A guide to senolytic intervention in neurodegenerative disease

Abstract

Cellular senescence is a potential tumor-suppressive mechanism that generally results in an irreversible cell cycle arrest. Senescent cells accumulate with age and actively secrete soluble factors, collectively termed the ‘senescence-associated secretory phenotype’ (SASP), which has both beneficial and detrimental effects. Although the contribution of senescent cells to age-related pathologies has been well-established outside the brain, emerging evidence indicates that brain cells also undergo cellular senescence and contribute to neuronal loss in the context of age-related neurodegenerative diseases. Contribution of senescent cells in the pathogenesis of neurological disorders has led to the possibility of eliminating senescence cells via pharmacological compounds called senolytics. Recently several senolytics have been demonstrated to elicit improved cognitive performance and healthspan in mouse models of neurodegeneration. However, their translation for use in the clinic still holds several potential challenges. This review summarizes available senolytics, their purported mode of action, and possible off-target effects. We also discuss possible alternative strategies that may help minimize potential side-effects associated with the senolytics approach.

1. Introduction

Aging is the most important risk factor for several neurodegenerative diseases, each characterized by a regionally-distinct pattern of neuronal loss in the brain. Multiple factors are believed to play a role in this neurodegeneration including increased oxidative stress, mitochondrial dysfunction, protein aggregation, and chronic inflammation to name but a few (Hou et al., 2019). Despite enormous efforts on the part of the biomedical community over the last two decades, the development of novel therapies for these disorders has yielded few success stories. In the case of Alzheimer’s disease (AD), after almost two decades of research, only recently has the US Food and Drug Administration (FDA) approved a drug for clinical use: the amyloid beta antibody aducanumab (Aduhelm; Biogen Inc). The overall disheartening outcomes in interventions for neurodegenerative disorders have led to consideration of a relatively novel approach, targeting basic aging mechanisms in lieu or in addition to approved interventions, in the hopes that this would lead to better overall therapeutic outcomes with broader applicability (Florence C. C. Tan, Emmette R. Hutchison, 2014; Saez-Atienzar and Masliah, 2020; Walton et al., 2020).

One such fundamental aging mechanism is cellular senescence. Senescence is traditionally considered an onco-suppressive mechanism that generally results in an irreversible cell cycle withdrawal (Campisi, 2001; Gorgoulis et al., 2019; Hayflick and Moorhead, 1961). Senescent cells can develop what is known as a senescence associated secretory phenotype (SASP) (Coppé et al., 2010, 2008). While irreversible cell cycle withdrawal is relatively harmless to the tissue microenvironment, the SASP involves the secretion of cytokines, chemokines, mitogenic factors and proteases that can lead to deleterious effects on surrounding tissues (Coppé et al., 2010). Senescence has been shown to spread via a process known as secondary senescence (also known as senescence-induced senescence or paracrine senescence) (Acosta et al., 2013; Hubackova et al., 2012; Nelson et al., 2012) (Figure 1). Whilst initially considered to be a relatively benign cell-autonomous onco-suppressive phenomenon, subsequent SASP production by senescent cells has now been widely shown to drive aging and age-related disease non-autonomously in a plethora of tissues.

Figure 1. Non-cellular autonomous features of cellular senescence.

Senescent cells can develop a SASP which can be deleterious for the surrounding microenvironment. Senescent cells can also induce cellular senescence in neighboring cells in what is known as secondary senescence, senescence-induced senescence, or paracrine senescence.

Arguably the strongest evidence supporting the potential benefits of senescent cell elimination comes from genetic models (Baker et al., 2011; Demaria et al., 2014; Hashimoto et al., 2016). Unlike senolytic compounds, genetic deletion targets the senescence machinery itself (Figure 2). This is achieved by placing genes that encode a protein that can be exogenously activated to trigger cell death under the same promoters that directly or indirectly drive p16 (Baker et al., 2011; Chinta et al., 2018; Demaria et al., 2014) or p53/p21 (Hashimoto et al., 2016), major regulators of cellular senescence (Donehower et al., 1992; Gorgoulis et al., 2019; Sharpless et al., 2001). The addition of compounds that activate the proteins will consequently result in the selective elimination of cells that have up-regulated senescence pathways. The result is the selective killing of senescent cells and proof-of-concept for senotherapies.

