferative capacity in response to oncogenic hyperproliferation [9]. Senescent cells stop dividing yet remain metabolically active, persist within renewing tissues, and contribute to progressive functional decline in chronic degenerative diseases, cancer, and organismal aging through DNA-damage signaling, senescence-associated secretory phenotype (SASP) activation, stem cell exhaustion, and tissue attrition [10,11,12]. Thus, while replicative senescence is a genetically encoded and epigenetically regulated mechanism that initially functions as an adaptive tumor-suppressive program, the persistence and accumulation of metabolically active senescent cells, along with their tissue-level effects, extend far beyond its original selective purpose, aligning with a regulated quasi-program model linking telomere dysfunction to organismal aging.

Attention subsequently shifted from telomere length to the chromatin architecture of telomeric and subtelomeric regions. In 2004, Fossel proposed that replicative aging is epigenetically programmed through progressive changes in telomeric heterochromatin, proposed as a protective “hood” of fixed size [13]. As telomeres shorten, this heterochromatic domain has been hypothesized to slide inward along the subtelomere, modulating transcriptional repression in a telomere-length-dependent manner. This model builds upon the phenomenon of telomere position effect (TPE), originally inspired by position-effect variegation in Drosophila melanogaster [14] and later demonstrated at yeast telomeres [15]. Wright and Shay extended these principles to human cells, proposing that telomere length and chromatin organization regulate subtelomeric gene expression during replicative aging [16]. Subsequent evidence of TPE in mammalian systems reinforced the concept that telomere shortening can influence transcriptional programs before terminal telomere dysfunction [17,18]. In this context, the fixed size of a shifting heterochromatin “hood” proportional to the original telomeric structure was proposed as a mechanism to maintain relative repression despite intercellular variation in telomere length [16,19]. Telomere shortening thus becomes not only a passive erosion of DNA repeats but a structural perturbation of chromatin boundaries that can reshape subtelomeric transcriptional landscapes. Thus, cells are not only “counting divisions” but also interpreting telomere state, which depends on both length and chromatin configuration [20,21].

Collectively, these findings form the basis of the subtelomere-telomere theory of aging, in which TPE-mediated regulation of subtelomeric loci, including telomeric repeat-containing RNA (TERRA), links telomere shortening and changes in gene expression to replicative senescence [16,19,22]. Building upon this concept, the present review develops the chromatin-sensing telomere theory of aging, integrating advances in telomere chromatin biology, TERRA regulation, and MLL/KMT2A lysine methyltransferase function into a unified mechanistic framework for replicative senescence and aging.

2. The Chromatin-Sensing Telomere Theory of Aging

The chromatin-sensing telomere theory of aging expands the classical telomere attrition model by proposing that telomere shortening progressively remodels the chromatin and regulatory environment at chromosome ends. Rather than acting solely through loss of telomeric DNA repeats, replicative aging is proposed to alter the recruitment and activity of chromatin regulators, thereby modifying subtelomeric transcriptional programs that influence the onset of replicative senescence.

The conceptual origins of this model trace back to the TPE model of cellular senescence proposed by Wright and Shay [16], and later reformulated by Libertini as the subtelomere-telomere theory of aging [23]. An intermediate formulation by Fossel introduced the idea of a protective heterochromatin “hood” of fixed length that progressively shifts inward over subtelomeric regions as telomeres shorten [13]. Evidence supporting progressive inward spread of TPE repression comes from studies of subtelomeric gene expression in fibroblasts derived from ring chromosome 17 syndrome patients, characterized by shortened telomeres at both chromosome 17 ends and significantly reduced expression of adjacent genes [24]. In addition, in isogenic human diploid fibroblasts (HDFs) approaching senescence, progressive telomere shortening is accompanied by reduced expression of genes proximal to telomeric repeats, consistent with inward spread of TPE repression, whereas more distal genes (within ~10 Mb) become upregulated, consistent with resolution of chromosome end looping [25].

