Past and Future Directions for Research on Cellular Senescence

Abstract

Cellular senescence was initially described in the early 1960s by Hayflick and Moorehead. They noted sustained cell-cycle arrest after repeated subculturing of human primary cells. Over half a century later, cellular senescence has become recognized as one of the fundamental pillars of aging. Developing senotherapeutics, interventions that selectively eliminate or target senescent cells, has emerged as a key focus in health research. In this article, we note major milestones in cellular senescence research, discuss current challenges, and point to future directions for this rapidly growing field.

The increase in numbers of older individuals (age > 65) to more than 20% of the population has created a global challenge in ensuring that health systems can meet the needs of older patients and foster healthier life in old age. Aging is a multifactorial and complex process that can begin even before conception (e.g., in Down syndrome [Meharena et al. 2022]), which over time leads to reduced resistance to multiple diseases, probability of survival, and capacity for self-regulation, repair, and adaptation to environmental insults (López-Otin et al. 2013; Cohen 2018). A deeper understanding of the fundamental mechanisms of aging processes is necessary to discover interventions that promote health span, the period during life free of disability, pain, cognitive impairment, and loss of independence.

The aging process is both characterized and driven by shared cellular and molecular mechanisms termed the pillars or hallmarks of aging. These include genomic instability, telomeric dysfunction, epigenetic alterations and other DNA changes, loss of proteostasis, dysregulated nutrient-sensing, mitochondrial dysfunction, progenitor (“stem”) cell dysfunction, altered intercellular communication, and cellular senescence, among others (López-Otin et al. 2023). These fundamental aging mechanisms progress at the molecular, cellular, tissue, and systemic levels and largely account for older age being the leading risk factor for multiple acute and chronic disorders and diseases. These include, among many others, cardiovascular diseases, cancers, osteoarthritis, diabetes, lung, liver, kidney, skin, and eye diseases, and neurodegenerative diseases such as Alzheimer's disease (AD) and Parkinson's disease (PD) (Niccoli and Partridge 2012; Hou et al. 2019). Given that aging is a shared underlying risk factor for multiple conditions, the geroscience hypothesis holds that manipulating fundamental aging mechanisms could delay, prevent, alleviate, or treat multiple disorders and diseases.

Cellular senescence, a cell fate that entails essentially stable growth arrest and resistance to apoptosis in response to cellular damage and stress, has been implicated as a mechanism contributing to development of many diseases and age-related conditions (Tchkonia et al. 2013). Senescent cells can develop at any point during life, tend to accumulate with aging (particularly in the context of dysfunction), and can appear at sites of pathology in multiple chronic as well as acute diseases, even in younger individuals. It appears that most other fundamental aging processes can induce cellular senescence, suggesting it might be widely involved in the genesis of aging phenotypes and multiple disorders and diseases. Consequently, targeting cellular senescence may provide protection against many pathologies and perhaps even attenuate or delay declines in health span in older individuals or younger people with cellular senescence-related disorders.

THE DISCOVERY OF CELLULAR SENESCENCE

In the early 1960s, Leonard Hayflick and Paul Moorhead reported that primary human fibroblasts isolated from embryonic lung tissue fail to further proliferate after a limited number of population doublings and thereafter exhibit a characteristic flattened and enlarged morphology (Hayflick and Moorhead 1961). This phenomenon of a maximum proliferative cellular life span in noncancerous cells is recognized as the Hayflick limit, and the replicatively exhausted cells were termed senescent, from Latin “senex” meaning “old.” Since then, the original definition of cellular senescence has expanded and become relevant in multiple other contexts, including aging in vivo and diseases.

INDUCTION OF CELLULAR SENESCENCE

Cellular senescence can be triggered in response to various intrinsic and extrinsic stimuli as well as developmental signals (Kumari and Jat 2021). The most frequent senescence inducers are replicative exhaustion, ionizing and nonionizing radiation, oncogene activation, genotoxic drugs, oxidative stress, metabolic dysfunction, demethylating and acetylating agents, mechanical or shear stresses, infections, and inflammation (Hernandez-Segura et al. 2018). Senescence is a highly dynamic and orchestrated process during which the properties of senescent cells continuously change and vary depending on their microenvironmental context (Kumari and Jat 2021). The different types of inducing stimuli prompt different types of senescence, such as telomere-dependent replicative senescence, developmentally programmed senescence, or non-telomeric stress-induced premature senescence, including oncogene-induced senescence (OIS), therapy-induced senescence (TIS), unresolved DNA damage-induced senescence, epigenetically induced senescence, and mitochondrial dysfunction-associated senescence (Kumari and Jat 2021).