Figure 2. Avenues for the selective killing of senescent cells.

A. Genetic deletion models encode a protein (skull) that can be exogenously activated by the addition of an activator compound (AP20187, GCV, DT) to trigger cell death under promoters that regulate major elements of the senescence signaling cascade. B. Resistance to cell death in senescent cells is bestowed by the up-regulation of pro-survival and or anti-apoptotic networks. Senolytic compounds target different elements of these networks to kill senescent cells. C. Chimeric antigen receptors cytotoxic T cells (CAR-T) have been implemented as a novel senolytic approach. CARs are design to target surface proteins, which then activates the cytotoxic function of the T-cell. By targeting membrane-bound ligands present only on senescent cells, CAR-T can be used as a live cell senolytic intervention strategy. The skull represents INK-ATTAC product FK506 binding protein-caspase 8 (FKB-Casp8) fusion protein, the p16-3MR product truncated herpes simplex virus thymidine kinase (HSV-TK), and the ARF-DTR product diphtheria toxin receptor (DTR). AP20187 is a synthetic drug that induces the dimerization of FKB-Casp8 and initiates apoptosis (Baker et al., 2012). GCV, ganciclovir, DT, diphtheria toxin. In p16-3MR mice, HSV-TK phosphorylates GCV, a nucleotide analogue, and converts it into a toxic DNA chain terminator (Laberge et al., 2013; Ray et al., 2004). In ARG-DTR mice, DT activated the DTR which initiates cell death (Hashimoto et al., 2016).

2. Senotherapies: senomorphics versus senolytics

Two basic ‘flavors’ of seno-therapeutics approaches have been developed—what are known as senomorphics and senolytics (Kim and Kim, 2019; Kirkland and Tchkonia, 2020; Short et al., 2019). Senomorphic compounds alter (“morph”) the phenotype of senescent (“seno”) cells into a non-deleterious phenotype by preventing the SASP without killing the senescent cells themselves. In this review we will focus on senolytics; for a description of senomorphics agents we recommend three more detailed reviews on the topic (Kim and Kim, 2019; Kirkland and Tchkonia, 2020; Short et al., 2019). The purpose of senolytic compounds is to selectively kill (“lytic”) senescent (“seno”) cells, thereby eliminating the accompanying SASP. Senescent cells can become resistant to apoptosis (Childs et al., 2014; Wang, 1995; Zhu et al., 2015) and senolytic compounds largely work by abrogating the up-regulation of anti-apoptotic pathways conferring this resistance (Kim and Kim, 2019; Kirkland and Tchkonia, 2020).

A fundamental difference between genetic deletion models and senolytic compounds is that the former target the senescence machinery itself and the latter largely target pro-survival signaling networks (Figure 2A & B). Senescent cells rely on different pro-survival networks, and accordingly, distinct compounds have been selected and tested (Table 1). For example, navitoclax is efficient at killing senescent HUVEC but not senescent human preadipocytes (Zhu et al., 2016). Fisetin also efficiently kills senescent HUVEC but not senescent IMR90 cells or human pre-adipocytes (Zhu et al., 2017). Dasatinib proved capable of eliminating senescent preadipocytes but not HUVECs whilst the opposite was true for quercetin; with a cocktail of the two eliminating both (Zhu et al., 2015).

Table 1. Heterogeneity of senolytics in vitro.