Within this framework, telomere shortening alters the recruitment of chromatin modifiers and transcriptional regulators to chromosome ends. A chromatin modifier and transcriptional regulator shown to associate with telomeric chromatin in a telomere length-dependent manner is lysine methyltransferase MLL/KMT2A. MLL binds telomeres preferentially when they are long and catalyzes histone H3 lysine K4 di- and trimethylation (H3K4me2/3) associated with transcriptional activation, thereby facilitating transcription of TERRA [26]. TERRA, first described by Azzalin and colleagues [27], consists of long non-coding RNAs transcribed primarily from subtelomeric promoters located within CpG-islands into telomeric G-rich repeats [28]. Evidence supports a conserved role for TERRA in telomere biology across eukaryotes, including regulation of telomeric chromatin structure, modulation of telomerase activity, and participation in the recombination-based alternative lengthening of telomeres (ALT) mechanism of telomere maintenance in cancer cells. TERRA promotes the formation of telomeric RNA:DNA hybrid structures (R-loops) between TERRA UUAGGG repeats and the complementary C-rich telomeric DNA strand, which facilitate homology-directed repair (HDR) at dysfunctional telomeres [29,30]. Additional findings indicate that TERRA participates in broader regulatory and signaling processes, including modulation of extratelomeric gene expression and signaling of dysfunctional telomeres to the cytoplasm and the extracellular environment [31,32].

In senescent HDFs, progressive telomere shortening is associated with reduced MLL occupancy at telomeres and a concomitant decline in TERRA levels [26]. Together with prior demonstrations that TERRA expression is directly influenced by chromatin-modifying enzymes [29], these findings provide a mechanistic link between telomere length, chromatin state, and transcriptional regulation. In this context, chromatin regulatory complexes such as the MLL complex of proteins associated with Set1 (MLL/COMPASS) and Polycomb repressive complex 2 (PRC2), together with components of the shelterin complex and the tumor suppressor p53, emerge as telomere length-sensitive regulators of TERRA transcription that integrate telomere state with chromatin remodeling, transcriptional control, and checkpoint activation in both cancer and aging [26,27,29,33,34,35]. Notably, the conceptual basis for a telomere length-dependent, MLL-regulated control of TERRA transcription as a determinant of chromatin state was articulated in earlier work, in which progressive telomere shortening was proposed to impair TERRA-mediated epigenetic protection and thereby contribute to the onset of replicative senescence [22].

Importantly, studies in telomerase RNA-knockout (Terc-KO) mice at the fifth generation and in TRF2 transgenic mice demonstrated loss of TERRA transcription and significant chromatin remodeling at shortened telomeres, including reduced telomeric and subtelomeric heterochromatin marks, increased histone acetylation, and unaltered or increased telomere-associated TRF1 and TRF2 levels (Table 1) [29,36,37]. Therefore, evidence does not support a strictly heterochromatin-based nature of the proposed protective “hood”. Instead, structural and biochemical data suggest shelterin, the principal structural and regulatory complex at chromosome ends [38,39], as the primary driver of TERRA repression following telomere shortening.

The chromatin-sensing telomere theory of aging posits a length-dependent chromatin-sensing mechanism in which telomere shortening progressively remodels the shelterin-dependent telomere-subtelomere chromatin environment, thereby reducing MLL and RNA polymerase II (RNA Pol II) accessibility to TERRA promoters and repressing TERRA transcription. In this framework, TERRA serves as a regulatory hub that integrates telomere structure with chromatin regulation, and its progressive reduction contributes to the loss of epigenetic stability. This chromatin remodeling promotes transcriptional reprogramming, for instance, by resolving chromosome end looping, resulting in replicative senescence. The progressive accumulation of replicative senescence over time serves as a quasi-programmed mechanism linking chromosome end dynamics to organismal aging.

1.1. TERRA as a Regulatory Hub in Aging

TERRA has emerged as a central regulator linking telomere biology, chromatin organization, and aging. In mammalian cells, TERRA transcripts have been shown to originate predominantly from specific loci, including the human 20q and Xp subtelomeres, the murine 18q and 9q chromosome ends, distal XqYq pseudoautosomal domains, and 2q and Xq intrachromosomal regions, which mainly contribute to the cellular pool of TERRA [40,41,42,43]. This non-uniform distribution suggests that individual chromosome ends may contribute disproportionately to global TERRA levels and therefore to telomere-dependent chromatin regulation.

TERRA expression is regulated by a network of shelterin proteins, transcription factors, chromatin remodelers, and epigenetic modifiers that couple telomere length and telomeric stress to subtelomeric transcriptional responsiveness. These include TRF1 and TRF2, the tumor suppressor p53 [26,29,33,44], the chromatin remodeler ATRX [45,46], DNA methyltransferases DNMT1 and DNMT3a/b, and the histone lysine methyltransferases SUV39H1/2 and SUV4-20H1/2, MLL, and PRC2 component EZH2 [26,29,34].

