Targeting Cell Senescence and Senolytics: Novel Interventions for Age-Related Endocrine Dysfunction.

Senescent cells have an enlarged and flattened morphology, frequently have high p53/p21 CIP1/WAF1 and/or p16 INK4a /Rb protein levels, and have DNA damage foci (particularly in telomeres) ( Fig. 1 ). Another biomarker for detecting senescent cells in culture or tissue samples is the lysosomal enzyme, senescence-associated-β-galactosidase (SA-βgal), with enzymic activity at pH 6.0 as opposed to lysosomal enzymes from nonsenescent cells ( 73 ). However, this marker is not highly sensitive or specific. A central feature of senescence is replicative arrest and lack of DNA replication. This can be detected using clonal plate dilution assays or by the absence of nucleoside analogue incorporation (eg, 5-bromodeoxyuridine or [3H] thymidine) ( 74 ). Immunostaining for proliferation markers, such as proliferating cell nuclear antigen and the marker of proliferation, Ki-67 (Ki-67), can also help in detecting senescent cells. Increased proinflammatory SASP factors, including interleukin-6 (IL-6), interleukin-1α (IL-1α), interleukin-8 (IL-8), monocyte chemoattractant protein 1, plasminogen-activated inhibitor 1 (PAI-1), plasminogen-activated inhibitor 2 (PAI-2), and matrix metalloproteinases (MMPs) can indicate increased senescent cell burden and can be analyzed in blood, tissues, or cells ( 64 ). However, not all senescent cells are proinflammatory and proapoptotic ( 63 , 75 ). Hence, there are many different senescence markers with varying degrees of specificity and sensitivity. As of today, there is no single marker for accurately measuring senescent cell accumulation, and the establishment of new cellular senescence markers or composite scores comprising key markers is a pressing issue. Ideally, these markers or scores should be detectible in noninvasively collected samples such as blood and urine. Senescent cell markers. Several senescence markers have been identified based on the molecular biology of senescent cells. Relatively new markers used in recent years are senescence-associated heterochromatin foci, telomere-associated foci, and senescence-associated distention of satellites. However, many of these senescence markers are nonspecific due to the heterogeneity of senescent cells, and to date, no single marker is a fully sensitive and specific indicator, especially for clinical use. Improvements have been made by developing composite scores or signatures comprising multiple markers. Refinement of indicators of senescent cell burden in vivo for clinical application, especially those that can be assayed noninvasively, is needed. Senescent cells generally have persistent DNA damage. Nuclear senescence-associated heterochromatin foci can be used to identify senescence induced by activated oncogenes such as H-RAS and BRAF and stressors that impede DNA replication ( 76 , 77 ). Increased activity of p38 mitogen-activated protein kinase (p38MAPK) or the γ phosphorylated form of the histone H2AX (γH2AX) reflects activated DNA damage responses, which along with depleted or irreparably damaged telomeres, can indicate cellular senescence. DNA damage response factors that persist at sites of damage and nuclear foci, which can be detected cytologically, can also serve as indicators of cellular senescence. Telomere-associated DNA damage foci (TAFs) that accumulate within telomeric sequences, as well as colocalization of γH2AX and p53-binding protein 1 (53BP1) with telomeres, are also indicators of the senescent state. Pericentromeric satellite heterochromatin undergoes decondensation in senescent cells, leading to senescence-associated distention of satellites (SADs). SADs appear earlier and more consistently than heterochromatin foci, reflecting an early and potentially key event in cellular senescence ( 76 , 78 ). Damage-associated molecular pattern factors, such as high-mobility group box 1 localization or molecules released by stressed cells undergoing cell death, such as mitochondrial DNA (mtDNA), reflect cellular damage and can indicate increased senescent cell abundance ( 79 ). The enzyme α-Klotho regulates multiple endocrine processes, such as insulin-like growth factor 1 (IGF-1) signaling ( 80 ), while also modulating mammalian target of rapamycin (mTOR) ( 81 , 82 ), cyclic adenosine monophosphate ( 83 ), p53/p21 CIP1/WAF1 ( 84 ), and Wnt protein levels ( 85 ). It is also involved in mineral metabolism, contributing to phosphate homeostasis. α-Klotho has come to light as a geroprotective factor that protects against physiological stresses such as oxidative damage and hypoxia. α-Klotho is also protective against the side effects of cytotoxic drugs. It is secreted by the distal renal tubule and so is present in urine. Importantly, α-Klotho is inversely and causally linked to senescent cell burden, and senolytics increase urinary α-Klotho in humans ( 86 ). This suggests that α-Klotho, along with other measures, could be a useful “gerodiagnostic” marker with respect to senescent cell abundance and other fundamental aging processes. Analysis of microRNAs (miRNAs) that may be specific to senescent cells ( 87 , 88 ), other nucleotides such as cell-free mtDNA ( 79 ), senescence-specific epigenetic profiles, and senescence-associated small extracellular vesicles (EVs) including exosomes ( 89 , 90 ), microsomes, or mitosomes could also be viable strategies for assessing senescent cell burden. Progress has been made with miRNAs, short (20-24 nt) noncoding RNAs that are involved in posttranscriptional regulation of gene expression. Several miRNAs that are differentially expressed with aging and by senescent cells have been reported ( 87 , 91 , 92 ). Since none of these markers can be used as a reliable senescence biomarker on its own, combinations of such indicators may be a more reliable reflection of senescent cell burden ( 74 , 93 , 94 ). Such composite scores could be useful for reflecting senescent cell abundance and following therapeutic efficacy in clinical trials of agents targeting senescent cells ( 95 , 96 ). Cell cycle arrest proteins (eg, p53, p21 CIP1/WAF1 and p16 INK4a ) can be measured in body fluids. Among these, p16 INK4a expression in peripheral blood T cells has been used to estimate senescent cell burden ( 97 , 98 ). Biopsy samples of, for example, adipose tissue or skin could be an alternative for assessing senescent cell burden, potentially along with imaging modalities.