Non-exhaustive compilation of in vitro experiments using senolytics. We have detailed the article and figure as well as a “Cell type” column, whether it is human or mouse in the “Species” column, the type of stressor that has been used to induce senescence in the “Stressor” column, and the senolytic dose in μM as well as treatment duration in the “Senolytic (μM)” and “Duration” columns, respectively. The percentage-range values are approximations extracted from the analysis of the articles and figures referenced in the table. “Non-senescent % killing” reflects the OFF-target killing of non-senescent cells. “Senescent % killing” reflect the ON-target killing of senescent cells. “% Selectivity” is “Senescent % killing” minus “Non-senescent % killing”. Selectivity is not calculated from the ranges depicted in the table but from more accurate data presented in the articles of reference. For example, the range 76-100% minus 51-75% can either be 100% - 51% or 76% minus 75%, resulting in 26-50% and 0-2% ranges, respectively. When data concerning the killing of non-senescent cells is not provided the “% Selectivity” will reflect “N.A.” for not available. The values are presented in generous ranges because they are not based on numerical values, as most articles do not provide them, but carefully extrapolated from graphs (e.g. bar charts). In addition, some studies have important variations in their experimental conditions. For example, proliferating and quiescent cells will result in different selectivity because the former mask OFF-target killing with newly generated cells. Generous percent ranges make it harder for senolytics to stand out and also easier to identify outlier reported killing when they do stand out and results are not replicated in other studies. Species: H, Human; M, Mouse. Stressors: “Bleo”, Bleomycin; “Bleo (IV)”, Bleomycin administered in vivo; “Doxo”, Doxorubicin; “Eto”, Etoposide; “Ercc1^−/Δ”, mouse model of human progeroid syndrome; “OS”, Oxidative Stress; “RS”, Replicative Senescence. Cell types: “BMDMC”, Bone Marrow Derived Mesenchymal Cells; “CPC”, Cardiac Progenitor Cells; “FT2AEC”, Fibrotic Type II Alveolar Epithelial Cells; “HUVEC”, human umbilical endothelial cell; “MEF”, mouse embryonic fibroblast; “MSC”, Mesenchymal Stem Cells; “OPC”, Oligodendrocyte Precursor Cells; “PA”, Preadipocytes; “REC”, Renal Epithelial Cells; “T2AEC”, Type II Alveolar Epithelial Cells