Evidence from murine embryonic stem cells (ESCs) and induced pluripotent stem cells (iPSCs), mouse embryonic fibroblasts (MEFs), and human cells demonstrates that disruption of major heterochromatin regulators alters TERRA expression and telomere homeostasis [29,47,48,49,50,51,52]. Loss of SUV39H1/2 or SUV4-20H1/2 activity results in elevated TERRA levels, reduced H3K9me2/3 or H4K20me3 heterochromatin marks, decreased heterochromatin protein 1 (HP1) recruitment, increased telomere recombination, and telomere elongation (Table 1) [29,47,48,49]. Likewise, human cells from individuals with immunodeficiency, centromeric instability, and facial anomalies (ICF) syndrome carrying DNMT3B mutations exhibit subtelomeric DNA hypomethylation, increased H3K4me2, reduced H3K9me3, elevated TERRA expression, and short telomeres (Table 2) [51,52]. Collectively, these studies establish TERRA as a sensitive indicator of subtelomeric chromatin organization.

A bona fide TPE also regulates TERRA expression. Early studies demonstrated that genes placed adjacent to newly formed telomeres become transcriptionally repressed [17]. Subsequent analyses in human cancer cells and primary fibroblasts transgenic for telomerase showed that telomere elongation represses TERRA transcription through increased H3K9me3 levels and HP1α activity at chromosome ends (Table 2), establishing a length-dependent chromatin environment that restricts telomeric transcription [53]. These findings provide evidence that TERRA expression responds to telomere length and local heterochromatin density.

Beyond its transcriptional regulation, TERRA functions as a molecular scaffold that coordinates the recruitment of chromatin modifiers to telomeres. RNA affinity purification, RNA immunoprecipitation, and quantitative proteomic studies identified SUV39H1, SETDB1, DNMT1, MLL, and PRC2 core components EZH2, EED, SUZ12, and RBBP4/7 among TERRA-associated proteins (Table 3) [31,34,42,44,54]. These interactions are consistent with a co-transcriptional gene-silencing model [55], in which TERRA transcription establishes a negative feedback loop that sustains the formation and maintenance of telomeric heterochromatin, thereby reinforcing a bona fide TPE.

A particularly informative example comes from ALT-positive U2OS osteosarcoma cells (Table 2), where the subtelomeric 20q locus contributes approximately 20–50% of total cellular TERRA [34,40]. Deletion of this locus (20q-TERRA KO) reduces H3K9me3, H3K27me3, H4K20me3, and HP1 occupancy at telomeres, increases the abundance of critically short telomeres, and impairs telomere maintenance [34]. These findings indicate that TERRA production is not equivalent across all chromosome ends and that specific subtelomeres can contribute disproportionately to total cellular TERRA levels. This supports the testable hypothesis that critical shortening of these TERRA-dominant telomeres may exert a global influence on telomeric chromatin regulation and the onset of replicative senescence.

Additional evidence supporting a role for TERRA in aging comes from telomerase-deficient systems. In Terc-KO MEFs from fifth-generation mice (Table 1), progressive telomere shortening is accompanied by a reduction in TERRA expression and concomitant heterochromatin relaxation, as evidenced by reduced H3K9me3, H4K20me3, and DNA methylation levels and increased histone acetylation [29,37]. A similar phenotype is observed in HDFs approaching replicative senescence (Table 2), where progressive telomere shortening leads to reduced MLL binding, decreased H3K4me2/3 levels, and downregulation of TERRA, a phenotype recapitulated by MLL depletion [26]. Together, these findings identify TERRA as a central regulatory hub linking telomere length, chromatin organization, and replicative senescence. Importantly, TERRA regulation cannot be explained solely by heterochromatin density but, instead, reflects the integration of activating and repressive chromatin pathways.

Beyond its local functions at chromosome ends, TERRA can also exert regulatory activity in trans at extratelomeric loci. In murine ESCs, TERRA-binding sites are enriched within gene bodies, a subset of which are also occupied by ATRX, with TERRA generally associated with gene activation and ATRX with transcriptional repression [31]. In Trp53-KO murine iPSCs, TERRA transcripts occupy Polycomb target genes genome-wide and modulate their transcriptional landscape through TRF1-dependent recruitment of PRC2 and deposition of H3K27me3 marks (Table 1) [56]. These findings extend TERRA’s role beyond a telomere-restricted chromatin scaffold and support a broader function in epigenetic regulation of gene expression.