Based on promising results in preclinical experiments in cultured human cells, human tissue explants, and animals, clinical studies are already underway ( 128 ). The first senolytic clinical trial published was an open-label pilot study in patients with idiopathic pulmonary fibrosis, conducted because idiopathic pulmonary fibrosis is a progressive and fatal disease with no highly effective treatment. Fourteen patients were treated with intermittent oral D + Q (100 mg/day of D and 1250 mg/day of Q) for a total of 9 doses: 3 days each week for 3 weeks in a brief, non–placebo-controlled, open-label study of safety and tolerability ( 305 ). The results suggested that senolytic treatment might alleviate physical dysfunction (gait speed, gait distance, chair stands, and short physical performance battery) in patients with idiopathic pulmonary fibrosis. However, the study was not placebo-controlled and phase 2 studies are needed. Post hoc analysis of another study involving 20 patients with idiopathic pulmonary fibrosis revealed higher levels of the geroprotective factor α-Klotho in urine after oral D + Q administration in each of the 20 participants ( 86 ). Another pilot study involving patients with diabetic kidney disease used, for the first time, a composite score developed for assaying senescent cell burden in humans. Nine individuals with diabetic kidney disease treated with a 3-day course of oral D + Q (100 mg/day of D and 1000 mg/day of Q) had decreased senescent cell burden in fat tissue compared to before the administration of senolytics ( 95 ). The fat tissue was biopsied 11 days after the last dose ( 95 ). There was also a decrease in circulating SASP factors 11 days after the last dose of D + Q compared to before D + Q was administered. Of note, D has a 3-hour and Q an 11-hour elimination half-life, so the agents were no longer present at the time of the second adipose biopsy and blood collection. This study indicated that the D + Q senolytic combination is effective in reducing senescent cell accumulation and associated inflammation in humans, suggesting that an intermittent, hit-and-run strategy may be a viable approach ( 95 ). After these promising results, more than 30 clinical trials of senolytic therapies for a variety of diseases are planned, ongoing, or completed, including phase 2 randomized, double-blinded, placebo-controlled trials ( 128 ). Other senolytics such as fisetin as well as D + Q will be used in upcoming or ongoing phase 2 trials. If positive, these trials will have to be followed by larger clinical trials examining the effects of senolytic drugs on senescence-related disorders and diseases. Although results from preclinical data appear promising with respect to certain endocrine and related diseases, there are limited human clinical study data at this point. This is because the senotherapeutic field is new and the clinical trials conducted so far have been limited in scale since they are pilot studies. Examples of clinical trials currently planned or underway for endocrine disorders, metabolic disorders, and related diseases are in Table 2 . Following the positive results from the pilot study, fisetin as well as D + Q will be used in phase 2 trials for diabetic kidney disease. One study will examine the effect of 20 mg/kg/day of fisetin on adipose tissue-derived mesenchymal stem/stromal cell function, kidney function, systemic inflammation, and physical function in individuals with advanced chronic kidney disease. Another study will evaluate whether targeting systemic senescent cell burden by 100 mg D + 1000 mg Q daily helps reduce markers of insulin resistance, inflammation, bone resorption, and physical dysfunction in older women with gait disturbance. Besides diabetic kidney disease, osteoporosis is a promising target for senolytic interventions based on preclinical findings, and a clinical trial is now underway to examine age-related osteoporosis. In this trial, markers of senescent cell burden alongside bone formation and resorption markers in older women are being measured. D + Q (D 100 mg/day and Q 1000 mg/day) or fisetin (20 mg/kg/day) for 3 consecutive days is being administered to older women. Type I collagen is being assayed as a serum bone metabolism marker after 20 weeks. The Alleviation by Fisetin of Frailty, Inflammation, and Related Measures (AFFIRM) study is currently being conducted to explore effects of senolytics on musculoskeletal aging and frailty. Fisetin (20 mg/kg/day) for 3 consecutive days is being administered, and serum inflammation markers, bone resorption, insulin resistance, and gait speed are being assayed as outcomes. In the planned “Safety and effectiveness of Quercetin & Dasatinib on epigenetic aging” study, 500 mg Q and 50 mg D for 3 days in a row per month, for a total duration of 6 months, will be administered to healthy individuals and epigenetic age will be assessed using blood samples. NAFLD, while not a primary endocrine disorder, is a complication of obesity and insulin resistance for which effective therapeutic agents are lacking. Senolytics may be effective, as suggested by analyses in murine models that indicated decreased hepatic steatosis after senolytic treatment ( 25 ). Neither NAFLD nor the related NASH have effective treatments that can reverse these disease processes. A planned double-blinded, randomized, proof-of-principle clinical study will examine the effect of D + Q on liver fibrosis in individuals with biopsy-proven NAFLD ( Table 2 ). In addition to senolytics, the TAME (Targeting Aging with Metformin) clinical trial is planned to test if metformin, a senomorphic drug, delays the appearance of a second age-related disease in patients without diabetes. Other clinical trials of senolytics and senomorphics are in Table 2 . Senolytics may also prove to be effective for cancer-associated conditions, potentially including endocrine cancers ( 199 , 200 ). As mentioned earlier, in some malignancies such as ACP, it is theorized that tumor cells can induce senescence in surrounding cells, which contributes to the malignant effects of the tumor due to SASP factor release ( 306 ). However, SASP factors can also contribute to tumor suppression by causing inflammation and recruiting immune cells that remove these damaged or oncogene-expressing cells ( 307 ). Senolytic drugs work very similarly to established anticancer treatments, and in some cases such as dasatinib and navitoclax, are repurposed anticancer drugs. There are 2 planned clinical studies using navitoclax as an adjunctive anticancer drug for cisplatin resistant/refractory ovarian cancer and metastatic castration-refractory prostate cancer. Navitoclax monotherapy has been shown to have poor activity against recurrent epithelial ovarian cancer, but there have been no unacceptable side effects. Another trial for prostate cancer was terminated early ( NCT01828476 ). Although further research is needed on this topic and these trials were not conducted with senolytic effects in mind, senolytic therapies may be an effective adjunctive cancer treatment after toxicity and long-term side effects have been examined. The senolytic treatments in many of the aforementioned clinical studies have to date had only mild side effects and off-target effects. To balance risk with potential benefits, so far clinical trials have focused on serious disorders and diseases for which few treatment options exist, rather than preventive studies in healthy individuals. In addition, there are 2 major issues to be solved before getting to the point where the clinical use of senotherapies can be advocated. First, even though some of the early clinical trials have suggested short-term safety and target engagement for senolytics, the field is still new. Larger, randomized clinical trials are needed for detecting any longer-term adverse effects and determining if these agents are effective. In our opinion, the only context for administering these agents is in carefully controlled clinical trials with data monitoring and safety oversight. Senolytics and other senotherapies should not be prescribed or used over the counter. There is a possibility that many of the early studies will fail. These trials should be viewed as a first step and a basis for continuous improvement of trial design so that as better treatments are developed, these trials can be used as a template. Several of the agents being administered in these early studies are approved for other uses in humans or are natural products with robust safety data. In our view, this needs to be the initial approach rather than beginning with new drugs rarely used in humans before. With these cautions, setbacks of the type that slowed clinical translation of gene therapies due to serious side effects in early trials might be avoided. The second issue is gerodiagnostic tests. To facilitate the development of interventions, further work in refining gerodiagnostic tests measuring senescent cell burden and the extent of other fundamental aging processes is required. Mere “aging clocks” will be less informative than signatures or composite scores of analytes that change in response to interventions, predict and track clinical changes, indicate which intervention to use, are reliable, reproducible, scalable, and inexpensive, and can be assayed in urine, saliva, blood, buccal swabs, or other easily obtained samples. There is not any single, universal, or fully specific biomarker for identifying senescent cells reliably in cell culture or tissue samples because the phenotype of senescent cells varies considerably among cell types and tissues. New approaches are being explored that take the initial cell type into account to subclassify senescent cells and focus on distinct senescent cell types. The initial trigger for senescence may determine the phenotype of senescent cells, an example being adipocyte senescence induced by high-fat diets, in which upregulation of p21 CIP1/WAF1 but not so much of p16 INK4a occurs ( 255 , 256 ). To characterize senescence fully and develop more sensitive and specific markers of different senescent cell subtypes, transcriptomic and proteomic studies down to the single-cell level across relevant cell and tissue types will be of importance, facilitating delineation of cell surface molecules that will allow detection and isolation of senescent cells from blood or tissue samples. Another solution to this problem is to combine sets of SASP factors and other senescence-associated analytes into a composite score to identify and quantify senescent cells. Several such panels have been developed, the first being one that successfully tracked decreases in senescent cell burden in a clinical trial of senolytics ( 95 ). Frequent components of these panels are p21 CIP1/WAF1 or p16 INK4a levels, along with inflammatory SASP factors such as TNFα or IL-1α. Another such composite score is the “SenMayo” panel, which appears to identify senescent cells across different tissues and species ( 96 ). This panel can be used at the tissue and single-cell level and identifies key signaling pathways. Panels such as SenMayo may facilitate monitoring senescent cell burden with aging and in diseases and for analyzing the effect of therapies targeting cellular senescence during clinical trials.