Article	Figure	Cell type	Species	Stressor	Senolytic (μM)		Duration	Non-senescent % Killing	Senescent % Killing	% Selectivity
Zhang et al., 2019	4a	N2A	M	IR	Quercetin	Dasatinib	24h
					5	0.5		<2%	3-25%	3-25%
						1.5		3-25%	3-25%	3-25%
	S15c	OPC	M	IR	Quercetin	Dasatinib	72h
					5	0.3		3-25%	3-25%	3-25%
						1		3-25%	51-75%	51-75%
Zhu et al., 2015	2c	PA	H	IR	Quercetin	Dasatinib	72h
					20	0		<2%	<2%	<2%
						0.05		3-25%	26-50%	3-25%
						0.1		3-25%	26-50%	3-25%
						0.2		3-25%	26-50%	3-25%
						0.4		15-35%	26-50%	3-25%
						0.6		15-35%	26-50%	3-25%
						0.8		15-35%	26-50%	3-25%
	2c	HUVEC	H	IR	Quercetin	Dasatinib	72h
					10	0		15-35%	<2%	<2%
						0.05		40-60%	26-50%	3-25%
						0.1		40-60%	26-50%	3-25%
						0.2		51-75%	76-100%	3-25%
						0.4		76-100%	76-100%	3-25%
						0.6		76-100%	76-100%	3-25%
						0.8		76-100%	76-100%	<10%
	2d	PA	H	IR	Quercetin	Dasatinib	48h
					15	0.1		3-25%	15-35%	3-25%
						0.2		3-25%	26-50%	26-50%
					30	0.1		26-50%	26-50%	3-25%
						0.2		26-50%	51-75%	26-50%
	2e	PA	H	IR	Quercetin	Dasatinib	12h
					20	0.2		<2%	51-75%	51-75%
	3a	MEF	M	Ercc1−/Δ	Quercetin	Dasatinib	48h
					50	0.25		N.A.	51-75%	N.A.
	3b	MDBMC	M	Progeria	Quercetin	Dasatinib	48h
					100	0.5		N.A.	26-50%	N.A.
Schafer et al., 2017	3g	Fibroblasts	H	IR	Quercetin	Dasatinib	72h
					15	0.5 (0.1)		N.A.	26-50%	N.A.
						1		N.A.	26-50%	N.A.
						5		N.A.	26-50%	N.A.
						10		N.A.	26-50%	N.A.
						20		N.A.	26-50%	N.A.
	S3	IMR90	H	Eto	Quercetin	Dasatinib	48h
					15	20		N.A.	76-100%	N.A.
Louis-McDougall et al., 2019	5a	CPC	H	Doxo	Quercetin	Dasatinib	24h
					20	0.5		15-35%	76-100%	51-75%
					20	1		26-50%	76-100%	51-75%
					20	5		26-50%	76-100%	40-60%
					20	10		40-60%	76-100%	15-35%
					20	20		76-100%	76-100%	3-25%
		CPC	H	Doxo	Quercetin	Dasatinib	24h
					0.5	0.5		15-35%	65-85%	40-60%
					1	0.5		15-35%	65-85%	40-60%
					5	0.5		15-35%	65-85%	40-60%
					10	0.5		15-35%	65-85%	40-60%
					20	0.5		15-35%	76-100%	40-60%
Lehman et al., 2017	4a	PMAET2	M	Bleo (IV)	Quercetin	Dasatinib	48h
					50	0.2		N.A.	26-50%	N.A.
Schafer et al., 2017	3g	Fibroblasts	H	IR	Navitoclax		72h
					0.4			N.A.	3-25%	N.A.
					1			N.A.	<2%	N.A.
					5			N.A.	3-25%	N.A.
					10			N.A.	3-25%	N.A.
					20			N.A.	3-25%	N.A.
	S3	IMR90	H	Eto	Navitoclax		48h
					10			N.A.	<2%	N.A.
Chang et al., 2016	1a	WI38	H	IR	Navitoclax		72h
					0.313			3-25%	26-50%	26-50%
					0.625			3-25%	51-75%	40-60%
					1.25			3-25%	65-85%	51-75%
					2.5			3-25%	76-100%	51-75%
					5			3-25%	76-100%	51-75%
		WI38	H	RS	Navitoclax		72h
					0.313			3-25%	3-25%	3-25%
					0.625			3-25%	15-35%	3-25%
					1.25			3-25%	40-60%	26-50%
					2.5			3-25%	51-75%	51-75%
					5			3-25%	65-85%	51-75%
		WI38	H	Ras	Navitoclax		72h
					0.313			3-25%	26-50%	26-50%
					0.625			3-25%	51-75%	51-75%
					1.25			3-25%	65-85%	51-75%
					2.5			3-25%	76-100%	51-75%
					5			3-25%	76-100%	51-75%
	1c	IMR-90	H	IR	Navitoclax		72h
					0.012			3-25%	15-35%	3-25%
					0.037			3-25%	26-50%	15-35%
					0.111			3-25%	51-75%	26-50%
