Therapy-induced senescence is finally escapable, what is next?

ABSTRACT

Several breakthrough articles have recently confirmed the ability of tumor cells to escape the stable cell cycle arrest imposed by Therapy-Induced Senescence (TIS). Subsequently, accepting the hypothesis that TIS is escapable should encourage serious reassessments of the fundamental roles of senescence in cancer treatment. The potential for escape from TIS undermines the well-established tumor suppressor function of senescence, proposes it as a mechanism of tumor dormancy leading to disease recurrence and invites for further investigation of its unfavorable contribution to cancer therapy outcomes. Moreover, escaping TIS strongly indicates that the elimination of senescent tumor cells, primarily through pharmacological means, is a suitable approach for increasing the efficacy of cancer treatment, one that still requires further exploration. This commentary provides an overview of the recent evidence that unequivocally demonstrated the ability of therapy-induced senescent tumor cells in overcoming the terminal growth arrest fate and provides future perspectives on the roles of TIS in tumor biology.

The long-standing paradigm that senescence is associated with an irreversible form of growth arrest has been recently subject to reconsideration. Recent evidence has shown that certain forms of senescence, with focus on Therapy-Induced Senescence (TIS), are coupled with a durable form of cell cycle arrest. TIS is a variant of senescence that develops in tumor (and non-tumor) cells in response to the exposure to an array of cancer chemotherapeutics, both conventional and targeted, and radiation [1] (Figure 1). TIS has always been described to impose a terminal cell fate where tumor cells succumb into an inescapable growth arrest which accounts for its desirable antitumor effect [24]. Despite the growing evidence that highlights the contribution of TIS in ominous malignant processes and the partial acceptance in the field that TIS might be “bad” or at least “not as good” (largely attributed to senescence-associated non-cell-autonomous mechanisms), there has been insistence that the definition of senescence is primarily based on the irreversibility of its growth arrest and that the observed untoward outcomes of senescence are not based on cell-autonomous processes. This has led to a significant uncertainty in the field in interpreting novel observations that argue against the stability of the Senescence-associated Growth Arrest (SAGA). Furthermore, it introduced confusing terminology into the literature, wherein an observation of durable senescence is experimentally established, the overall cellular phenotype, despite consistently exhibiting a spectrum of senescence-associated biomarkers, can only be coined as “senescence-like” or “pseudosenescence” [7,23,25]. Finally, and based on an evolving body of evidence, the potential for escaping TIS is gradually being adopted [26–29], but now, the meaning of this new conception needs to be put into perspective. In that, concepts like senescence bypass, reversal, and escape, as laid out by Evangelou et al. [30], must be considered to avoid confusion. Evangelou et al. defines senescence reversal as the ability of cells to return to a state of “presenescence” after a senescent state has been established, and senescence escape as the potential for cells to resume proliferation from a completely growth-arrested senescent state [30]. Indeed, the fundamental focus of our manuscript is on escape from TIS, since reversal of senescence, or senescence reversibility, at least as defined by Evangelou et al., has not been well documented, and most previous evidence (discussed below) suggests events of escape from TIS rather than reversal.

Figure 1.

Timeline of major events in the evolution of TIS escape. Years following the discovery of Replicative Senescence (RS) by Hayflick and Moorhead in human fibroblasts [2], it was possible to identify senescent cells in culture based on the upregulation of the Senescence-associated β-galactosidase (SA-β-gal) [3]. Late in the twentieth century, separate observations by the Wahl, Ames, and the Adami groups suggested for the first time that DNA damaging agents such as ionizing radiation, bleomycin, or actinomycin can trigger an accelerated (or premature) form of senescence in human fibroblasts [4–6], which was then shown to occur in a similar fashion in tumor cells which conserve the senescence machinery despite being immortalized [7]. At the same time, evidence of the existence of SA-β-gal positive cells was demonstrated by te Poele et al. in human breast cancer tissue exposed to chemotherapy suggesting that Therapy-Induced Senescence (TIS) occurs in vivo. The Bernards group, soon after, has provided first clue on the possibility of RS reversal upon deactivation of major senescence regulators such as p53 or p16^INK4a [8], which paved the way for the consideration of the possibility of senescence escape [9]. Work by the Wu group, then, provided observations confirming the ability of tumor cells to escape TIS [10,11], which was supported by similar findings by other groups [12–15]. As more complex roles of TIS in tumor microenvironment [16,17], anticancer immune response [18], and the development of several forms of therapy-related adverse effects were established [19], the discovery of senolytics has paved the way for considering their potential use as adjuvant cancer treatments in the so-called “one-two punch” approach coined by Wang and Bernards [20–22]. Recently, the contribution of TIS in the development of cancer recurrence in patients with acute myeloid leukemia has been demonstrated by Duy et al. [23]. Collectively, escape from TIS, its contribution to cancer recurrence, immunoevasion, and therapy-induced adverse effects should invite for further investigation of the role of senescence in tumor dormancy and encourage for the development of more effective and safer senescence-eliminating pharmacotherapy for cancer treatment.