The cell-cycle exit during development of senescence occurs at defined cell-cycle checkpoints. Briefly considering these checkpoints provides context for use of several senescence “markers,” including p16^INK4a, p21^CIP1/WAF1, cyclin-dependent kinase inhibitor 1A (P21, Cip1), and p53 (Blagosklonny and Pardee 2002). The first is the G₁/S checkpoint, also known as the restriction point. Retinoblastoma protein, Rb, is a primary regulator of the G₁ phase of the cell cycle and an important tumor suppressor. The active form of Rb is the un- or hypophosphorylated form. During the normal progression of G₁, the CDK4/cycD complex catalyzes the initial monophosphorylation of Rb and then CDK2/cycE hyperphosphorylates and inactivates Rb to release the transcription factor, E2F. E2F is itself a transcription factor for cycE. Hence, phosphorylation of Rb reinforces its inactivation through a CDK2/cycE-mediated hyperphosphorylation feedback loop. E2F then activates the transcription of genes facilitating S phase progression.

Alternatively, when a cell does not pass the restriction point, p16^INK4A prevents the inactivation of Rb by inhibiting CDK4/cycD and preventing the initial Rb monophosphorylation. Active Rb recruits HDAC, a histone deacetylase, which causes chromatin compaction and thereby blocks binding sites that would have been activated by the E2F transcription factors, resulting in continued repression of E2F-controlled genes and cell-cycle arrest. This can be further reinforced and stabilized in response to DNA damage by p53-mediated activation of p21^CIP1/WAF1, which inhibits the G to S phase transition by inhibiting CDK2/cycE. Under sustained stress, such as persistent DNA damage signaling or hypoxia, this temporary cell-cycle arrest can turn into permanent senescence.

The next major checkpoint is the G₂/M DNA damage checkpoint. The critical target of this checkpoint is CDK1/cycB, which can be inhibited in response to DNA damage by p21^CIP1/WAF1 through p53-dependent p21^CIP1/WAF1 up-regulation. Normally, CDK1/cycB activates Cdc25. Cdc25 is a phosphatase that dephosphorylates and thereby deactivates CDK1/cycB inhibitors such as Wee1. But in the event of DNA damage, p21^CIP1/WAF1 inhibition of CDK1/cycB ensures Cdc25 is not activated and Wee1 is not inhibited, further stabilizing the inhibition of CDK1/cycB by p21^CIP1/WAF1. In this way, p53 and p21^CIP1/WAF1 regulate both the G₂/M DNA damage checkpoint and the G₁/S checkpoint.

HALLMARKS OF SENESCENCE

Cellular senescence is associated with multiple phenotypic and molecular changes that can help to distinguish senescent cells from quiescent or terminally differentiated cells. In the mid-1990s, an increased lysosomal β-galactosidase activity (SA-βgal) at pH 6.0 was reported by Dimri et al. (1995) as a biomarker of cellular senescence both in situ and in vitro. Nowadays, either chemogenic or fluorescent probes can be used to detect SA-βgal by staining or flow cytometry. However, increased SA-βgal activity is neither fully sensitive (not all types of senescent cells express it) nor specific to senescent cells. It can be detected in macrophages and multiple types of postmitotic cells, such as neurons (de Mera-Rodríguez et al. 2021). Despite this, due to its technical simplicity, SA-βgal remains a commonly used assay for estimating senescent cell abundance in vivo and in vitro. Senescent cells not only have increased SA-βgal but also increased size and number of lysosomes, which result in the enhanced granularity that is sometimes used as another morphologic feature for identifying senescent cells (Robbins et al. 1970; Gorgoulis et al. 2019).

Major alterations occur not only in lysosomes but the nucleus as well. Heterochromatin reorganization occurs during progression of senescence. The heterochromatin associated with this process is most prominent among domains of facultative heterochromatin that often form in senescent human cells, termed senescence-associated heterochromatin foci (SAHF) (Narita et al. 2003). DNA segments with chromatin alterations reinforcing senescence (DNA-SCARS) are other markers indicative of senescence (Rodier et al. 2011). DNA-SCARS are dynamically formed, unique chromatin structures that functionally regulate multiple aspects of the senescent phenotype.

Dysfunctional telomeres, which are ordinarily protective structures at the ends of linear DNA comprised of specific repeated DNA elements, is another feature of senescence. Cell stresses such as hypoxia or DNA damage trigger a DNA damage response (DDR). The activation of the DDR at telomeres results in the formation of telomere-associated DDR foci (TAFs) or telomere-induced DNA damage foci (TIFs), which are sites of DNA damage that normal DNA repair mechanisms cannot repair. Thereafter, persistent DNA damage signaling initiates the senescence process. Cellular TAFs and TIFs, which are among the more sensitive and specific markers for senescence, can be quantified in cell cultures and tissues.