Within the chromatin-sensing telomere framework, progressive loss of TERRA is proposed to impair chromatin maintenance at telomeres, promoting a self-reinforcing cycle of chromatin destabilization and telomere dysfunction. Beyond chromosome ends, progressive TERRA dysregulation may also alter distal transcriptional networks through its extratelomeric epigenetic functions. Although no direct evidence currently links these activities to the induction of replicative senescence, they suggest that TERRA dysregulation during telomere shortening contributes to broader transcriptional reprogramming.

1.2. MLL as a Chromatin Sensor at Telomeres

MLL, also known as MLL1/KMT2A (Table 3), originally identified as the mammalian ortholog of the Drosophila melanogaster’s Trithorax (Trx) protein, is the founding member of the KMT2 family of H3K4 methyltransferases (MLL1-4/KMT2A-D, SET1A-B/KMT2F-G) all of which bind to a multiprotein complex evolutionarily conserved in yeast Saccharomyces cerevisiae termed COMPASS, for complex of proteins associated with Set1, with Set1 representing the only H3K4 methyltransferase in yeast [57]. MLL/COMPASS was initially recognized for maintaining transcriptional activation of Homeobox and developmental genes [58], thereby antagonizing Polycomb repressive complexes such as PRC1 and PRC2. PRC2-catalyzed H3K27me3 and the resulting chromatin compaction ensure stable propagation of repressive states through cell divisions [59]. The balance between H3K4me3 and H3K27me3 represents a fundamental chromatin regulatory axis.

Evidence from our group and others extended the function of MLL/COMPASS to telomeric and subtelomeric chromatin, positioning MLL as a chromatin sensor capable of translating telomere shortening into transcriptional and checkpoint responses. Chromatin immunoprecipitation analyses in HDFs demonstrated that MLL binds telomeric chromatin in a telomere length-dependent manner, promotes H3K4 methylation, and cooperates with p53 in the transcriptional upregulation of TERRA at TRF2-depleted dysfunctional telomeres, as part of the DNA damage response (DDR) [26]. MLL/COMPASS physically interacts with p53 and enhances p53-dependent transcription in vitro [60]. Within the COMPASS core subunits, WDR5 positively regulates p53 stability by inhibiting p53 ubiquitination [61]. These findings link MLL-dependent chromatin modification to checkpoint activation pathways. In MEFs, Mll KO prevents TERRA induction and abrogates DNA-damage-triggered p53 phosphorylation on Ser15 [p-p53 (Ser15)] under telomeric stress, while also causing telomere elongation [22,26]. These results indicate that MLL-dependent TERRA transcription couples chromatin remodeling to both DDR signaling and control of telomerase activity.

Proteomic analyses further reinforce MLL integration into telomeric regulatory networks (Table 3). RNA affinity purification and quantitative mass spectrometry studies in human and murine cells identify MLL, COMPASS core subunits RbBP5, WDR5, and DPY30, and complex-specific partners menin, PSIP1, HCF1, and WDR82 among TERRA-associated proteins [42,54]. COMPASS subunits WDR5, menin, PSIP1, and WDR82 are also among the proteins associated with mammalian telomeric chromatin, with menin initially localized at meiotic telomeres [62,63]. Among the KMT2 family members, only MLL shows association with telomeric chromatin and the TERRA interactome. Other paralogs (MLL2-4/KMT2B-D and SET1A-B/KMT2F-G) broadly regulate gene and enhancer activity but lack direct evidence of association with TERRA or telomeric localization (Table 3). This distinction underscores MLL’s unique role in coupling telomere structure to transcriptional control.