Several reports have supported the potential for escaping TIS in tumor cells and proposed molecular mechanisms that facilitate this escape. These mechanisms range from the spontaneous to artificial (experimental interventions) and include loss of the growth control of central cell cycle regulators that promote the escape from the SAGA including p16^INK4a [31], p21^Cip1 [32], Cdc2/Cdk1, Cdk4, cyclin D1, cyclin B1 and c-Myc [33–35]. Moreover, the resolution of oxidative stress caused by gene expression changes mediated by the nuclear relocalization of p65 has been shown to facilitate escape from the SAGA [36]. Extensive work by the Sikora group has also suggested that polyploidy-driven endoduplication is permissive, and somewhat obligatory, for senescent tumor cells to give rise to proliferating progeny [12,37,38]. Furthermore, senescent tumor cells, which frequently exhibit increased levels of autophagy, can overcome metabolic distress and repopulate [13,39], often coupled with their ability to cannibalize on neighboring cells [40]. In addition, escape from senescence can also be driven by constant non-cell-autonomous stimulation by components of the Senescence-associated Secretory Phenotype (SASP) in both autocrine and paracrine fashions [41]. Escaping senescence, of course, is certainly not a stochastic event, and requires extensive genetic and epigenetic reprogramming that does not merely describe the resolution of the state, but rather, gives rise to new phenotypes that are more malignant, aggressive, and potentially resistant to treatment [42–44]. Lastly, escaping senescence was not restricted to TIS tumor cells models and was also recorded in Oncogene-induced Senescence (OIS) and Replicative Senescence (RS) cell models, strongly suggesting that it is a potentially universal event that can occur under more stable SAGA settings [8,45,46].

A major concern in some of these observations was the possibility of non-senescent cell variants in a heterogeneous therapy-exposed tumor cell population of giving rise to proliferative recovery. This issue has been addressed, in part, by our previous work which employed a rigorous senescent cell enrichment based on the expression of the Senescence-associated β-galactosidase (SA-β-gal) where the division of individual, sorted was unequivocally visualized [14]. However, the heterogeneity of TIS in tumor cells remains a concern [47], since, even at rigorous enrichment (cell sorting) condition, it is still possible that a certain subpopulation of SA-β-gal-positive tumor cells is not in a complete senescent state, especially that the expression of SA-β-gal can also be detected in non-senescent cells raising questions regarding its specificity [48–50].