As discussed previously, another feature of cellular senescence is certain process-associated pathways. In response to cellular stress, diverse effector programs, including those related to cell cycling, DNA repair, and cell death, are activated to mitigate cellular and tissue damage. If the stress or damage cannot be resolved, the cellular senescence program can be initiated and regulated by the p16^INK4a/Rb and/or p53/p21^CIP1/WAF1 tumor suppressor pathways. They simultaneously or independently maintain the senescent state by inducing widespread changes in gene expression, since p53 and pRb are master transcriptional regulators. The p53/p21^CIP1/WAF1 and the p16^INK4a/Rb pathways interact with and influence each other through multiple cross-talk mechanisms. p21^WAF1/CIP1 acts downstream of p53, whereas p16^INK4a acts upstream of pRb. They are crucial cyclin-dependent kinase inhibitors (CDKIs) and act as negative regulators of cell-cycle progression. Thus, p16^INK4a and/or p21^CIP1/WAF1 are frequently, but not always, increased in senescent cells.

Cytoplasmic DNA, such as mtDNA released from stressed mitochondria or cytoplasmic chromatin fragments (CCFs) released from damaged nuclei, trigger inflammation response in senescent cells (Miller et al. 2021). Recently, the activation of endogenous retrotransposable elements, line 1 elements, and the ensuing activation of an IFN-I response, have been reported in late-stage senescent cells (De Cecco et al. 2019; Gorbunova et al. 2021).

THE SENESCENCE-ASSOCIATED SECRETORY PHENOTYPE

A functional output of senescent cells is the senescence-associated secretory phenotype, or SASP, which encompasses a variety of secreted bioactive factors that can include cytokines, chemokines, growth factors, proteases, profibrotic factors, prothrombotic factors, lipids and other metabolites and bioactive small molecules, miRNAs, other noncoding nucleotides, and extracellular vesicles (Zhu et al. 2014; Basisty et al. 2020). The SASP was initially characterized using antibody arrays in human fibroblasts and epithelial cells (Coppé et al. 2008). Given the complexity of senescence induction and differences in the cell types that become senescent, senescent cell functional output is highly diverse and context-dependent. Many stress response pathways are involved in establishing and maintaining the SASP. These pathways can also be interlinked, leading to prolonged activation, such as in the cases of the p38MAPK (Freund et al. 2011), mTOR/nutrient sensing (Laberge et al. 2021), JAK/STAT (Xu et al. 2016; Kandhaya-Pillai et al. 2022), NF-κB (Salminen et al. 2012), and cGAS-STING (Yang et al. 2017) pathways.

The effects of the SASP can be pleiotropic and complex. The SASP can be tissue damaging and proinflammatory or can promote tissue regeneration, and the extent of expression of different elements of the SASP depends on the type of cell that became senescent, the inducer of senescence, how long the cell has been senescent, and the microenvironment (Tripathi et al. 2021a). For example, SARS-CoV-2 S-antigen can accentuate release of tissue-damaging SASP factors (Camell et al. 2021; Tripathi et al. 2021b; Kandhaya-Pillai et al. 2022; Schmitt et al. 2023). The SASP appears to be important for many biological processes, such as embryogenesis, immune surveillance, and, in some cases, wound repair. Wound or tissue repair occurs highly dynamically and throughout life, and it involves extensive cellular communication orchestrated by cytokines, chemokines, growth factors, and components of the extracellular milieu. The SASP might sometimes be needed for new tissue formation and tissue remodeling during the wound-healing process, although there are situations in which generation of senescent cells actually impedes acute wound repair (Dańczak-Pazdrowska et al. 2023) (e.g., in muscle [Moiseeva et al. 2023] or chronic skin wounds [Wyles et al. 2023]). During embryonic development, the SASP recruits immune cells to fine-tune the embryogenesis process and enhances plasticity and stemness in surrounding cells to promote tissue regeneration (Muñoz-Espin and Serrano 2014). In addition, senescent cells can play a role in immune surveillance. They can attract, anchor, and activate immune cells at sites of damage through the SASP and act as immunogenic targets for immune clearance (Prata et al. 2018).