Accumulating evidence from in vitro and in vivo studies using the senescent markers mentioned earlier indicate that senescent cells accumulate in multiple endocrine organs with aging, especially in older individuals with impaired function, decreased physical resilience, and/or multimorbidity. Senescent cell accumulation may contribute to the onset and progression of several endocrine diseases ( Fig. 2 ). Senescent cell accumulation in endocrine tissues with aging. With increasing age, senescent cells can accumulate in tissues, as demonstrated by assays of p16 INK4a /p21 WAF1/CIP1 expression, increased senescence-associated β-galactosidase activity, or shortened telomeres. Another approach is to monitor levels of senescence-associated secretory phenotype factors. Senescent cells can disrupt tissue homeostasis in endocrine and other organs and dysregulate hormone production and target organ effects, contributing to worsening health outcomes. The prevalence of T2DM increases with age ( 99 , 100 ), together with the hallmarks of this condition such as increased insulin levels and peripheral insulin resistance. Adipose tissue is a metabolically dynamic organ that is the primary site of excess energy storage, but it also serves as an endocrine organ capable of synthesizing a number of biologically active compounds that regulate metabolic homeostasis, such as tumor necrosis factor-α (TNFα), IL-6, IL-8, leptin, adiponectin, angiotensin, resistin, and PAI-1 ( 101-103 ). Two decades ago, it was found that aging is linked to increased adipose tissue inflammation and decreased capacity of cloned adipose progenitors and mesenchymal stem cells for differentiation and replication. This decline in adipogenesis correlated with insulin resistance ( 104 ). These findings suggest that presenescent and senescent cells accumulate in adipose tissue with aging since impaired replication is a hallmark of cellular senescence. It was then found that insulin resistance correlates with increased markers of cellular senescence in fat tissue, including adipose tissue β-galactosidase, which is a senescence-linked marker of increased lysosomal activity, as well as increased levels of PAI-1, p53, and cyclin D kinase inhibitors, including p16 Ink4a ( 37 , 105 , 106 ). Excessive caloric intake leading to insulin resistance increases SA-βgal activity with increased p53 and p21 CIP1/WAF expression in adipose tissue compared to normal caloric intake ( 105 ). These changes were related to accumulation of reactive oxygen species (ROS), which can drive senescence or be a product of senescent cells ( 105 ). In adipose tissue–specific p53 knockout (KO) mice, insulin resistance caused by a high-fat diet was significantly attenuated, and senescence markers in adipose tissue were decreased even during increased caloric intake ( 105 ). These findings suggest that insulin resistance linked to obesity can be mediated by cellular senescence in adipose tissue. Complications of insulin resistance, including hepatic steatosis progressing to nonalcoholic fatty liver disease (NAFLD) and nonalcoholic sclerosing hepatitis (NASH), diabetic/high-fat diet–induced kidney disease, diabetic cardiovascular dysfunction, macular degeneration, and obesity-related neuropsychiatric dysfunction are associated with senescent cell accumulation in the affected organs ( 34 , 38-42 , 52 , 107 , 108 ). Examples of hormones affected by aging include insulin and IGF-1, which regulate metabolic balance, anabolic activity, and replication and differentiation of multiple types of cells ( 109 ). In 1997, it was reported that a mutation leading to decreased expression of daf-2 (Dauer formation-2), which encodes an insulin/IGF-1 receptor in the nematode Caenorhabditis elegans , results in a 2- to 3-fold increase in lifespan ( 110 , 111 ). This gene is conserved across species ranging from yeast to mammals ( 112 ). Other genes involved in insulin/IGF-1 signaling have also been shown to extend lifespan when suppressed. Examples include the transcription factor, daf-16 , a C elegans homologue of the forkhead box O ( FOXO ) gene in humans that is downstream in the IGF-1/Akt signaling pathway ( 113 , 114 ), and Sch9 , a gene homologous to protein kinase B ( Akt ) ( 115 , 116 ). In Drosophila , mutations leading to decreased expression of the insulin/IGF-1 receptor or in insulin receptor substrate (IRS) 2-like molecules have also been shown to be associated with prolonged lifespan ( 117 ). Increases in lifespan from reduced insulin signaling are conserved in more complex organisms. In mammals, in which insulin and IGF-1 have separate receptors, heterozygous IGF-1 KO mice live an average of 30% longer than controls ( 118 ) and mice lacking adipose tissue insulin receptors live 18% longer, as do mice lacking brain IRS2 ( 119 , 120 ). Mice with genetic GH deficiency or humans with isolated GH deficiency are protected from aging phenotypes and show longevity ( 121 ). The National Institute on Aging's Primate Aging Research Project found that caloric restriction extends lifespan in primates. Caloric restriction of rhesus macaques over a 20-year period resulted in animals that were biologically “younger,” with shinier hair and less likelihood to develop age-related diseases ( 122 ). These monkeys also had low circulating insulin ( 123 , 124 ). These findings suggest that insulin/IGF-1/Akt signaling might have a role in primate aging ( 122 ). In the Baltimore Longitudinal Study of Aging, which began in 1958, approximately 700 older men were followed for 25 years. Low insulin levels were associated with longevity ( 125 ), further suggesting that insulin/IGF-1/Akt signaling is involved in the development of age-related dysfunction and diseases ( 126 , 127 ). Consistent with our Unitary Theory of Fundamental Aging Mechanisms ( 128 , 129 ), age-related changes in IGF-1 and senescent cell abundance appear to be interlinked ( 130 ). The zinc metalloproteinase plasma protein-A (PAPP-A) increases IGF-1 bioavailability by cleaving insulin-like growth factor binding proteins (IGFBPs), particularly IGFBP-4 ( 131 ). IGFBPs bind IGF-1, hindering receptor activation. After cleavage of IGFBP-4 by PAPP-A, IGF is liberated from the binding protein and IGF signaling is initiated in the pericellular environment. Inhibiting PAPP-A activity by gene deletion, and thereby decreasing IGF-1 signaling, appears to extend lifespan by up to approximately 30% in naturally aging mice ( 132 ). This was also observed in PAPP-A KO mice on a high-fat diet, and even in already adult mice after PAPP-A gene expression was knocked-down ( 132-134 ). In the conditioned medium from senescent (senescence induced by etoposide) adult human primary preadipocytes, PAPP-A proteolytic activity was more than 12-fold higher compared to conditioned medium from nonsenescent cells, indicating that PAPP-A is an SASP component. Proteolytically active PAPP-A was also abundant on the surface of EVs secreted by senescent preadipocytes. These findings link increased IGF-1 activity to cellular senescence ( 130 ). Perhaps paradoxically, despite increases in lifespan related to low IGF-1 levels, low levels of circulating IGF-1 in plasma have also been linked to low muscle mass, a key component of frailty ( 135 ), although this has not been found consistently and requires further study. Frailty indices include grip strength, gait speed, or other indirect measurements of muscle mass, strength, and function. IGF-1, along with its systemic anabolic effects, directly enhances muscle protein synthesis, driving hypertrophy ( 136 ), as well