Cellular senescence: Neither irreversible nor reversible

Abstract

Cellular senescence is a critical stress response program implicated in embryonic development, wound healing, aging, and immunity, and it backs up apoptosis as an ultimate cell-cycle exit mechanism. In analogy to replicative exhaustion of telomere-eroded cells, premature types of senescence-referring to oncogene-, therapy-, or virus-induced senescence-are widely considered irreversible growth arrest states as well. We discuss here that entry into full-featured senescence is not necessarily a permanent endpoint, but dependent on essential maintenance components, potentially transient. Unlike a binary state switch, we view senescence with its extensive epigenomic reorganization, profound cytomorphological remodeling, and distinctive metabolic rewiring rather as a journey toward a full-featured arrest condition of variable strength and depth. Senescence-underlying maintenance-essential molecular mechanisms may allow cell-cycle reentry if not continuously provided. Importantly, senescent cells that resumed proliferation fundamentally differ from those that never entered senescence, and hence would not reflect a reversion but a dynamic progression to a post-senescent state that comes with distinct functional and clinically relevant ramifications.

Figure 1.

Senescence escape versus bypass. (A) A variety of pro-senescent triggers (blue bolt) that account for RS, OIS, TIS, or VIS, respectively (left), and senescence-overriding capacities (red oval) that disable the cellular capacity to respond to pro-senescent triggers (as in A) with senescence entry, or promote cell-cycle reentry out of manifest senescence (right). Specific examples include lost expression of the two gene products p16^INK4a and p14/p19^ARF (the latter operating as a p53 upstream activator) encoded by the CDKN2A locus (compromised by mutations, deletions, and/or promoter hypermethylation), p53 inactivation (typically by missense mutations and/or allelic deletion), Rb inactivation (due to mutations or deletions), Myc overexpression (in conjunction with Ras/Braf type oncogenes), or overexpression of histon H3-lysine 9 (H3K9)-active demethylases (such as JMJD2c or LSD1). (B) Transcriptionally repressive H3K9me3-decorated senescence-associated heterochromatin foci (SAHF) formation in the vicinity of S-phase entry-promoting E2F-driven target gene promoters upon recruitment of an H3K9 methyltransferase capacity (such as Suv39h1) that binds in conjunction with heterochromatin protein 1 (HP1) to G1-phase-typical hypophosphorylated Rb protein complexed to E2F transcription factors, thereby firmly blocking the cell in G1 (left). Additional components of the G1/S border control are cyclin-dependent kinase (CDK) four and six inhibitors p16^INK4a and p21^CIP1, the latter a p53 target gene, that counter CDK4/6-cyclin D1-mediated cell-cycle progression (right). Senescence-specific upstream activation of ARF/p53 and p16^INK4a is not entirely clear but involves DNA damage signaling, FoxO transcriptions factors, and the MAPK/ETS cascade (not shown). Key barriers to cell-cycle reentry out of senescence are highlighted in red (and can be overridden by the indicated gene moieties): the H3K9me3 status (disrupted by elevated H3K9 demethylase activities such as JMJD2c or LSD1), repressed CDK4/6 activity (de-repressed by CDK4 amplification and/or inhibitor-insensitive mutations such as CDK4-R24C, or reduced p16^INK4a or p21^CIP1 inhibitor expression), or enforced S-phase entry (e.g., via Myc overexpression). Notably, inducible gene moieties—e.g., a doxycycline-controlled p53-targeting small-hairpin RNAs or a 4-OHT-responsive JMJD2C:ER^TAM fusion—were successfully experimentally employed to enforce a senescence exit. (C) Distinct cellular journeys in which pro-senescent triggers were encountered by senescence-capable versus a priori senescence-incapable cells. Senescence-capable cells respond to pro-senescent triggers with senescence entry, typically associated with wound healing-reminiscent reprogramming into a latent transcriptional stem-like state (e.g., elevated Wnt and Notch signaling). Cells may experience senescence-overriding gene alterations (i.e., either overexpression of senescence-disabling or lost expression of senescence maintenance-essential gene moieties as outlined in A and B) while being in senescence (e.g., due to DNA replication-independent CDKN2A promoter hypermethylation or an inability to continuously reestablish repressive H3K9me3 marks [which are subject to nucleosome turnover] at proliferation-promoting target gene promoters). These senescent cells with their senescence-associated stemness may resume proliferation (i.e., escape) out of senescence with particularly aggressive growth properties as “post-senescent” or “previously senescent” cells due to retained marks of senescence-associated epigenomic remodeling (top). In contrast, exposure of cells with an a priori senescence defect (as outlined in A and B, potentially resulting in immortalization, i.e., the ability to divide indefinitely) to a pro-senescent trigger will not lead to senescence and associated epigenomic remodeling, hence will bypass senescence and produce trigger-specific remodeled or even -transformed cellular conditions without a history in senescence (“never senescent”; bottom). Note that neither escape nor bypass cells appear to exhibit growth properties similar to their proliferating ancestors; especially escape cells are no mere senescence revertants but distinctly different from their pre-senescent counterparts. See main text for additional details and references.

Figure 2.

Transient senescence passaging during stress response journeys alters cell fate. Epigenetic remodeling and chromatin dynamics shape a Waddington-reminiscent landscape of senescence-associated stemness and plasticity. (A–C) Fate models refer to normal differentiation from a progenitor (A), iPSC-like pluripotent reprogramming (B), and direct conversion from one to another terminally differentiated cell type (C). (D) Stress response journeys (bold arrows) of senescence-incapable (immortal) cells that bypass (b) cellular senescence upon oncogenic stress exposure on their path to a (pre-)malignant state as compared with normal cells transiently encountering oncogenic stress-induced senescence and undergoing profound epigenetic changes including stem-like reprogramming and altered lineage commitment (i.e., phenotypic plasticity) before ultimately escaping (e) from the arrest as post-senescent cells. The profound senescence-related epigenetic changes determine a distinctly different cell fate, as depicted by the transdifferentiated cellular offspring now found in a different Waddington valley (adapted from Takahashi and Yamanaka, 2016; Waddington, 1957).