The SASP is an example of antagonistic pleiotropy, the evolutionary hypothesis that a trait advantageous early in life, when natural selection pressure is highest, can be selected for despite deleterious consequences of the same trait in old age. Despite the beneficial physiological functions of the SASP, persistent senescence or unresolvable senescent cell accumulation can be detrimental. Given the detrimental effects of accumulated senescent cells and the SASP (in those cases when it is tissue-damaging), senescent cells are likely causally implicated in multiple phenotypes and pathologies such as osteoarthritis-like phenotypes, frailty, impaired metabolic function, and shortened health span. In the past few years, there have been extensive studies to elucidate the role of senescence in the onset of multiple pathologies and diseases. For instance, despite cellular senescence being initiated as a protective mechanism against cancer, senescence markers have been observed in numerous human cancers. They are linked to an increased risk of recurrence and poor survival outcomes. The SASP was identified as a driver of cell growth, promoting tumorigenesis in the tumor microenvironment through a paracrine route (Gonzalez-Meljem et al. 2018)—the proteases and matrix-modifying enzymes of the SASP can alter the microenvironment, resulting in tumor progression (Coppé et al. 2010; Liu et al. 2022). Furthermore, certain cytokines released by senescent cells or that are produced by immune cells as a consequence of the presence of senescent cells such as IL-17, IL-21, and IL-23, appear to further modulate the immune response, especially adaptive T-cell subtypes, and predispose to cancers, cardiac dysfunction, neurodegeneration, infection, and frailty. In addition to the SASP factors directly released by senescent cells, tissue-damaging or disease-promoting factors such as these that are released by immune or other cell types in response to senescent cells could serve as useful gerodiagnostics (see below).

Extracellular matrix (ECM) production can be another effect of the SASP. The ECM is crucial in maintaining normal tissue structure and cell-to-cell communication, and many pathological conditions arise from dysregulated ECM remodeling because of aging or damage (Coppé et al. 2010; Blokland et al. 2020). The SASP can drive pathological increases in ECM deposition and remodeling, resulting in tissue fibrosis (Blokland et al. 2020). Additionally, the SASP can impair the functionality of progenitor cells by altering the ECM, which progenitor cells depend upon for maintaining self-renewal capability and capacity for differentiating into specialized cell types. SASP proinflammatory factors and growth factors can cause progenitor cells to lose multipotency (or “stemness”) by causing sustained mitotic stimulation.

Multiple diseases and disorders involve senescent cells, and the abundance of senescent cells correlates with the onset of pathologies, including idiopathic pulmonary fibrosis (IPF), neurodegenerative diseases, liver diseases, kidney-related diseases, and cardiovascular diseases. The high burden of senescent cells in multiple metabolic and endocrine tissues was found to be associated with disease progression and dysregulated glucose homeostasis in type 2 diabetes. Removal of accumulated senescent cells increased health span in murine models, suggesting that the SASP can play a deleterious role in age-related multimorbidity. The SASP can disrupt tissue homeostasis by promoting chronic inflammation, fibrosis, and progenitor cell dysfunction. Chronic inflammation is associated with the dysfunction that can develop with aging, and the SASP can directly initiate chronic inflammation in surrounding tissues or indirectly induce secondary inflammation through chronic activation of immune cells and spread of senescence locally and systemically (Prata et al. 2018; Xu et al. 2018; Schmitt et al. 2023). Emerging evidence suggests that proinflammatory SASP factors can be amplified, potentially contributing to the high susceptibility to hyperinflammation frequently observed in the elderly in the course of multiple pathological conditions, such as infectious diseases (e.g., COVID-19 [Camell et al. 2021; Tripathi et al. 2021b]). Collectively, this evidence highlights that targeting senescent cells and/or the SASP using “senotherapeutics” might be a strategy for delaying, preventing, alleviating, or treating multiple disorders and diseases across the life span (Fig. 1).

Figure 1.

Features of cellular senescence. Various stimuli such as DNA damage, oncogenes, mitogens, reactive metabilites, proteotoxic stress, inflammation, and DAMPs (damage associated molecular patterns) can induce cellular senescence across the life span. Senescent cells and their senescence-associated secretory phenotype (SASP) exert both physiological and pathological roles in age-related dysfunction and multiple diseases. (Figure created with BioRender.com.)

THE GROWING FIELD OF SENOLYTICS

Given that cellular senescence is a stable cell growth arrest that often entails enhanced expression of p21^CIP1/WAF1, p16^INK4a, both (or sometimes neither), a genetic senescent cell ablation method of fusing the p16^INK4A or p21^CIP1/WAF1 promoter region with apoptosis machinery was used to target senescent cells in tissues (Baker et al. 2011; Demaria et al. 2014; Wang et al. 2022). As those senescent cells with a proapoptotic SASP rely on one or more senescent cell antiapoptotic pathways (SCAPs) to evade cell death, a hypothesis-driven and mechanism-based drug discovery approach was used to seek small molecules that selectively induce apoptosis in the senescent cells that are tissue-damaging by deactivating these SCAPs. Based on this approach, the first senolytic drugs discovered, dasatinib (D) and quercetin (Q), target key nodes of the SCAP network, including ephrins/dependence receptors, PI3Kδ/Akt/metabolic factors, and Bcl-xL; p21^CIP1/WAF1/serpines (PAI-1/PAI-2). The combination of D + Q has been used to test the effectiveness of senolytic treatment across multiple conditions (Zhu et al. 2015; Wissler Gerdes et al. 2020).