as increasing the abundance of the satellite cells that surround and support skeletal muscle ( 137 ). IGF-1, through Akt, increases mTOR activity, which in turn increases protein synthesis ( 138 ). It also decreases FOXO levels and thus protein breakdown ( 139 ). In a 1990 study of men older than 60 years with low IGF-1 levels, exogenous GH administration increased bone density and lean muscle mass while decreasing adipose tissue mass ( 140 ). However, studies with recombinant human GH and GH secretagogues failed to demonstrate benefits that outweigh risks such as increases in insulin resistance linked to weight gain. Resistance training may decrease frailty and improve health in older patients ( 141-144 ) and stimulates the hypothalamic-pituitary GH–IGF-1 axis through stimuli from muscles ( 145 ). Thus, relationships among IGF-1, muscle function, frailty, health span, and lifespan are complex. With increases in the number of older people, the prevalence of osteoporosis has been increasing. Reduction in bone quality brings with it reduced quality of life, mainly due to pathological fractures ( 146 , 147 ). The pathophysiology of osteoporosis is multifactorial, and vitamin D deficiency is a risk factor. With age, the capacity of the skin to synthesize vitamin D decreases ( 148 ) and senescent cells accumulate ( 149 ), but whether this association is causal remains to be determined. Detrimental effects of senescent cell accumulation in bone have recently come to light. The senescence biomarkers, p16 INK4a and p21 CIP1/WAF1 were elevated in iliac crest needle biopsies from older postmenopausal compared to younger premenopausal women. Despite the heterogeneous nature of the biopsy samples, including variations in cellular composition and fat abundance, SASP factors were also increased in biopsies from the older participants ( 78 ). Accumulation of senescent cells in the bone microenvironment is linked to increased bone resorption by osteoclasts and reduced bone formation by osteoblasts, leading to reduced bone density ( 150 ). The senescent osteocytes that accumulate with aging in mice have increased expression of multiple SASP factors compared to young mice, as well as age-associated upregulation of SASP factor production in bone marrow myeloid cells. These findings suggest that senescent osteocytes and their SASP may contribute to age-related bone loss and that their removal may be a therapeutic strategy for age-related (as opposed to postmenopausal) osteoporosis ( 151 ). The hypothalamus and the pituitary gland are key regulatory organs of the endocrine system, and this regulatory function can become disrupted with aging. Age-related loss of hypothalamic regulation might be linked to SASP factors, as hypothalamic pro-opiomelanocortin neurons were increased by rapamycin, a known SASP inhibitor ( 152 ). In rodents, the hypothalamus becomes less sensitive to several feedback processes with aging ( 153 , 154 ). Hypothalamic arcuate nucleus GH-releasing hormone secretion is decreased in older individuals while paraventricular nucleus somatostatin secretion is increased, contributing to decreased GH levels related to reduced frequency and lower amplitude of secretory pulses ( 155-157 ). In a recent study, age-related changes were detected in the hypothalamic expression of the SASP factors IL-6, IL-1β, TIMP metallopeptidase inhibitor 1 (Timp1), Mmp12, Cxcl1, and Cxcl2. However, the senescent cell markers p16 Ink4a and p21 Cip1/Waf1 were not increased ( 64 ). This needs to be explored further at the single-cell level in the hypothalamus because in some tissues, for example, muscle, it was recently discovered that a very small, difficult to detect compartment of senescent cells contributes to substantial dysfunction ( 158 ). Deleterious effects of this small senescent cell fraction on muscle function were dramatically alleviated by senolytics. The diurnal rhythmic release of melatonin is regulated, in part, by a circadian clock located in the suprachiasmatic nucleus of the hypothalamus. This release can become dysregulated with increasing age, leading to decreased levels of melatonin and poor sleep quality ( 159 ). In mice, disruption of circadian clock genes such as Bmal1 or Clock reduced lifespan ( 160-162 ) and accelerated development of age-related phenotypes. Interestingly, poor sleep quality may be correlated with increased senescent cell burden, evidenced in humans by increased numbers of circulating senescent T cells ( 163 ) and increased p16 INK4a gene expressing peripheral blood mononuclear cells ( 164 ). In addition, senescent cells can accumulate in the aorta after induced sleep fragmentation ( 165 ). More work needs to be conducted to evaluate the connection between sleep disturbances and senescent cells, perhaps focusing on effects of hypothalamic senescent cell burden. The pituitary gland is prone to adenoma formation with aging, with an estimated prevalence of clinically silent adenomas in older individuals of up to 20% ( 166 ). Increased SA-βgal activity has been found in pituitary adenomas compared with normal tissue, possibly due to oncogene-induced senescence (OIS) ( 167 , 168 ). In this instance, the appearance of senescent cells may be beneficial, since OIS can slow tumor growth. Perhaps this could be the case for other endocrine tumors, but further research into this is needed. With increasing age, the ovary and uterus in females and the testis in males become dysfunctional ( 169 ), with decreased sexual function, infertility, sleep, and mood disturbances, and loss of muscle mass ( 170 ). Gonadal dysfunction with aging may be related to cellular senescence and mitochondrial dysfunction ( 171 ), and senolytic or senomorphic interventions may have the potential to delay menopause and extend the female reproductive window ( 172 ). In aged rats, p16 Ink4a levels increase in the ovaries and testes ( 173 ). In aged compared to young dogs, a 4-fold increase in p21 CIP1/WAF1 expression in testicular fibroblasts and 8 times more senescent Leydig cells in the testes were found ( 174 ). Senescent cells may interfere with the differentiation of endometrial stromal cells into decidual cells, impeding embryonic implantation and placentation, leading to infertility ( 175 ). However, a recent study using p16 Ink4a -KO mice found no improvement of infertility in an alkylating agent-induced primary ovarian insufficiency model on inhibiting p16 Ink4a -induced senescence ( 176 ). More work needs to be performed to determine if the accumulation of senescent cells in gonadal tissues causes dysfunction. The adrenals secrete cortisone, aldosterone, and dehydroepiandrosterone (DHEA), among other hormones. The connection between DHEA and longevity has been investigated in several epidemiological studies ( 177-179 ). In the previously mentioned Baltimore Longitudinal Study, high DHEA levels were associated with longevity ( 125 ). Dysregulated function of the adrenals can have wide-reaching consequences, even affecting brain function and mood. The daily cortisol rhythm is disrupted in patients with dementia, along with decreased DHEA levels. High levels of circulating cortisol may induce apoptosis of hippocampal neurons, while high levels of DHEA protect them ( 180 , 181 ). Senescent cell accumulation in the adrenals has been suggested to have detrimental effects, with senescent cells in the zona glomerulosa being linked to hyperaldosteronism ( 182 , 183 ). In a recent paper investigating potassium inwardly rectifying channel subfamily J member 5 (KCNJ5)-mutated aldosterone-producing