					0.333			3-25%	65-85%	51-75%
					1			3-25%	76-100%	51-75%
					3			15-35%	76-100%	51-75%
		REC	H	IR	Navitoclax		72h
					0.004			3-25%	3-25%	3-25%
					0.012			3-25%	26-50%	26-50%
					0.037			3-25%	51-75%	40-60%
					0.111			15-35%	76-100%	51-75%
					0.333			26-50%	76-100%	40-60%
					1			76-100%	76-100%	3-25%
		MEF	M	IR	Navitoclax		72h
					0.313			3-25%	26-50%	26-50%
					0.625			3-25%	51-75%	26-50%
					1.25			26-50%	65-85%	26-50%
Pan et al., 2017	3a	T2AEC	M	IR	Navitoclax		24h
					0.5			3-25%	3-25%	<2%
					1			3-25%	3-25%	3-25%
					3			3-25%	40-60%	26-50%
					5			26-50%	65-85%	26-50%
Zhu et al., 2016	2a	HUVECs	H	IR	Navitoclax		72h
					0.001			<2%	3-25%	3-25%
					0.005			<2%	3-25%	3-25%
					0.1			<2%	26-50%	26-50%
					0.5			3-25%	51-75%	26-50%
					1			3-25%	76-100%	51-75%
		IMR90	H	IR	Navitoclax		72h
					0.4			3-25%	3-25%	3-25%
					1			3-25%	15-35%	3-25%
					4			3-25%	26-50%	3-25%
					8			15-35%	26-50%	3-25%
					10			15-35%	26-50%	3-25%
		PA	H	IR	Navitoclax		72h
					0.4			3-25%	<2%	<2%
					1			<2%	<2%	<2%
					4			<2%	<2%	<2%
					8			3-25%	<2%	<2%
					10			15-35%	<2%	<2%
	2c	HUVECs	H	IR	TW-37		72h
					0.4			<2%	<2%	<2%
					1			<2%	<2%	<2%
					4			<2%	3-25%	3-25%
					8			3-25%	<2%	<2%
					10			51-75%	26-50%	<2%
		IMR90	H	IR	TW-37		72h
					0.4			<2%	<2%	<2%
					1			<2%	3-25%	3-25%
					4			<2%	3-25%	3-25%
					8			3-25%	3-25%	<2%
					10			26-50%	3-25%	<2%
		PA	H	IR	TW-37		72h
					0.4			<2%	<2%	<2%
					1			<2%	<2%	<2%
					4			<2%	<2%	<2%
					8			26-50%	<2%	<2%
					10			26-50%	3-25%	<2%
Yosef et al., 2016	2a	IMR 90	H	Eto	Obatoclax		24h
					0.01			3-25%	3-25%	<2%
					0.1			15-35%	3-25%	<2%
					1			51-75%	26-50%	<2%
				Ras	Obatoclax
					0.01			3-25%	3-25%	3-25%
					0.1			15-35%	26-50%	3-25%
					1			51-75%	15-35%	<2%
	2b	IMR 90	H	Eto	Venetoclax		24h
					1			3-25%	3-25%	<2%
					5			3-25%	3-25%	3-25%
					10			3-25%	3-25%	3-25%
				Ras	Venetoclax
					1			3-25%	3-25%	<2%
					5			<2%	3-25%	3-25%
					10			<2%	15-35%	15-35%
Li et al., 2020	2	WI-38	H	IR	EF24		72h
					0.5			3-25%	3-25%	<2%
					1			3-25%	3-25%	3-25%
					2			3-25%	26-50%	3-25%
					4			26-50%	65-85%	26-50%
				RS	EF24
					0.5			3-25%	3-25%	3-25%
					1			3-25%	51-75%	26-50%
					2			3-25%	51-75%	40-60%
					4			26-50%	76-100%	51-75%
				Ras	EF24
					0.5			3-25%	3-25%	3-25%
					1			3-25%	76-100%	51-75%
					2			3-25%	76-100%	65-85%
					4			26-50%	76-100%	51-75%
Yousefzadeh et al., 2018	1b	Ercc1−/− MEFs	M	OS	Fisetin		48h
					1			<2%	<2%	<2%
					5			<2%	3-25%	3-25%
					10			<2%	26-50%	26-50%
					20			<2%	51-75%	51-75%
	1c	IMR90	H	ETO	Fisetin		24h
					1			<2%	<2%	<2%
					3.75			<2%	<2%	<2%
					7.5			<2%	15-35%	26-50%
					15			3-25%	51-75%	15-35%
Zhu et al., 2017	1e	PA	H	IR	Fisetin		12h
					5			<2%	3-25%	3-25%
					10			<2%	<2%	<2%
					20			<2%	3-25%	3-25%
					40			3-25%	3-25%	<2%
					60			26-50%	3-25%	<2%
	1f	IMR90	H	IR	Fisetin		12h
					0.001			<2%	<2%	<2%
					0.005			<2%	<2%	<2%
					0.025			<2%	3-25%	3-25%