Evolutionary conservation of this principle is evident in yeast. Deletion of Set1 causes loss of H3K4 methylation, derepression of subtelomeric genes, and telomere shortening [64,65]. Deletion of COMPASS components Spp1 and Sdc1 similarly reduces H3K4 methylation and disrupts subtelomeric repression, confirming that intact COMPASS function is required for telomeric chromatin maintenance and TPE [65,66]. Set1 domain analyses demonstrate that the SET catalytic domain is essential for these effects, whereas deletion of RNA recognition motifs RRM1 and RRM2 only partially reduces H3K4 methylation and causes mild derepression. Importantly, yeast cells lacking both Set1 and telomerase activity senesce more rapidly than telomerase single mutants despite comparable telomere length, indicating that Set1 may contribute to telomere stability independently of bulk telomere length. In fission yeast Schizosaccharomyces pombe, deletion of SpSet1 results in telomeric derepression and telomere elongation [67], paralleling the extended telomeres observed in Mll-KO MEFs. These findings support an evolutionarily conserved role for H3K4 methyltransferases at telomeric boundaries. Although the precise mechanism by which H3K4 methylation enforces telomeric silencing in yeast remains unclear, subtelomeric and telomeric RNAs, including TERRA, have been identified in budding and fission yeast, where they interact with telomerase and chromatin regulators and form R-loops to influence telomere stability [68,69,70]. While direct Set1 regulation of TERRA-like transcripts has not been demonstrated, Set1 localization at telomeres and its control over TPE suggest an evolutionary precursor to the mammalian MLL-TERRA axis.

In Drosophila melanogaster, the MLL ortholog Trx further links these H3K4 methyltransferases to lifespan regulation. Trx acts as an antagonist of Polycomb-group protein silencing. In flies heterozygous for mutations in the PRC2 core subunits E(z) and Esc, lifespan is significantly extended. Mutations in Trx suppress this lifespan extension by restoring high H3K27me3 levels in PRC2 mutants [71]. These findings demonstrate that the balance between Trx-catalyzed H3K4me3 and E(z)-catalyzed H3K27me3 influences organismal longevity.

Taken together, evidence from yeast Set1, fly Trx, and mammalian MLL reveals H3K4 methyltransferases as chromatin boundary regulators that integrate transcriptional activation with repressive chromatin control at genomic edges, including telomeres. In mammals, MLL uniquely links telomere length, TERRA transcription, and p53-mediated checkpoint activation. These evolutionary parallels support the interpretation that MLL functions as a chromatin sensor, translating telomere structural state into transcriptional and aging-related outcomes.

1.3. Shelterin-mediated Repression of TERRA

Wright and Shay proposed that TPE shifts inward as telomeres shorten [16]. However, the concurrent loss of TERRA transcription and telomeric chromatin marks observed in Terc-KO MEFs at the fifth generation and in late-passage HDFs, along with their unchanged telomere-associated TRF1 and TRF2 levels despite significant telomere shortening, suggests that a protective “hood” repositioning over subtelomeric regions cannot be due merely to heterochromatin [26,37]. Instead, the shelterin complex is proposed as a plausible structural mediator of this shift, a specialized six-protein assembly composed of TRF1, TRF2, RAP1, POT1, TIN2, and TPP1 that binds telomeric DNA to preserve chromosome end capping [38,39]. These observations suggest that telomere shortening may remodel the shelterin-dependent telomeric chromatin topology, progressively modifying the accessibility of subtelomeric promoters to transcriptional regulators such as MLL and RNA pol II. However, the precise mechanistic coupling between telomere shortening, shelterin organization, and MLL recruitment remains to be fully elucidated.

TRF1 and TRF2 interact with TERRA, with TRF2 binding via its amino-terminal GAR and carboxy-terminal Myb domains, promoting proper TERRA localization to telomeres [35]. Functional studies show increased TERRA transcription after TRF2 knockdown (KD) in HDFs and cancer cell lines (Table 2) and Terf2 KO in MEFs (Table 1) [26,44,72]. Mechanistically, TRF2 represses TERRA transcription via its TFRH homodimerization domain, which promotes chromatin compaction and reduces telomere accessibility to RNA pol II [44,73]. Terf2 KO in SV40-immortalized MEFs, TRF2 KD in Trp53-KO mouse iPSCs (Table 1), and TRF2/p53 double-KD in HDFs (Table 2) fail to induce TERRA upregulation, indicating that p53 transactivation activity is required for dysfunctional telomere transcription in normal or pluripotent cells [26,42,56]. In contrast, TRF2 depletion induces global TERRA upregulation in both wild-type and Trp53-KO HCT116 cancer cell lines, whereas etoposide-induced DNA damage in the same cells triggers a p53-dependent increase in TERRA transcripts from specific subtelomeres (Table 2) [33,44]. These findings indicate that TRF2-depleted telomeres induce TERRA transcription across multiple cellular contexts, whereas p53 transcriptional activity is specifically required in normal or pluripotent cells but is partially dispensable in cancer cells.