adenomas, p21-induced cell cycle arrest was correlated with higher aldosterone-to-renin ratios ( 184 ). In KCNJ5-mutated and wild-type aldosterone-producing adenomas, compact tumor cells were more likely to be senescent than intratumor clear cells ( 184 ). More work is needed to test if and how cellular senescence affects adrenal function and tumor formation. Since DHEA replacement therapy in older women did not lead to major benefits ( 177 , 185 ), widespread use of DHEA supplementation as a gerotherapeutic intervention is not supported as of today. Studies in human and animal models have suggested an inverse relation between TH levels and longevity. Older men appear to have decreased sensitivity to thyrotropes, possibly related to decreased thyrotropin-releasing hormone 24-hour rhythmicity or increased somatostatin ( 186 , 187 ). In vivo studies in several long-lived small mammalian species suggest that lower TH levels are associated with extended longevity, such as a recent study in which 4 small mammalian species were studied ( 188 ). This indicated there is an inverse relation between thyroxine (T4) levels and species maximum lifespan. Increased mitochondrial activity related to TH may link increased TH to accelerated development of aging phenotypes, perhaps related to increased generation of DNA-damaging ROS. This appears to be associated with the accelerated appearance of aging changes in vitro and in vivo ( 189 ). ROS and cellular senescence are causally linked. ROS can induce cells to become senescent, and senescent cells activate macrophage degradation of nicotinamide adenine dinucleotide (NAD), which then leads to increased ROS generation, especially by those innate immune cells that can induce other types of nonsenescent cells to become senescent ( 190 , 191 ). Studies are needed to determine if these interlinked cellular-senescence–mediated processes are accentuated by TH. Also pointing to potential links between senescence and the effects of thyroid function on the progression of aging phenotypes is the observation that caloric restriction, which can delay the development of aging phenotypes, both reduces circulating TH and decreases senescent cell burden ( 173 ). Progressive telomere shortening, which can be both a cause and biomarker of cellular senescence, has been noted in the thyroid and parathyroid glands of humans. Thyroid and parathyroid telomeric erosion becomes evident in people older than 50 years, which is later than for other tissues in which telomere shortening has been reported ( 192 ). Perhaps this is due to the slow turnover of thyroid cells. Older women tend to have higher thyrotropin (TSH) levels than younger women and, in both sexes, triiodothyronine (T3) tends to be lower and circulating antithyroid antibodies higher in older than younger populations ( 193 ). In geographic areas where iodine intake is high, TSH levels tend to increase with age, whereas in areas with lower iodine intake, circulating TSH levels generally decrease with increasing age ( 193 , 194 ). Long-term residency in areas with high iodine content in the drinking water has been associated with increased longevity ( 195-197 ). Although epidemiological data suggest a relationship between iodine intake and longevity, little is currently known about whether this is an indirect effect of changes in TH or iodine homeostasis. The most aggressive thyroid neoplasm is anaplastic thyroid cancer (ATC). Historically, the mean survival time after this diagnosis has been established is 4 months ( 198 ). In the thyroid gland, type 2 deiodinase (D2) converts T4 into the more metabolically active T3, which is crucial for ATC cell proliferation. Interestingly, treatment by the D2 inhibitor, reverse T3, induces cellular senescence in these tumor cells ( 184 ). Together with senolytic therapy to remove these now-senescent cancer cells in a 2-step process (“1:2 punch” approach) ( 199 , 200 ), this may become a valid treatment strategy for patients suffering from this type of cancer. The conversion of the vitamin D precursor, 7-dehydrocholesterol, into previtamin D 3 occurs in the skin with exposure to ultraviolet radiation in sunlight. The capacity of the skin to synthesize previtamin D 3 decreases by up to 50% with increased age ( 148 , 201 , 202 ). Cellular senescence plays a major role in skin aging, with senescent cells accumulating that produce collagenases and elastase that disrupt skin architecture and contribute to altered skin pigmentation ( 203 , 204 ). As such, senescent cell accumulation in the skin may theoretically contribute to reduced vitamin D production, a possibility that needs to be tested. The ability of the kidneys to complete the synthesis of metabolically active vitamin D is also reduced with age ( 205-208 ). Vitamin D deficiency is widespread worldwide ( 209 ) and has been connected to obesity ( 210 ) and osteoporosis ( 211 ), among other endocrine disorders. Interestingly, vitamin D has potent geroprotective effects, with higher vitamin D levels correlating with longer telomeres in humans ( 212 ). Vitamin D has been shown to reduce senescent cell burden by inhibiting the p16 and p53 pathways ( 213 ). The thymus is effectively an endocrine organ as well as a vital component of the immune system since it produces hormones such as thymulin, thymosin, and thymopoietin, and the thymus is the first organ in the body that exhibits age-associated involution. Studies have shown that cellular senescence occurs in thymic epithelial cells due to high oxidative stress, especially during advanced stages of human thymic involution ( 214 ). The senescence indicators, p16 Ink4a , p53, p21, enhanced SA-βgal activity, and γ-H2AX were detected by immunohistochemistry in the aged human and mouse thymus ( 214-217 ). Age-related thymic involution contributes to immunosenescence and inflammaging declines due to the capacity to establish central tolerance, thereby causing increased self-reactive T cells to escape to the periphery ( 218 ). Much work remains to be done to test whether there are causal links between cellular senescence and endocrine and immune function of the thymus. Vascular endothelial cells (ECs) are abundant throughout the body and produce vasoactive peptide hormones, growth factors, coagulation factors, and adhesion molecules, effectively functioning in an endocrine-like manner. In a recent in vitro study, EC-conditioned media were collected from cultures treated with radiation to induce senescence in nonsenescent ECs. The senescent ECs had a more robust SASP than senescent epithelial cells or myoblasts. Senescent ECs also exhibited functional abnormalities, including decreased expression of endothelial nitric oxide synthase and increased ROS production ( 219 ). Production of prostenoids such as prostaglandin I2 was decreased, whereas production of PAI-1, thromboxane-A2, and endothelin-1 (was increased ( 220-222 ). Furthermore, atherosclerotic coronary arteries have increased SA-βgal ( 223 ) and p16 INK4a and p53/p21 CIP1/WAF1 expression ( 224-228 ). Senescent cells accumulate in the placenta during pregnancy ( 229 , 230 ), and senescence can spread locally and systemically ( 231 ). Vascular senescence is prominent in preeclampsia, which is linked to a disrupted fetal-maternal barrier that in turn is linked to metabolic dysfunction, hypertension, recurrence of preeclampsia during subsequent pregnancies, and accelerated development of aging phenotypes in women ( 232-234 ). The effect of senescent vascular cells on the action of vasoactive and other hormones requires further study.