					0.05			3-25%	<2%	<2%
					0.5			<2%	3-25%	3-25%
					1			<2%	<2%	<2%
					10			3-25%	3-25%	3-25%
					50			51-75%	76-100%	3-25%
	1g	HUVEC	H	IR	Fisetin		12h
					0.25			<2%	<2%	<2%
					0.5			<2%	<2%	<2%
					1			<2%	<2%	<2%
					5			<2%	<2%	<2%
					10			3-25%	3-25%	3-25%
					20			3-25%	26-50%	3-25%
					50			26-50%	51-75%	3-25%
	2e	PA	H	IR	A1331852		12h
					0.001			<2%	3-25%	3-25%
					0.005			<2%	<2%	<2%
					0.05			<2%	<2%	<2%
					0.25			<2%	<2%	<2%
					0.5			<2%	<2%	<2%
					1			<2%	<2%	<2%
					10			<2%	<2%	<2%
					50			51-75%	26-50%	<2%
	2f	IMR90	H	IR	A1331852		12h
					0.001			<2%	3-25%	3-25%
					0.005			3-25%	40-60%	26-50%
					0.05			3-25%	51-75%	51-75%
					0.25			3-25%	51-75%	40-60%
					0.5			3-25%	51-75%	51-75%
					1			26-50%	51-75%	26-50%
					10			26-50%	65-85%	26-50%
					50			76-100%	76-100%	<2%
	2g	HUVEC	H	IR	A1331852		12h
					0.001			3-25%	3-25%	3-25%
					0.005			3-25%	3-25%	3-25%
					0.05			3-25%	15-35%	3-25%
					0.25			26-50%	26-50%	3-25%
					0.5			26-50%	26-50%	3-25%
					1			26-50%	40-60%	3-25%
					10			76-100%	65-85%	<2%
					50			76-100%	51-75%	<2%
	3e	PA	H	IR	A1155463		12h
					0.001			<2%	<2%	<2%
					0.005			<2%	<2%	<2%
					0.05			<2%	3-25%	3-25%
					0.25			<2%	3-25%	3-25%
					0.5			3-25%	3-25%	3-25%
					1			<2%	3-25%	3-25%
					10			<2%	3-25%	3-25%
					50			51-75%	26-50%	<2%
	3f	IMR90	H	IR	A1155463		12h
					0.001			<2%	3-25%	3-25%
					0.005			<2%	15-35%	15-35%
					0.05			<2%	26-50%	26-50%
					0.25			<2%	51-75%	51-75%
0.5					<2%	51-75%		51-75%
1					<2%	65-85%		51-75%
10					26-50%	65-85%		26-50%
50					51-75%	65-85%		3-25%
3g	HUVEC	H	IR	A1155463		12h
				0.001			3-25%	15-35%	3-25%
				0.005			3-25%	3-25%	3-25%
				0.05			3-25%	15-35%	3-25%
				0.25			3-25%	3-25%	3-25%
				0.5			3-25%	40-60%	26-50%
				1			3-25%	26-50%	15-35%
				10			76-100%	51-75%	<2%
				50			76-100%	65-85%	<2%
Baar et al., 2017	3A	IMR90	H		FOXO-DRI
					5-7.5			<2%	26-50%	26-50%
					10-15			<2%	76-100%	76-100%
					25			<2%	76-100%	76-100%
					17-18			26-50%	76-100%	51-75%
	S3B	IMR90	H	IR	FOXO-DRI
					2.5-5			<2%	3-25%	3-25%
					5-7.5			<2%	3-25%	3-25%
					12-13			<2%	26-50%	26-50%
					17-18			<2%	51-75%	51-75%
	4D	IMR90	H	Doxo	FOXO-DRI
					5-10			<2%	65-85%	65-85%
					10-15			3-25%	76-100%	76-100%
					25			3-25%	76-100%	76-100%
					50			26-50%	76-100%	51-75%
	4E	IMR90	H	Doxo	FOXO-DRI
					8-10			<2%	26-50%	26-50%
					18-20			<2%	51-75%	51-75%
Fuhrmann-Stroissnigg et al., 2017	5a/b	MEFs	M	OS	17-DMAG		48h
					0.03			3-25%	3-25%	3-25%
					0.1			3-25%	15-35%	3-25%
					0.3			3-25%	26-50%	3-25%
					1			3-25%	26-50%	3-25%
	5a/b	MEFs	M	OS	Geldanamycin		48h
					0.03			3-25%	3-25%	3-25%
					0.1			3-25%	26-50%	15-35%
					0.3			3-25%	26-50%	15-35%
					1			15-35%	76-100%	51-75%
	6d	MSC	M	OS	17-DMAG		24-48h
					0.1			N.A.	51-75%	N.A.
		IMR90	H	Eto	17-DMAG		24h
					0.1			N.A.	26-50%	N.A.
		WI38	M	RS	17-DMAG		110d
					0.1			N.A.	15-35%	N.A.
		MEFs	M	OS	17-DMAG		24-48h
					0.1			N.A.	40-60%	N.A.

