Telomere shortening is not sufficiently counteracted in most human cells and this physiological event contributes to limit cell proliferation (Allsopp et al., 1992). Indeed, critically shortened and/or dysfunctional telomeres elicit a DNA damage response (DDR), which triggers cell death or an irreversible arrest in cell division, the latter known as “replicative senescence” state or the “M1 stage” (IJpma and Greider, 2003; d’Adda di Fagagna et al., 2003; Galli et al., 2021). It is noteworthy that mitochondrial defects, non-telomeric DNA damage, chromatin changes, and oncogenes activation also contribute to replicative senescence onset (Gorgoulis et al., 2019).

The exact mechanisms through which telomeres trigger replicative senescence onset are still not fully understood. In yeast cells, a single very short telomere is sufficient to trigger replicative senescence, while in mammalian cells multiple (from 5 to 10) very short telomeres are required (Zou et al., 2004; Kaul et al., 2011; Xu et al., 2013). However, the exact threshold length and, more importantly, the exact state of the shortest telomere(s) triggering replicative senescence are still unclear. Moreover, i) how short or damaged telomeres activate and maintain the DDR, ii) whether and how cells decide to repair them, and iii) how DNA repair contributes to replicative senescence onset are still open questions.

Interestingly, single-cell studies in the model organism S. cerevisiae have shown that telomerase-negative yeast cells often undergo transient periods of cell division arrest followed by resumption of normal growth before finally entering a “terminal” state (replicative senescence) (Xu et al., 2015; Coutelier et al., 2018). Importantly, “non-terminal” arrests start to occur early after telomerase inactivation, when telomere shortening is not yet critical, thus indicating that this event is not the triggering signal. However, the frequency of transient arrests increases with ongoing cells divisions and telomere shortening, thus suggesting that short/dysfunctional telomeres contribute to reach the terminal arrest (Figure 1). Interestingly, transient proliferation arrests preceding replicative senescence have been described in human cells too (Ghadaouia et al., 2018; Ghadaouia et al., 2021).

These transient arrests indicate the presence of DNA damage and DDR activation. Several hypotheses could explain how cells resume growth: telomeric damage i) is resolved, ii) activates a weak DDR that is tolerated, iii) remains unrepaired and the DDR is switched off in a process called “adaptation” (Figure 1). These findings also suggest that different signals reflecting different telomere length/state exist and they dictate whether the arrest is transient or permanent.

The exact trigger of the terminal arrest is still unclear, especially in mammalian cells. Here, multiple short/dysfunctional telomeres might be required to pass a DDR activation threshold.

Interestingly, genome instability increases upon subsequent non-terminal arrests both in yeast and in mammals (Coutelier et al., 2018; Ghadaouia et al., 2018; Ghadaouia et al., 2021). An intriguing hypothesis suggests that dysfunctional telomeres are not direct inducers of replicative senescence, but they rather lead to a wider genome instability and a strong DDR activation that together trigger senescence onset (Figure 1). Accordingly, adaptation to DNA damage is a potent source of genome instability (Pizzul et al., 2022), which may accumulate upon successive transient arrests. In human cells, telomere uncapping can be tolerated leading to a transient arrest (Ghadaouia et al., 2018; Ghadaouia et al., 2021), but subsequent cell divisions with dysfunctional telomeres likely fuel genome instability and lead to a stable senescence-associated proliferation arrest (Ghadaouia et al., 2021).

The path to replicative senescence onset must be tightly regulated to ensure a non-pathogenic balance. In fact, the premature accumulation of senescent cells contributes to progeroid syndromes (reviewed in Armanios, 2022). As previously mentioned, replicative senescence affects both cancer suppression and development. For example, solid tumors are rare in individuals with short telomere syndromes associated with premature senescence, whereas long telomeres seem to promote cancer development, likely by postponing senescence onset and increasing the chance for oncogenic mutations to occur (McNally et al., 2019; Armanios, 2022). Thus, since telomere length and protection are critical determinants for senescence onset, their control is essential.

Telomeres constitute a barrier against DNA repair mechanisms. Nonetheless, in yeast and mammals, homology-directed DNA repair (HDR) mechanisms constitute a TMM besides telomerase enzyme (Lundblad and Blackburn, 1993; Bryan et al., 1997; Sobinoff and Pickett, 2017). Telomeres can be lengthened by the recombination-mediated synthesis using the DNA sequences of other telomeres as a template. HDR is important for telomere maintenance in a subset of cancer cells (ALT cells) and in yeast cells (called “survivors”) that bypass replicative senescence and upregulate/deregulate these mechanisms at telomeres (Lundblad and Blackburn, 1993; Le et al., 1999; Apte and Cooper, 2016).

Importantly, HDR is also crucial before senescence onset, but these mechanisms are still poorly understood, especially in mammals. In yeast cells, HDR impairment in the absence of telomerase leads to a drastically premature senescence (Claussin and Chang, 2016). Notably, in telomerase-negative cells, many important HDR factors bind telomeres long before senescence onset (Khadaroo et al., 2009; Pérez- Martínez et al., 2020). Recently, RAD51 defects have been shown to cause premature aging in mouse models (Matos-Rodrigues et al., 2022). HDR activity was also shown to decrease as human fibroblasts get closer to replicative senescence (Mao et al., 2012). Thus, under certain circumstances, recognition of a telomere as a DNA damage is not triggering senescence but rather contributes to its physiological delay.