Cellular senescence has an important role in several conditions related to aging as well as multiple diseases across the lifespan ( 30 , 128 , 129 , 204 , 235 ). Dietary and exercise interventions can prevent the accumulation of senescent cells by reducing DNA damage, mitochondrial dysfunction, excessive ROS, and inflammation ( 236 ) or reduce deleterious properties of senescent cells, such as attenuating the SASP ( 237 ). Caloric restriction was one of the first lifespan-extending interventions to be identified. Other dietary interventions including methionine restriction ( 238 ) and ketogenic diets ( 239 , 240 ) promote a favorable metabolic state that may limit the accumulation of senescent cells. Exercise is another promising strategy to delay aging phenotypes with numerous health benefits by preventing the accumulation of senescent cells ( 241-243 ). It has been shown that low-magnitude vibration, which mimics exercises, can alleviate age-related bone loss by inhibiting senescence of osteogenic cells in aged rats ( 244 ). However, whether these interventions have direct effects on senescent cells requires further study, as these interventions are multifaceted. In addition, since many reports showed that exercise activates intrinsic immune cells ( 245-247 ), these improvements on aging phenotypes may be due to the indirect removal of senescent cells by intrinsic immune cell activation ( 248 ). Besides these nonpharmacological approaches, currently 4 approaches for targeting senescent cells are being investigated in preclinical studies ( Fig. 3 ): (1) inhibition of senescent cell formation; (2) suppression of the SASP (senomorphics); (3) elimination of persisting senescent cells (senolytics); and (4) reprogramming of senescent cells. Scheme of senotherapies. Four strategies for attenuating detrimental effects of senescent cells. The first, targeting the senescence program itself, may lead to increased tumor formation due to apoptosis-resistant, damaged, cancerous mutation–harboring cells continuing to proliferate, and as such is not being as widely investigated currently as other approaches. The second strategy is to modify the characteristics of senescent cells using senomorphic agents that decrease production of tissue-damaging senescence-associated secretory phenotype (SASP) factors. This can be achieved using already available agents such as metformin, rapamycin, or ruxolitinib, but as the senescent cells remain in tissues, more continuous administration of senomorphic drugs may be required compared to senolytics. The third option is to target the apoptosis resistance mechanisms operative in the 30% to 70% of senescent cells that are tissue-damaging with senolytic drugs, leading to their removal. Due to the time it takes for new senescent cells to form and acquire a proapoptotic SASP, senolytics are effective even if administered intermittently using a “hit-and-run” approach. The fourth option is to alter the epigenetic programming of senescent cells by induction of Yamanaka Factors (OSKM), reverting them into a nonsenescent, replicating state. This approach, if perfected, could be promising, but since many senescent cells can harbor or develop oncogenic mutations, it could lead to cancers, including teratocarcinomas. Targeting the cell cycle inhibitors that enforce cellular senescence directly, such as p53, which is upstream of p21 CIP1/WAF1 , may prevent cells from becoming senescent. Using p53-conditional KO mice, inhibiting cellular senescence in an adipose and vascular EC-specific manner attenuated obesity and alleviated glucose intolerance ( 105 , 224 , 225 , 249 ). However, systemically reducing p16 Ink4 a or p53-p21 Cip1/Waf1 expression is likely to increase cancer risk since p53 , p21 Cip1/Waf1 , and p16 Ink4a global KO mice have a high prevalence of cancer and hence a shorter lifespan than control mice ( 250-253 ). Local or intermittent administration of agents that inhibit senescent cell formation might be an alternative, but even a few DNA-damaged cells with revived proliferative potential might still be sufficient to cause cancer. Hence, interfering with the capacity of cells to become senescent, as opposed to removing already senescent cells (which can include senescent cells harboring cancerous mutations), may not become a viable strategy because of the role of induction of senescence in preventing replication of cancerous cells. As some SASP factors have a role in chronic inflammation and the progression of multiple disorders, SASP inhibitors, also called senomorphic agents, can break this link between proinflammatory, proapoptotic senescent cells and disease without directly eliminating senescent cells. Several senomorphics target the transcription factor nuclear factor (NF)-κB, Janus kinase (JAK), or the JAK signal transducer and activator of transcription (STAT) signaling pathways. Other senomorphics target rapamycin complex 1 (mammalian target of rapamycin complex 1; mTORC1), mitochondrial complex 1- or 4-related (eg, metformin), or p38 mitogen-activated protein kinase (MAPK) family members. Further possibilities for inhibiting the SASP include modulating NAD + /NADH metabolism, inhibiting heat shock protein 90, or neutralizing SASP factors or their receptors ( 254 ). A recently described p21 Cip1/Waf1 -Cre mouse model, which has a p21 Cip1/Waf1 promoter driving inducible Cre , enables direct targeting of p21 Cip1/Waf1 -highly expressing (p21 high ) senescent cells ( 255 ). Using this model, it was demonstrated that inactivating the NF-κB pathway in p21 high cells attenuated insulin resistance in obese mice, as did the removal of p21 high cells, suggesting that SASP factors, over and above senescent cells themselves, may be a central contributor to insulin resistance ( 256 ). Rapamycin is approved as an immunosuppressant at high doses. It or related mTORC1 inhibitors used at much lower, less immune-system–suppressing doses, appear to be senomorphic and decrease frailty ( 257 ), heart failure ( 258 ), cancer formation ( 259 ), cognitive impairment ( 260 ), immune dysfunction ( 261 ), and age-related adipose tissue loss. Rapamycin also appears to increase the maximum lifespan in mice (at least in animals raised under ideal, constant, nonstressed, pathogen-free conditions that may not reflect the real world) ( 262 ). Ruxolitinib is an inhibitor of the JAK1/JAK2-STAT3 pathway, which has been implicated in cellular senescence ( 263 ). Ruxolitinib is in clinical use for various disorders (eg, polycythemia rubra vera, myelofibrosis, and graft-versus - host disease). Ruxolitinib inhibits production of some SASP factors in vitro and in vivo in aged mice ( 264 ). In rodents, it alleviates age-related adipose tissue dysfunction, decreases insulin resistance, reduces age-related osteoporosis and frailty, alleviates critical illness myopathy, and decreases progenitor cell dysfunction ( 35 , 265 ). In older myeloproliferative syndrome patients, ruxolitinib partially attenuates frailty and increases appetite, body weight, and skeletal muscle strength, although it does not directly affect the hematological disorder itself. This indicates that ruxolitinib may alleviate frailty and geriatric phenotypes through mechanisms such as SASP inhibition independently of its effects on hematological function ( 266 ). Metformin, an inexpensive drug that has been used to treat diabetes for more than 60 years, reduces the release of multiple proinflammatory SASP factors by senescent cells, with NF-κB inhibition playing a key role ( 267 ). Metformin has been shown to delay, prevent, or alleviate multiple age-related disorders, including cardiovascular diseases ( 268 ), cognitive dysfunction ( 269 ), and diabetes in animals and humans ( 270 ). A retrospective analysis of patients with diabetes who received metformin suggested there could be an increase in lifespan compared to individuals without diabetes ( 271 ). The proposed TAME (Targeting Aging with Metformin) clinical trial will test if metformin delays the appearance of a second age-related disease in patients who already have a single age-related condition ( 272 , 273 ). Disentangling the effects due to SASP modulation by these agents from other “off-target” effects that could lead to side effects is difficult. Examples of potential adverse off-target effects of