Transgenic studies further clarify the distinct roles of TRF2 in chromosome-end protection and regulation of telomeric chromatin. In HDFs, TRF2 overexpression accelerates telomere shortening while protecting critically short telomeres from end-to-end fusions, without altering the timing of replicative senescence [74]. Yet, cells ultimately undergo replicative senescence despite preserved chromosome-end protection. Maintenance of the canonical TRF2 capping function alone is therefore insufficient to prevent the senescence program. In transgenic mouse keratinocytes (Table 1), increased TRF2 levels similarly accelerate telomere shortening while repressing TERRA transcription and remodeling telomeric chromatin, leading to reduced telomeric and subtelomeric heterochromatin marks, increased histone acetylation, decreased nucleosome density, and greater nucleosomal spacing [36]. Conversely, in ICF lymphoblastoid cell lines and fibroblasts carrying DNMT3b mutations (Table 2), increased TERRA levels are accompanied by decreased TRF2 binding at subtelomeres, shortened telomeres, and premature senescence at a low population doubling [51,52,75]. Collectively, these findings support a model in which TRF2 contributes not only to chromosome-end protection but also to regulation of the transcriptional and chromatin state of the telomere-subtelomere domain, with TERRA expression reflecting its functional organization rather than telomere length alone.

TRF1 also plays a context-dependent role in TERRA regulation (Table 1). In mouse C2C12 myoblasts, TRF1 depletion reduces TERRA levels without altering RNA pol II recruitment, suggesting regulation at the level of elongation or transcript stability [29]. No change in global TERRA levels is observed in Terf1-KO MEFs, whereas Terf1-KO mouse keratinocytes show increased TERRA [56,76]. Unlike TRF2, TRF1 depletion in Trp53-KO mouse iPSCs increases TERRA levels, supporting a DDR-independent mechanism [56]. Similarly, in human ESCs, HeLa and A549 cancer cells, TRF1 loss results in TERRA upregulation (Table 2), indicating a TRF1 repressive role in all these contexts [44,77,78].

Among the other shelterin components, RAP1, POT1, and TPP1 do not robustly immunoprecipitate TERRA, yet RAP1 and POT1 bind to TERRA in RNA affinity purification and iDRiP assays [31,35]. RAP1 deletion produces species- and cell-type-specific effects. In MEFs and human cancer cells, RAP1 KO minimally affects TERRA, H3K9 methylation, or HP1 binding at telomeres [79,80]. In contrast, RAP1-deficient yeast and human mesenchymal stem cells (hMSCs) exhibit increased TERRA levels, with RAP1-KO hMSCs showing reduced H3K9me2 at telomeres, telomere elongation (Table 2), and delayed replicative senescence [81,82]. POT1 depletion does not alter TERRA levels in cancer cell lines [44], yet in Caenorhabditis elegans, POT1 functions as a negative regulator of TERRA, where its depletion increases telomere-specific TERRA expression [83]. Additional evidence for shelterin-mediated transcriptional restraint comes from herpes simplex virus 1 infection of HDFs, where TERRA expression increases alongside proteolytic degradation of TPP1 and the dissociation of TRF1, TRF2, and histone H3 from telomeric DNA [84]. Although direct TPP1 KD was not tested, coordinated shelterin stripping correlates with transcriptional derepression, supporting a dynamic shelterin-TERRA regulatory axis responsive to telomere stress. Collectively, TRF1, TRF2, RAP1, and POT1 exert context-dependent repression of TERRA and influence telomeric chromatin integrity across species, cell types, telomere states, and epigenetic environments.

These findings refine the proposed TPE inward-shift concept during telomere shortening. Experimental perturbation of shelterin components, particularly TRF2, is sufficient to induce telomere uncapping and cellular senescence [85,86], whereas progressive telomere shortening similarly drives replicative senescence in mammalian cells [87,88]. However, although reduced shelterin occupancy at shortened telomeres has been reported [89], direct quantitative evidence that physiological telomere shortening produces sufficient endogenous shelterin depletion to fully account for replicative senescence remains limited [90]. Interestingly, Karlseder and colleagues demonstrated that TRF2 overexpression preserves chromosome-end protection, delaying senescence despite accelerated telomere shortening, indicating that the functional state of telomeres rather than telomere length alone determines the onset of replicative senescence [74]. Available evidence suggests that substantial shelterin occupancy is retained at shortened telomeres, yet TERRA expression declines as telomeres approach critical length [26,37]. Complementing this observation, MLL positively regulates TERRA transcription [26], whereas TRF2 overexpression represses TERRA transcription and remodels telomeric chromatin, supporting a role for shelterin in regulating the transcriptional accessibility of subtelomeric TERRA promoters [36].