The exact role of HDR mechanisms in regulating replicative senescence onset is still unclear. Since HDR can elongate telomeres, this event may be directly important in delaying senescence onset (Kockler et al., 2021). By telomere sequencing, rare telomere addition events can be detected in telomerase-negative yeast cells before senescence onset (Claussin and Chang, 2016; Sholes et al., 2022). Moreover, recent nanopore sequencing approaches have shown that, in telomerase-negative yeast cells, subpopulations with sets of elongated telomeres are drastically reduced in the absence of HDR (Sholes et al., 2022), in contrast with previous findings (Claussin and Chang, 2016).

Alternatively, HDR might be required to protect/resolve telomeres under replication stress (Claussin and Chang, 2016; Matos-Rodrigues et al., 2022) and/or limit the signal(s) triggering senescence onset, with no telomere elongation (Figure 2).

As previously mentioned, both in yeast and in human cells, transient cell division arrests have been observed, thus suggesting the existence of a “reversible” telomeric damage. Here is where HDR likely acts (Figure 2), thus preventing premature and frequent terminal arrests. In yeast, indeed, the lack of HDR machinery strongly reduces the frequency of transient arrests (Xu et al., 2015).

Nonetheless, an important question is still open: is HDR a safe process that helps to avoid a premature senescence or is it a backup mechanism that lays the groundwork for senescence onset? A recent work by Kockler and others proposed that, in yeast cells, shortened telomeres undergo HDR events that prevent a premature senescence onset by elongating telomeres, but, at the same time, as cells undergo subsequent cell divisions and arrests, genome instability increases. Accordingly, recent findings in human cells suggest that HDR might occur at uncapped telomeres and these events contribute to increase genome instability and promote senescence onset (Ghadaouia et al., 2018; Ghadaouia et al., 2021).

Altogether these data suggest that HDR delays premature senescence onset, but at the same time it might promote genome instability and set the starting point for senescence onset. Further studies are required to better understand these processes.

Important mechanisms of HDR regulation at telomeres rely on R-loop formation. R-loop is a three-stranded nucleic acid structure where the formation of a RNA-DNA hybrid, mainly during transcription, displaces the second DNA strand. R-loops are receiving huge attention because they play important biological functions but, on the other hand, their dysregulation contributes to genome instability and to the development of different diseases [reviewed in García-Muse and Aguilera (2019) and in Rinaldi et al. (2021)].

R-loops form in many genomic regions, including telomeres. Here, their formation is mainly ascribed to the transcription of TERRA, a long non-coding RNA conserved from yeast to mammals that plays important roles in regulating both TMMs (reviewed in Fernandes et al., 2021; Zeinoun et al., 2023). It is still debated whether R-loops form only in cis during TERRA transcription or even in trans when the transcribed TERRA anneals back to the telomere (Feretzaki et al., 2020). Nonetheless, these structures clearly play important roles in the regulation of HDR at telomeres, both before senescence onset and in ALT cells.

Indeed, several studies in yeast and mammals have shown that TERRA and R-loops levels are high both in pre-senescent and in ALT cells, and they are important for telomere maintenance by HDR (Arora et al., 2014; Graf et al., 2017; Misino et al., 2018; Guh et al., 2022; Misino et al., 2022). Moreover, they regulate telomere length dynamics and senescence onset in pre-senescent yeast cells by promoting HDR: accumulation of R-loops delays senescent onset and this effect requires HDR, while an increased removal of RNA-DNA hybrids at telomeres accelerates senescence onset (Balk et al., 2013; Graf et al., 2017).

How exactly R-loops promote HDR is still unclear. The most likely hypothesis is that, as in other genomic regions, R-loops hamper DNA replication, leading to replication stress and the formation of DNA double-strand breaks (DSBs) that are repaired by HDR. R-loops may also act by recruiting repair factors to telomeres and promote HDR initiation (D’Alessandro et al., 2018). TERRA has recently been shown to interact with RAD51, a key player in HDR, and RAD51 is important for TERRA R-loop formation at telomeres (Feretzaki et al., 2020). Thus, it is also possible that increased levels of TERRA and R-loops may increase the local concentration of RAD51 to sustain DNA recombination. Accordingly, TERRA has been recently shown to interact with proteins involved in several DNA repair pathways (Guh et al., 2022).

Several studies have shown that R-loops must achieve an optimal balance to stimulate HDR. In fact, their excess causes an overload of replication stress or hampers the DNA repair machinery access, thus destroying telomere integrity (Arora et al., 2014; Balk et al., 2014; Silva et al., 2021).

These findings suggest that telomere maintenance via HDR requires several factors regulating R-loop levels, both positively and negatively. Indeed, new factors are continuously identified or shown to be involved at telomeres in addition to other genomic regions (Silva et al., 2019; Guh et al., 2022; Kaminski et al., 2022; Pérez-Martínez et al., 2022).

In this scenario, the role of important factors, like RNA-DNA and RNA helicases, still need to be investigated at telomeres. Indeed, in yeast, two important helicases involved in R-loop regulation, Sen1 and Pif1, orthologs of mammalian SETX and PIF1, have been shown to interact with telomere specifically during replicative senescence (Pérez-Martínez et al., 2022). Many RNA helicases, especially those belonging to the DEAD-box family, have been identified as telomere interactors and their role need to be deeply addressed (Pérez-Martínez et al., 2022).