senomorphics may include suppression of cytokine secretion by nonsenescent immune cells (eg, in the case of rapamycin), potentially hindering physiologically necessary inflammation in the face of an infection, or inhibition of anabolism, potentially contributing to reduced myogenesis under some conditions. In addition, since the senescent cells themselves remain in the body, continuous administration of SASP inhibitors may be necessary, potentially leading to more side effects than senolytics, which appear to be effective even if administered intermittently. However, it should be noted that some senomorphics may be effective intermittently ( 274 ), perhaps because they decrease the SASP-induced spread of senescence (discussed next). The idea of selectively targeting and removing senescent cells was suggested by a study that indicated interventions such as caloric restriction that enhance health span or lifespan are associated with decreased senescent cell burden in mice ( 173 ). Based on that earlier study and after work to develop senolytic agents that selectively eliminate senescent cells had already begun, this speculation was reinforced by the observation that removing cells highly expressing p16 Ink4a , many or perhaps most of which are senescent cells, enhances function in transgenic mice with an accelerated aging-like state due to transgenic expression of a gene predisposing to DNA mutations ( 275 ). In these INK-ATTAC mice, administration of a drug that has little or no effect on cells lacking the ATTAC transgene (first used in 2005 in FAT-ATTAC mice to remove fat cells) ( 276 ), highly expressing p16 Ink4a cells undergo apoptosis. This observation in progeroid mice suggested that cellular senescence might be implicated in generating age-related phenotypes in naturally aging mice and that removal of senescent cells might prevent or delay cellular senescence-related dysfunction ( 277 , 278 ). This was first demonstrated across multiple age-related disorders in naturally aged mice treated with senolytic agents (discussed next) that clear senescent cells ( 75 ) and confirmed in naturally aged INK-ATTAC mice, in which targeting highly expressing p16 Ink4a cells alleviated senescence-related metabolic dysfunction ( 35 ). Fulfilling Koch's postulates of causality, it was shown that transplanting small numbers of senescent cells into middle-aged mice is sufficient to cause frailty, physical dysfunction, and the premature onset of most or all of the diseases that older nontransplanted mice normally die from ( 231 ). It was further shown that transplanting senescent cells causes the recipient's own cells to undergo senescence, even at a distance. Hence, senescence can spread not only in a paracrine but also distantly in an “endocrine” fashion. Additionally, intermittently eliminating senescent cells alleviated physical dysfunction after senescent cell transplantation, all pointing toward senescent cells as a causal contributor to aging phenotypes and diseases ( 231 ). Possible benefits of selectively eliminating senescent cells were subsequently indicated in mouse models using other markers of aging such as p21 Cip1/Waf1 ( 255 , 256 ) and p19 ( 279 ). These findings support a potential role for targeting senescent cells as a therapeutic strategy to delay, prevent, alleviate, or treat multiple age-related pathologies, including endocrine and metabolic diseases and disorders ( 68 , 129 ). As senescent cells are more apoptosis-resistant than nonsenescent cells ( 280 ), it was theorized that they may have increased expression of prosurvival signals, defending them against their own proapoptotic SASP factors ( 75 ). To test this, proteomic data from different types of human senescent cells were analyzed and compared to nonsenescent cells. Using bioinformatics approaches, these data were further analyzed with the goal of identifying specific senescent cell antiapoptotic pathways (SCAPs). It was discovered that different SCAPs are upregulated depending on the senescent cell type. Next, key identified SCAP nodes were inhibited using small interfering RNAs (siRNAs). This induced apoptosis in the 30% to 70% of senescent cells that are proapoptotic and tissue-damaging ( 75 ). These SCAP components included ephrins (EFNB1 or 3)/SRC kinases, the phosphatidylinositol-4,5-bisphosphate 3-kinase delta catalytic subunit (PI3KCD), the cyclin-dependent kinase inhibitor 1A (CDKN1A; p21 CIP1/WAF1 ), BCL-xL, mitochondrial pathways, and plasminogen-activated inhibitor-2 (PAI-2) ( 75 , 281-284 ). Next, bioinformatics approaches were used to select small molecules that could target the identified SCAPs. Dasatinib (D), a drug approved by the Food and Drug Administration since 2006 for hematologic malignancies, is a kinase inhibitor that interferes with the SRC kinase/EFNB1/3-dependent apoptosis resistance of senescent human fat cell progenitors, leading to their selective apoptosis ( 285 , 286 ). Quercetin (Q) is a natural flavonoid in apple skin, capers, and red onions that is available in the United States and Europe as a health supplement. Q is known to interfere with PI3K, and this mechanism was shown to induce death in tissue-damaging senescent human ECs ( 75 ). The combination of dasatinib and quercetin (D + Q) was theorized to have wider senolytic activity than either alone and this combination successfully reduced senescent cell burden in chronologically aged mice ( 75 ). Based on the findings regarding BCL-xL in the aforementioned siRNA studies, 8 months later 2 groups found that navitoclax (ABT-263) also exhibited senolytic activity against certain types of senescent cells and could alleviate age-related disorders and dysfunction ( 287 , 288 ). To date, multiple additional senolytics have been identified using the original hypothesis-based drug discovery approach and, later, by high-throughput screening. Examples include the specific BCL-xL inhibitors A1331852 and A1155463 ( 289 ), the flavonoid fisetin ( 289 , 290 ), piperlongumine ( 282 ), procyanidin C1 ( 291 ), and FOXO4-related peptide ( 292 ), among others. Several dozen small-molecule senolytics have now been identified ( 128 ). Since senescent cells can take from 1 to 6 weeks to fully develop, at least in vitro, and senescent cells do not replicate, senolytics can be administered once every couple of weeks or once a month. In mouse models, this appears to be as or more effective than administering these agents continuously ( 30 , 128 , 150 ). Most current senolytics act on SCAP pathways. Second-generation senolytics are now being identified using high-throughput library screens and other approaches, including cardiac glycosides, such as digoxin, which was found to decrease the number of senescent cells after senescent cell implantation in mice ( 293 , 294 ). Other recently identified senolytic targets include glutaminase 1 ( 295 ), GPNMB (glycoprotein nonmetastatic melanoma protein B) ( 226 , 296 ), CD153 ( 297 ), uPAR (urokinase-type plasminogen activator receptor) ( 298 ), and intravenous zoledronic acid, a bisphosphonate that is both senolytic and senomorphic ( 299 ). Another strategy targeting aging processes being tested by some research groups is cellular reprogramming using the “Yamanaka factors,” Oct4, Sox2, Klf4, and c-Myc (OSKM) ( 300 ). This strategy, which uses these transcription factors to bring senescent cells closer to a stem cell–like state, may convert the senescent cells accumulated in endocrine organs into functional cells. This reprogramming of senescent cells may be a viable treatment for several endocrine diseases, such as diabetes. In a murine model, with mice modified with an OSKM polycystronic cassette and in which OSKM expression can be induced on doxycycline treatment, OSKM induction increased pancreatic β-cell regenerative capacity and reduced metabolic dysfunction caused by streptozotocin. In older wild-type mice, reprogramming enhanced muscle regeneration after cardiotoxin-induced muscular injury ( 301 ), indicating that reprogramming may alleviate sarcopenia, much like intermittent oral senolytic D + Q administration ( 158 ). The field is still new, and it remains to be seen if it can be made safer and if creating cancers, such as malignant teratomas, by allowing potentially cancer-harboring senescent cells to reenter the cell cycle can be avoided.