Together, these findings support a model in which progressive telomere shortening induces a shelterin-dependent telomere-subtelomere chromatin remodeling, restricting access of MLL and RNA pol II to subtelomeric promoters and repressing TERRA transcription.

1.4. A Biphasic Model of TERRA Regulation in Aging

Work in telomerase-negative yeast has highlighted a protective function for TERRA in somatic cells undergoing telomere shortening, where moderate TERRA and the resulting R-loop accumulation promote HDR of dysfunctional telomeres, thereby facilitating telomere maintenance and delaying senescence [91,92]. Kyriacou and Lingner, along with others, review this duality, emphasizing that TERRA accumulation can counteract telomere loss by recruiting repair machinery and mitigating the deleterious effects of progressive shortening [93,94].

Increased TERRA levels have been documented in non-pathological human aging tissues, including human blood cells, brain tissue, fibroblasts, and alveolar type II epithelial cells, along with increased R-loop formation, DNA damage, and chronic activation of cGAS-STING signaling, in inverse correlation with telomere length [95,96]. This presenescent upregulation suggests an adaptive response to progressive telomere shortening, in which elevated TERRA levels contribute to chromatin stabilization and repair, thereby buffering the effects of replicative event accumulation and delaying the onset of senescence. In this regard, persistent DNA damage signaling, SASP, and chronic innate immune activation, as reported in aging tissues, appear to arise primarily from the accumulation of senescent cells, the release of their cytoplasmic DNA, and chronic activation of cGAS-STING signaling [10,97], rather than directly from TERRA-dependent R-loop accumulation.

Previous studies have shown that a small number of critically short telomeres is sufficient to trigger replicative senescence [98,99]. Telomere length is highly heterogeneous, and individual chromosome ends exhibit distinct shortening kinetics and marked allelic heterogeneity, as reported for XpYp telomeres, showing progressive as well as abrupt shortening events during replicative aging [100]. Subsequently, TERRA transcription in mammalian cells has been shown to originate preferentially from specific loci, like the human telomeres 20q and Xp [40,41]. Therefore, the present model predicts that the telomeres most likely to trigger senescence are those that both reach critical shortening and primarily contribute to TERRA expression.

In support of this selective TERRA-dominant telomere-shortening model, recent work has revealed specific length distributions of human telomeres, with 20q consistently among the shorter chromosome ends in the aging population [101]. As telomeres progressively shorten, critical erosion of chromosome ends leads to transcriptional repression of TERRA. Because TERRA-dominant telomeres account for a major fraction of total TERRA expression, as shown for 20q in the ALT-positive U2OS osteosarcoma cell line [34], their repression results in a global decrease in TERRA levels rather than a localized effect restricted to individual chromosome ends.

These findings support a biphasic model of TERRA regulation in aging, with MLL identified as a core effector of this transition (Figure 1). In presenescent cells, MLL maintains subtelomeric chromatin in a transcriptionally permissive state, enabling TERRA upregulation in cooperation with p53 as an adaptive response to progressive telomere shortening, thereby supporting telomere maintenance and genome stability, likely through R-loop accumulation. Elevated TERRA in aging human tissues, therefore, may reflect a protective phase rather than a primary cause of chronic inflammation. Critical shortening of TERRA-dominant telomeres subsequently results in impairment of TERRA transcription by MLL/RNA pol II displacement from subtelomeric promoters, compatible with a shelterin-dependent telomere-subtelomere chromatin remodeling. The resulting global decrease in TERRA levels would impair recruitment of chromatin-modifying complexes to telomeres, compromising restoration of heterochromatin and thereby accelerating chromatin relaxation, telomere shortening, and persistent DDR activation, ultimately contributing to p53-dependent replicative senescence. According to this model, replicative senescence is therefore initiated not by isolated dysfunction at random chromosome ends, which tend to be protected by R-loop accumulation, but by the critical shortening of a subset of telomeres with prevalent TERRA production that function as critical nodes in telomere biology.