Following the initial discovery via RNA interference approaches that the BCL-2 family is a prosurvival pathway in certain types of senescent cells (Zhu et al. 2015), the Bcl-2/Bcl-xL inhibitors navitoclax, A-1331852, and A-1155463 were confirmed to be senolytic in many but not all senescent cell types (Zhu et al. 2016, 2017). However, off-target effects of Bcl-2 family inhibitors on platelets and neutrophils limit their translational potential. Later, multiple senolytic compounds targeting different SCAPs were developed, including HSP90 inhibitors (Fuhrmann-Stroissnigg et al. 2017), FOXO4-p53 inhibitors (Baar et al. 2017), and Na⁺/K⁺-ATPase inhibitors (Triana-Martínez et al. 2019). In addition, natural flavonoids and related compounds such as fisetin, piperlongumine, curcumin, and procyanidin C1 showed certain senolytic activities, although their exact molecular mechanisms of action are not yet fully clear in all cases (Wang et al. 2016; Zhu et al. 2016; Yousefzadeh et al. 2018; Li et al. 2019; Xu et al. 2021).

Another approach for eliminating senescent cells is to use proteolysis-targeting chimeras (PROTACs): fused molecules comprising a ligand specific to a target protein and a ligand capable of recruiting an E3 ubiquitin ligase (Sakamoto et al. 2001). PROTACs can recognize and degrade targeted SCAP proteins through the ubiquitin–proteasome system, which induces cell death in the senescent cells that rely on those particular SCAPs. Despite the potential to reduce off-target effects using PROTACs, they are large molecules comprising fused ligands and further pharmacokinetic and pharmacodynamic studies will be needed to enable efficient delivery in vivo (He et al. 2020).

Senomorphics, also known as senostatics, reduce SASP factor release without inducing senescent cell death. Given the complexity of senescence induction, the SASP can be blunted by interfering in one or multiple SASP regulatory pathways, such as p38-MAPK inhibitors, NF-κB inhibitors, mTOR inhibitors, and JAK/STAT3 inhibitors. Alternately, senomorphic effects may be achieved through specific neutralizing antibodies against individual SASP factors, such as IL-1α, IL-8, Activin A, or IL-6 (Zhang et al. 2023). Senomorphics may not only attenuate release of tissue-damaging SASP factors, but also delay or prevent senescence induction due to the spread of senescence caused by certain SASP factors. Maintaining the effects of senomorphics can entail a need for more continuous administration than senolytics to achieve SASP-inhibiting blood levels, which might be problematic due to the lack of specificity for senescent cells of some senomorphic drugs (Wissler Gerdes et al. 2020; Zhang et al. 2023).

Recently, other novel senotherapeutic strategies have been devised based on the modulating immune system. Surveillance by senescent cells can activate immune cell responses and recruit particular classes of immune cells (Kale et al. 2020; Chen et al. 2023), such as macrophages, natural killer (NK) cells, or cytotoxic T cells. Aging or tissue damage can impede immune cell function through a separate, non-cell-cycle-based process, immunosenescence, partly explaining the accumulation of senescent cells in tissues with aging. Developing immunotherapies targeting senescent cells might be a tractable senotherapeutic strategy. For example, the chimeric antigen receptor T (CART)-cell therapy approach was developed to specifically target senescent cells through a senescent cell-specific surface marker, urokinase-type plasminogen activator receptor (uPAR), and may be of benefit for treating liver adenocarcinoma and liver fibrosis in preclinical models (Amor et al. 2020). DPP4 (dipeptidyl-peptidase 4) is another potential senescent cell-surface target expressed by certain types of senescent cells (Kim et al. 2017). Anti-DPP4 antibodies have been used to recognize DPP4-positive senescent cells and lead to their preferential elimination by NK cells (Kim et al. 2017). Other immunotherapies were subsequently developed based on senescent cell-surface markers, including B2M, CD9 receptor, NOTCH receptors (Yoshioka et al. 2021), and others (Rossi and Abdelmohsen 2021). Vaccination is another potential approach: mice were immunized against the antigen, glycoprotein nonmetastatic melanoma protein B (GPNMB), a senescent cell-surface marker (Suda et al. 2022). Although immune clearance of senescent cells has potentially promising outcomes, a deeper understanding of the interplay between senescent cells and the immune system is critical, warranting more studies to achieve insight relevant for designing immune cell–based interventions to treat age-related dysfunction and chronic diseases. Also, methods need to be developed for stopping continued senescent cell removal due to immune therapy–mediated interventions should senescent cells be needed (e.g., during pregnancy for placental mediation of parturition).