Escaping TIS carries several significant consequences on its overall contribution to tumor cell fate. First, the acute tumor suppressor role of TIS as a favorable response to cancer therapy should now only be considered as temporary and unnecessarily a sought outcome of treatment. Second, the ability of senescent tumor cells to generate phenotypes with more malignant traits and stem-cell-like behavior indicates that the long-term effects of TIS are detrimental [51–54]. Third, if senescent tumor cells can truly resume proliferation after a period of growth stagnation, then TIS becomes a plausible phenotype that explains how dormant tumor cells can then reemerge and contribute to cancer recurrence, and could replace the currently prevailing quiescence-based tumor cell dormancy theory [55]. What makes senescence a more credible form of tumor cell dormancy is that it can share several established hallmarks of dormancy, including escape from immunosurveillance [18,56–58], promotion of angiogenesis [59,60], and upregulation of autophagy [61–63]. Moreover, TIS, whether of tumor malignant cells or stromal nonmalignant cells, has been directly linked to the promotion of metastasis and emergence of recurrence [16,23,64–67], an effect that is largely attributed to the SASP [19]. However, in this aspect, a clear distinction of how tumor cell SASP vs. stromal cell SASP vary in their contribution to immunoevasion, plasticity, and recurrence requires further investigation [68], especially that senescence is associated with substantial transcriptomic heterogeneity, even at the single-cell level, and can give rise to a versatile SASP profile based on the cell type [69–71]. Despite that escape from TIS has been demonstrated in vivo (in animal models of cancer senescence), it is yet to be proven of whether it occurs in humans especially that recent evidence has strongly shown that TIS is a cardinal component of tumor biology and a primary response to the chemotherapy [72–76]. Lastly, although unrelated directly to the escape from TIS but contributes to its ultimately harmful effects, TIS, mainly induced in nonmalignant cells, has been strongly linked to the adverse effects of chemotherapy [19], including neurotoxicity such as peripheral neuropathy [77–80] and cognitive dysfunction [81,82], cardiotoxicity [83], ovarian injury [84,85], and cancer cachexia [86].

The escape from TIS, its potential connection to dormancy and recurrence, and contribution to chemotherapy-induced adverse effects, have all provided the basis for the use of senescence-eliminating or -modulating pharmacotherapy, namely senolytics or senomorphics, for cancer treatment [20]. The use of these compounds has been heavily investigated in several TIS models and showed promising outcomes both in vitro and in vivo with the premise of eliminating senescent tumor cells prior to their ability to regain proliferative capacity [87–93]. This approach is famously known as the “one-two” punch strategy for cancer treatment [20,94]. Senolytic treatment not only successfully culls the number of senescent tumor cells but directly results in the reduction of metastases in tumor bearing animals, or occasionally, the complete eradication of senescent tumor cells and metastases [95]. Moreover, the use of senolytics has been extended to eliminating premalignant senescent cells, senescent immune cells, and senescent nonmalignant cells with clear benefit to preventing the initiation and progression of cancer and ameliorating several chemotherapy-induced adverse effects [96–99]. However, the translation of senolytics for cancer treatment is still premature, as the most widely used compounds are complicated by issues of universality, selectivity, and toxicity [100–102]. For example, after careful reexamination of the data, some frequently tested senolytics have shown only weak or largely nonselective senescence-killing effect in different tumor cell models with a significantly less robust efficacy in comparison with their efficacy against aging-associated senescent cells [103–107]. Moreover, some senolytics exhibit senescence-independent effects [108], succeed in interfering with senescence but with little or no overall benefit [109], or, surprisingly, produce opposing, detrimental effects that lead to functional deterioration [110]. Subsequently, alternative senolytic approaches are being considered including the use of senescence-targeted CAR T cells [111], antibody-targeted senolytics [112], pharmacological conjugation of senolytics to senescence-specific substrates [113], and proteolysis-targeting chimera (PROTAC) [114]. Efforts must continue to investigate and explore the use of current senolytics or identify new ones for cancer treatment taking advantage of high throughput drug screening [115], machine learning [116], and the design of multi-target compounds [117]. Lastly, the escape from TIS should pave the way for exploiting molecular targets that contribute to the ability of tumor cells to escape senescence as a priority since it has the potential to mitigate cancer recurrence. Although targeting nonmalignant senescent cells can also be of value, it can be associated with untoward adverse effects, and considered secondary to the use of senotherapeutics to target senescent tumor cells.

Overall, the discovery of TIS escape has opened the door for a serious reconsideration of the definition of senescence and corroborated the recent efforts that have been proposing senescence as an unfavorable and unwanted outcome of cancer treatment (Figure 1). Further, it should promote the serious consideration of TIS as a mechanism or one form of tumor dormancy which would certainly require the design of state-of-the-art experimental models. Finally, the reversibility of TIS must be the primary target for newly developed senotherapeutics, to interfere with the most detrimental outcomes of TIS: cancer relapse.

Funding Statement

Dr Tareq Saleh is currently funded by the Deanship of Scientific Research, The Hashemite University (grant no. 743/51/2022).

Disclosure statement

No potential conflict of interest was reported by the author(s).