The plethora of survival pathways are behind the use of combined Dasatinib-quercetin (DQ) as a senolytic cocktail (Lehmann et al., 2017; Lewis-McDougall et al., 2019; Schafer et al., 2017; Zhu et al., 2015). Dasatinib targets tyrosine kinases including Src family kinases and Bcr-Abl (Dorsey et al., 2000; Louis J. Lombardo et al., 2004) while quercetin targets B-cell lymphoma 2 (Bcl2) antiapoptotic signaling and AKT survival signaling (Granado-Serrano et al., 2006) amongst others (Murakami et al., 2008). In addition to quercetin, plant-derived natural substances like fisetin (Yousefzadeh et al., 2018; Zhu et al., 2017), piperlongumine (PL) and analogues (Liu et al., 2018; Wang et al., 2016; Zhang et al., 2018), and the curcumin analog EF24 (Li et al., 2019; Yang et al., 2013) also have senolytic properties. Like quercetin, fisetin is a flavonoid present in vegetables and fruits and has been argued to be superior to quercetin in clearing senescent cells (Yousefzadeh et al., 2018; Zhu et al., 2017). Fisetin was reported to induce cell death in breast cancer MCF-7 cells in the absence of p53 or caspase 3 via induction of caspase 7 and inhibition of autophagy (Yang et al., 2012). PL is compound isolated from piper species, has been shown to induce apoptosis in senescent cells on its own and to be synergic with navitoclax (ABT-263) (Wang et al., 2016). In senescence studies, PL was shown to increase the amount of reactive oxygen species (ROS) but senescent cell killing was shown to be ROS-independent (Wang et al., 2016). PL and some of its analogues were subsequently shown to destabilize oxidation resistance 1 (OXR1), indicating that impaired antioxidant mechanisms sensitized senescent cells to oxidative stress (Liu et al., 2018; Zhang et al., 2018). E2F4 is a curcumin analogue that improves bioavailability and therapeutic efficacy and was shown to achieve senolytic effects in a ROS-independent manner by reducing the expression of B-cell lymphoma extra-large (Bcl-XL) and Myeloid Leukemia 1 (Mcl1) in senescent cells (Li et al., 2019). E2F4 was found to be synergic with navitoclax, enabling a dose reduction of the later to reduce its toxicity.