HUMAN STUDIES

The first in situ evidence that senescent cells present and accumulate with age in human skin tissues was reported in 2009 by taking advantage of increased SA-βgal activity (Dimri et al. 1995). Since then, many senescence hallmarks have been used to identify senescent cells across human tissues and organs, not only during the aging process but also related to multiple diseases (Tuttle et al. 2021). Indeed, senescent cell accumulation with aging in healthy individuals is sometimes slight: those older individuals with frailty and multimorbidity may have a higher senescent cell burden than healthy older individuals. Greater expression of the senescence markers, γH2AX, p16^INK4A, p21^CIP1/WAF1, and p53, was observed in glial cells, progenitor cells, and damaged neurons with tau tangles in humans with neurodegenerative diseases than those without (Jurk et al. 2012; Shanbhag et al. 2019; Vazquez-Villasenor et al. 2020). These markers were also present in the context of eye pathologies, and some, not all, of these markers are associated with specific stages of eye diseases, including endothelial corneal dystrophy, glaucoma, cataracts, and retinal microaneurysms (Lee et al. 2021a; López-Otin et al. 2023). p16^INK4A was increased in the human livers with fibrosis, and p16^INK4A correlated with histological stages of fibrosis (Wandrer et al. 2018). Other senescent cell markers, including telomeric dysfunction (Amsellem et al. 2011), SASP regulators (Houssaini et al. 2018), and Lamin B1 (Saito et al. 2019), have been used to track senescent cell abundance in lung tissues from patients with chronic obstructive pulmonary disease (COPD). Recently, patients with SARS-CoV-2 infection were shown to have senescence markers in their airway mucosa and increased circulating SASP factors (Lee et al. 2021b). Despite extensive evidence for increased senescent cell burden in human tissues in multiple disorders and diseases, the use of robust and universal senescence markers to precisely characterize the dynamics and function of senescent cells in situ remains challenging due to the heterogeneous nature of senescent cells across tissues and disease states.

To extend successful preclinical findings to humans, more than 20 clinical trials of senolytic therapies have been initiated and some of them have been completed. Although much is known about side effects of the senolytic agents in these trials from experience during their use over many years for treating other conditions or their being natural products, much remains to be learned about them in the context of being administered as senolytics, especially potential long-term side effects. To explore their efficiency and effectiveness, pilot clinical trials have proceeded in patients with serious conditions, such as diabetic kidney disease, AD, advanced frailty, and IPF. The results of the first-in-human, single-arm, open-label clinical trial of senolytics (D + Q) were reported in 2019 (Justice et al. 2019). Fourteen subjects with IPF, a senescence-associated disease, showed improved physical performance 5 days after nine doses of D + Q over 3 weeks (Cope 1986; Justice et al. 2019). However, as this study was not placebo-controlled and was small, a phase 2 trial is necessary before firm conclusions can be drawn. Intriguingly, a post hoc analysis of a continuation of this study involving 20 subjects showed increased urine levels of the geroprotective factor α-Klotho after compared to before senolytic treatment (Zhu et al. 2022).

The first evidence for senescent cell clearance by senolytics in humans was reported in an open-label phase 1 pilot study in which nine subjects with diabetic kidney disease were recruited and administered 3 days of D + Q (Hickson et al. 2019). A significant reduction of p16^INK4A and p21^CIP1/WAF1 highly expressing senescent cells and SA-βgal-positive cells was observed in adipose tissue. Circulating SASP factors, including IL-1α, IL-6, and matrix metalloproteinases (MMPs), were decreased after compared to before senolytic administration. Given encouraging preclinical results of senolytics in AD mouse models, an open-label pilot trial (SToMP-AD) tested safety, feasibility, and efficacy of D + Q in five subjects with early-stage AD (Gonzales et al. 2022). This pilot study was completed recently, and preliminary results are about to be published. Not all initial studies of senolytics have been successful: a pilot human trial by a biotech company reported that treatment with an MDM2 inhibitor-based senolytic compound, UBX101, in patients with osteoarthritis failed to meet its primary end point. It needs to be anticipated that many trials will fail, adding to the imperative to view current trials as initial steps along a potentially long path, particularly before senolytics are considered for preventing conditions in relatively healthy individuals. The pilot clinical trials have provided valuable data about the feasibility of testing senolytics in future larger randomized controlled trials for serious senescence-related diseases. Growing numbers of clinical studies are underway or planned to examine the effectiveness of senolytics for senescence-associated conditions, including COVID-19, age-related skeletal health, frailty in cancer survivors, diabetic macular edema, neovascular age-related macular degeneration, and others (Table 1).

Table 1.