3. Future Directions

Despite significant advances in understanding the interplay among telomeres, subtelomeres, shelterin, and chromatin regulators, several key mechanistic questions remain important to validate the chromatin-sensing telomere framework. The most important challenge is to define how progressive telomere shortening is converted into loss of MLL occupancy at subtelomeric TERRA promoters. Whether shortening-induced changes in shelterin-dependent telomeric chromatin topology progressively restrict access of MLL and RNA pol II to subtelomeric TERRA promoters remains unknown. Direct experimental testing of the proposed shelterin-mediated remodeling process, together with analyses of higher-order telomeric chromatin structure and MLL recruitment dynamics during replicative aging, will be essential to establish this mechanistic link between telomere shortening and senescence.

A second major question concerns the existence and functional importance of TERRA-dominant telomeres. A limited subset of chromosome ends has been shown to contribute disproportionately to total cellular TERRA levels in mammalian cells [40,41,42,43]. This framework predicts that critical shortening or epigenetic repression of these loci may exert a global influence on telomeric chromatin regulation and the onset of replicative senescence. Future studies that combine locus-specific telomere editing, chromatin profiling, and TERRA quantification will be required to determine whether specific chromosome ends serve as critical nodes in telomere biology.

Another priority will be to define how TERRA dynamics are generated across different stages of aging. While replicative senescence is associated with TERRA repression, several aging tissues exhibit elevated TERRA levels. Determining whether these changes arise from altered transcription, RNA stability, or R-loop persistence, as compensatory responses to telomere shortening, will be critical for understanding how TERRA participates in aging. Longitudinal studies integrating telomere length measurements, chromatin-state analyses, TERRA expression, and senescence markers across tissues and species will be particularly valuable for assessing the proposed biphasic model of TERRA regulation during aging.

Finally, the MLL-TERRA axis represents a potential target for interventions aimed at promoting healthy longevity. An unconventional experimental approach arises from observations in MLL-rearranged acute lymphoblastic leukemias, where TERRA levels remain elevated independently of telomere length and ploidy [102]. These findings raise the possibility that oncogenic MLL fusion proteins may bypass the telomere-length-dependent repression of TERRA. Engineered recruitment of MLL-rearranged transcriptional complexes to TERRA-dominant subtelomeres could provide an experimental strategy to evaluate whether sustained TERRA expression delays replicative senescence. Although direct therapeutic application of MLL fusion proteins would be precluded by their oncogenic activity, dissecting the molecular basis of their sustained activation of TERRA may reveal strategies for preserving subtelomeric transcriptional competence during aging.

4. Conclusions

By integrating the telomere position effect model of cellular senescence with the subtelomere-telomere theory of aging and experimental evidence linking MLL to TERRA regulation, this review proposes a chromatin-sensing telomere framework in which progressive telomere shortening is translated into replicative senescence through epigenetic and transcriptional mechanisms operating at chromosome ends. Within this model, telomere shortening functions as a replication-dependent timing mechanism that initiates a regulated chromatin response involving shelterin components, chromatin modifiers, and TERRA. Evidence from mammalian cells supports a biphasic model of TERRA regulation during aging, consistent with an initial adaptive response to telomere shortening followed by transcriptional repression during replicative senescence (Figure 1).

This framework suggests that replicative senescence is not simply a direct consequence of critically short telomeres, but the outcome of progressive changes in telomere-associated chromatin and transcriptional regulation. We propose that late-stage critical telomere shortening may remodel the shelterin-dependent telomeric chromatin topology, progressively restricting access of transcriptional regulators such as MLL and RNA pol II to subtelomeric promoters. A further testable hypothesis, arising from the integration of current evidence, is that a restricted subset of TERRA-dominant telomeres overly contributes to global TERRA output and may therefore be the critical determinant of senescence initiation.

Because senescent cells can persist for prolonged periods within tissues, repeated activation of replicative senescence throughout life results in their progressive accumulation. The progressive telomere shortening documented in multiple human and animal tissues [103], together with the accumulation of senescent cells during aging [104], supports the view that replicative senescence contributes to organismal aging beyond its established tumor-suppressive function. Within this framework, telomere shortening serves as the timed initiator of replicative senescence, whereas the resulting accumulation of senescent cells drives a quasi-programmed decline in tissue function, further amplified by stochastic molecular damage and chronic inflammatory signaling.

Testing the predictions generated by the chromatin-sensing telomere framework will determine whether progressive telomere shortening is translated into replicative senescence through a chromatin-dependent regulatory program centered on the MLL-TERRA axis. Validation of this model would establish telomeric chromatin remodeling as a mechanistic link between cellular senescence and organismal aging, while providing a basis for future strategies to preserve subtelomeric transcriptional competence and promote healthy longevity.