In recent years, research targeting fundamental aging processes has made remarkable progress, and senescent cell-targeting therapies including SASP inhibitors, senolytics, and reprogramming are promising. With accumulating evidence from preclinical studies, some of these therapies have progressed to the point of early clinical trials. Diabetes, metabolic syndrome, macular degeneration, and osteoporosis are among endocrine and endocrine-related disorders for which clinical senolytic trials are planned or ongoing. Senotherapies are a new treatment strategy and have the potential to significantly change existing treatments. Targeting senescent cells is an intervention against a root-cause contributor to the diseases of aging and may improve age-related conditions in a comprehensive manner, rather than one disease at a time. On the other hand, this field is new, and there are many issues such as long-term outcomes, side effects, and the establishment of gerodiagnostics, so we must proceed carefully. Adoption by the Food and Drug Administration or the World Health Organization International Classification of Diseases code for impaired functional capacity would help in obtaining and tracking data at an epidemiological level and facilitating regulatory oversight. If interventions or combinations or interventions (eg, lifestyle plus drug) can be developed that target interlinked fundamental aging processes additively or synergistically, their first use will likely be for serious diseases and disorders linked to these processes. Eventually, if safe and effective, senotherapies may move toward use for less-serious conditions, as agents coadministered with other interventions already developed to treat particular diseases, and perhaps, eventually, for secondary prevention and even primary prevention to enhance health span.

Cellular senescence has been implicated in various endocrine disorders, and preclinical data showing improvement of endocrine and related disorders with senolytics are accumulating ( Table 1 ). Ongoing research in the senescence field may, apart from these common conditions, also possibly yield new treatment options against relatively rarer diseases, such as hypothalamic and pituitary gland disorders and aldosterone-secreting adenomas, along with several other endocrine disorders, but little is currently known about these rarer diseases as there are either no adequate preclinical models or existing models are difficult to evaluate. The evidence summarized here is for senolytics in animal models of more common conditions. Results from preclinical studies of senolytics in endocrine diseases Abbreviation: D + Q, dasatinib and quercetin. In diet-induced obese mice, senescent cell clearance has beneficial effects on adipose tissue function and systemic metabolism, with the ratio of subcutaneous to visceral fat being increased without reducing total body weight ( 30 , 34 ). Senescent cell clearance also decreased lipid deposition in muscle and liver ( 34 , 42 ). These changes and the extent of senescent cell clearance correlated with enhanced insulin sensitivity ( 34 ). Senescent cell clearance in aged mice mitigated age-related subcutaneous fat tissue atrophy by enhancing adipogenesis ( 35 ). In addition to these metabolic benefits, senolytics also prevented or alleviated some complications of diabetes in obese mice, including diastolic cardiac dysfunction, hepatic steatosis, microalbuminuria, and obesity-induced anxiety ( 34 , 35 , 39 , 42 ). Clinical trials of senolytics for diabetes, adipose dysfunction, and their complications are underway ( Table 2 ). Clinical studies of senotherapies for endocrine-related diseases Abbreviation: D + Q, dasatinib and quercetin. Intermittent senolytic treatment in old mice reduced p16 Ink4a expression and SADS in osteocyte-enriched bone samples ( 78 ). Both genetic and pharmacologic removal of senescent cells as well as the SASP inhibitor, ruxolitinib, suppressed cortical bone resorption, increased bone formation on endocortical surfaces, and enabled bone maintenance on trabecular surfaces of some bones. This resulted in improved bone microarchitecture and increased bone strength ( 44 , 150 ). Senolytics increased osteoblast numbers and bone formation rates while reducing bone marrow adipose tissue. This was due to alterations in progenitor commitment toward osteoblast formation and away from osteoclast and adipocyte formation following senescent cell clearance. Senolytics also alleviated frailty symptoms in these mice ( 75 , 302 ), indicating far-reaching effects of senescent cell removal, even when targeting a single disease. Clinical trials of senolytics for age-related osteoporosis are underway (see Table 2 ). In the pituitary tumor, adamantinomatous craniopharyngioma (ACP), cells that carry oncogenic β-catenin mutations and have increased Wnt signaling, form cell clusters that become senescent and promote SASP factor production. Evidence supporting paracrine signaling by senescent cells as a risk factor for tumorigenesis across different tumors and cancer models is accumulating ( 303 ). In ex vivo cultures of pretumoral pituitary glands, the senolytic compounds navitoclax and ABT-737 tended to reduce the average size of the clusters ( 304 ) and the cardiac glycosides ouabain and digoxin, which also are senolytic compounds, killed oncogenic β-catenin–expressing cells and reduced the levels of senescence markers and SASP factors ( 293 ). These data suggest that senolytics may synergize with established anticancer drugs by eliminating OIS cells. In a small study of endometrial tissue from patients with endometriosis, D alone, Q alone, and especially D + Q reduced senescence markers and enhanced decidualization marker expression, suggesting further studies for infertility are warranted ( 175 ).