Examples of ongoing clinical trials of senolytics

ClinicalTrials.gov number	Trial title	Current status	Diseases	Senolytics	Phase
NCT04785300	ALSENLITE: senolytics for Alzheimer's disease	Enrolling by invitation	Mild cognitive impairment; Alzheimer's disease	Dasatinib + quercetin	Phases 1 and 2 (PMID:34366147)
NCT04733534	An open-label intervention trial to reduce senescence and improve frailty in adult survivors of childhood cancer	Recruiting	Frailty; childhood cancer	Dasatinib + quercetin; fisetin	Phase 2
NCT04537299	COVID-FIS: pilot in COVID-19 (SARS-CoV-2) of fisetin in older adults in nursing homes	Enrolling by invitation	Covid19; SARS-CoV infection	Fisetin	Phase 2 (PMID:34375437)
NCT04476953	COVID-FISETIN: pilot in SARS-CoV-2 of fisetin to alleviate dysfunction and inflammation	Enrolling by invitation	Covid19	Fisetin	Phase 2
NCT04771611	COVFIS-HOME: COVID-19 pilot study of fisetin to alleviate dysfunction and decrease complications	Completed	Covid19	Fisetin	Phase 2
NCT04313634	Targeting cellular senescence with senolytics to improve skeletal health in older humans	Recruiting	Healthy	Dasatinib + quercetin; fisetin	Phase 2
NCT04063124	Senolytic therapy to modulate progression of Alzheimer's disease	Completed	Alzheimer's disease	Dasatinib + quercetin	Phases 1 and 2 (PMID:34687726)
NCT03675724	Alleviation by fisetin of frailty, inflammation, and related measures in older adults	Recruiting	Frail elderly syndrome	Fisetin	Phase 2
NCT03430037	Alleviation by fisetin of frailty, inflammation, and related measures in older women	Recruiting	Frail elderly syndrome	Fisetin	Phase 2
NCT03325322	Inflammation and stem cells in diabetic and chronic kidney disease	Enrolling by invitation	Chronic kidney diseases; diabetic nephropathies	Fisetin	Phase 2
NCT02874989	Targeting proinflammatory cells in idiopathic pulmonary fibrosis: a human trial	Completed	Idiopathic pulmonary fibrosis	Dasatinib + quercetin	Phase 1 (PMID:30616998)
NCT02848131	Senescence in chronic kidney disease	Enrolling by invitation	Chronic kidney disease	Dasatinib + quercetin	Phase 2 (PMID:31542391)
NCT02652052	Hematopoietic Stem Cell Transplant Survivors Study (HTSS)	Recruiting	Stem cell transplant	Dasatinib + quercetin	Not applicable
NCT04771611	COVFIS-HOME: COVID-19 pilot study of fisetin to alleviate dysfunction and decrease complications	Enrolling by invitation	Covid19; coronavirus infection	Fisetin	Phase 2
NCT04685590	Senolytic therapy to modulate the progression of Alzheimer's disease (SToMP-AD) study	Recruiting	Alzheimer's disease; early onset of mild cognitive impairment	Dasatinib + quercetin	Phase 2 (PMID:34366147; 34687726)
NCT04210986	Senolytic drugs attenuate osteoarthritis-related articular cartilage degeneration: a clinical trial	Active, not recruiting	Osteoarthritis	Dasatinib + quercetin	Phases 1 and 2
NCT05416515	A study of fisetin to treat carpal tunnel syndrome	Recruiting	Carpal tunnel syndrome	Fisetin	Phase 2
NCT05593588	Senolytics treatment of interstitial lung disease in common variable immunodeficiency	Not yet recruiting	Common variable immunodeficiency	Fisetin	Phase 2
NCT05595499	Treatment of frailty with fisetin (TROFFi) in breast cancer survivors	Recruiting	Breast cancer	Fisetin	Phase 2

TRANSLATIONAL GEROSCIENCE NETWORK

The Translational Geroscience Network (TGN) was established to accelerate testing the geroscience hypothesis that targeting fundamental aging mechanisms, including cellular senescence, can delay, prevent, alleviate, or treat multiple diseases and disabilities in humans. Envisioning a time- and cost-efficient approach to conducting geroscience-guided trials targeting the biology of aging, the national TGN collaborative was initiated by eight U.S. research institutes—Mayo Clinic, Harvard, John Hopkins, Wake Forest, Universities of Minnesota, Michigan, and Connecticut, and University of Texas Health Sciences Center at San Antonio—with additional collaborators at Northwestern University, St. Jude Children's Cancer Center, City of Hope, the Steadman Philippon Clinic, and others (Fig. 2).

Figure 2.

Translational Geroscience Network structure, partners, and governance. (UMich) University of Michigan, (UMN) University of Minnesota, (UCONN) University of Connecticut, (UTHSCSA) University of Texas Health Sciences Center at San Antonio.

The TGN provides expertise, infrastructure, and coordination needed for geroscience-based studies to be conducted in parallel and with comparable designs, as well as sharing data generated by these studies. It is hoped this will accelerate the discovery and translation of interventions targeting fundamental aging mechanisms into clinical application. Beyond fundamental scientific and regulatory support for the trials, the TGN also created both a Facility for Geroscience Analysis (FGA), which assays >150 factors in body fluids and other biospecimens across the trials, and a biobank to facilitate reverse translation from bedside to bench to fuel discovery science. The FGA is developing, standardizing, and implementing sensitive, reproducible, and reliable methods to measure candidate gerodiagnostics and composites of tests that assay all of the pillars of aging, predict which interventions to use and when, follow responses to geroscience interventions, and predict clinical improvement (Fig. 3). The TGN has a subcommittee that manages data acquisition, analysis, and development of predictive algorithms. The TGN may help to accelerate exploration of the translational potential of senotherapeutics for challenging diseases and disorders across the life span.

Figure 3.

Translational Geroscience Network (TGN) activities for facilitating geroscience-focused clinical trials.

CONCLUDING REMARKS AND FUTURE PERSPECTIVES

The rapidly growing field of cellular senescence has opened up many challenging yet unsettled questions. First, the most widely used senescence markers, SA-βgal and the cell-cycle inhibitors, p16^INK4A and p21^CIP1/WAF1, are neither fully sensitive nor specific for identifying senescent cells. Given that the dynamics of senescent cell occurrence and accumulation are cell origin-, causation-, stage-, and physiological context–dependent, it may be difficult to pinpoint a single universal marker for detecting all senescent cells. There is a need to identify and characterize different types of senescent cells across organs, various health conditions, and the life span. Delineating biological functions and consequences of different types of senescent cells will require precise spatial and temporal information about senescent cells and their changes in response to interventions in humans. Advances in single-cell technologies provide opportunities to profile features of DNA methylation, chromatin accessibility, mRNA, and protein expression at single-cell resolution. Integrative single-cell multiomics analysis holds potential for characterizing senescent cell subtypes, possibly offering new insights into their function and pathophysiological processes. Importantly, these technologies open an avenue for identifying novel factors that can potentially be used for developing new mouse models to study the biology of cellular senescence and identify additional senolytic targets. Application of artificial intelligence algorithms to integrative single-cell multiomics analysis may also be useful in identifying new interventions.

A threshold hypothesis about senescent cell accumulation has been proposed (Chaib et al. 2022). This posits that accumulation of whole-body senescent cell burden results in paracrine and endocrine spread of senescence that outpaces the capacity of the immune system to clear them, resulting in accumulation. This triggers and accelerates pathogenic changes. Importantly, accelerated accumulation of persistent senescent cells may predispose to multiple age-related phenotypes and diseases, perhaps contributing to age-related multimorbidity. Despite the detrimental roles of senescent cells, they also have necessary physiological functions. Carefully distinguishing physiological or “beneficial” from pathological accumulation of senescent cells and developing interventions specifically targeting detrimental cells could be of benefit.

Genome instability, inflammation, low NAD⁺, mitochondrial dysfunction, oxidative stress, protein aggregates, and lipotoxicity can induce cellular senescence. Conversely, senescent cells appear to causally contribute to and exacerbate many of other fundamental aging processes, such as inflammation, fibrosis, NAD⁺ depletion, decreased α-Klotho, accumulation of aggregated and misfolded proteins, progenitor cell dysfunction, adaptive immune cell dysfunction, and others. Hence, fundamental aging processes appear to be interlinked, supporting a “unitary theory of targeting fundamental aging mechanisms”: that targeting one fundamental aging mechanism may alleviate many or most of the others. This may explain in part why targeting a single fundamental aging mechanism, such as cellular senescence, can delay, prevent, or alleviate multiple age-related phenotypes as a group. If the unitary theory is correct, it will be critical to test whether combining individual interventions targeting different aging mechanisms results in less-than-additive, additive, or synergistic effects. More needs to be understood about nondividing, differentiated cell types, such as neurons, adipocytes, osteocytes, or cardiomyocytes, which can acquire a senescent-like state, and the SASP of such cells. Whether and how such cells contribute to tissue dysfunction and to what extent these cells should be eliminated is as yet unclear.

There has been initial success in very small, early human pilot trials of senolytics. However, the impact of senescent cell clearance needs much further study in double-blind, placebo-controlled clinical trials, including for repurposed agents or natural products with safety profiles already known in other contexts. Unless such trials show target engagement, minimal short- and long-term adverse events, and clinical utility, in our view these interventions should not be prescribed in routine clinical practice or used over-the-counter by the general public. Given that cellular senescence is a common contributing factor to multiple diseases and disorders, future clinical trials combining senolytics with disease-specific interventions hold additional potential for treating currently incurable diseases such as many neurodegenerative diseases and cancers.