Together with DQ, navitoclax is also a popular choice as a senolytic (Chang et al., 2016; Pan et al., 2017; Zhu et al., 2016). Navitoclax is part of compounds that inhibit Bcl-2 family of anti-apoptotic proteins that also include ABT-737 and venetoclax (Yosef et al., 2016), A-115463 (Tao et al., 2014; Zhu et al., 2017), A-1331852 (Wang et al., 2020; Zhu et al., 2017), and obatoclax (Nguyen et al., 2007; Yosef et al., 2016). ABT-737, which is not orally bioavailable, and navitoclax, which is orally bioavailable, target Bcl-2 and Bcl-XL (Oltersdorf et al., 2005; Tse et al., 2008). A dose-limiting toxicity for navitoclax in cancer intervention is that it kills platelets (thrombocytopenia), attributed to Bcl-XL inhibition (Mason et al., 2007; Shoemaker et al., 2006; Zhang et al., 2007). In response, and in the context of cancer treatment, venetoclax was developed to selectively inhibit Bcl-2 (Souers et al., 2013). Finally, A-1155463 and A-1331852 are selective for Bcl-XL (Leverson et al., 2015; Tao et al., 2014).

Other examples of senolytics include HSP90 inhibitors geldanamycin, 17AAD, and 17-DMAG (Fuhrmann-Stroissnigg et al., 2017), the interference peptide FOXO4-DRI (Baar et al., 2017), cardiac glycosides such as ouabain, digoxin (Guerrero et al., 2019; Triana-Martínez et al., 2019), and galactose-modified duocarmycin (GMD) derivatives (Guerrero et al., 2020). In cancer cells, HSP90 has been shown to stabilize pro-survival client proteins such as Akt and Raf-1 as well as shunting apoptotic signaling by sequestering apoptotic protease activating factor 1 (Apaf-1) away from caspase 9 (Basso et al., 2002; Pandey et al., 2000; Schulte et al., 1995). Foxo4-DRI induces p53-dependent cell death of senescent cells selectively by interfering with FOXO4-p53 binding (Baar et al., 2017). Cardiac glycosides are reported to killed senescent cells by inhibiting Na⁺/K⁺ ATPase on the plasma membrane (Guerrero et al., 2019; Triana-Martínez et al., 2019). GMD are pro-drugs used in antibody-directed pro-drug enzyme therapy (ADEPT). ADEPT entails the conjugation of a tumor-specific antibody to an enzyme that can catalyze the activation of a pro-drug into a highly toxic drug (Sharma and Bagshawe, 2016). Tumor-targeting antibodies combined with β galactosidase are designed to activate pro-drugs at the site of tumors. GMD include β galactosidase-activated duocarmycin; which is thought to exert cytotoxicity by causing DNA damage (Boger et al., 2002). Senescence associated β galactosidase (SA-β-Gal), a widely used marker of senescence, reflects increased lysosomal β-gal in senescent cells (Debacq-Chainiaux et al., 2009; Dimri et al., 1995; Kurz et al., 2000; Lee et al., 2006). This was capitalized on to selectively activate GMD in senescent cells which was argued to induce cell death independent of DNA damage associated with hyper-replication (Guerrero et al., 2020).

Nano-based carriers can also be used to increase the efficacy and or selectivity of existing senolytics (Adamczyk-Grochala and Lewinska, 2020). For example, navitoclax was functionalized to galacto-oligosaccharide capped mesoporous silica nanoparticles to condition its release to the presence of SA-β-gal, thereby reducing its off-target toxicity (Agostini et al., 2012; Galiana et al., 2020). Other alternatives are also possible. PZ15227 (PZ) is bispecific proteolysis-targeting chimera (PROTAC) that targets Bcl-XL by including a modified binding moiety of navitoclax linked to a cereblon (CRBN) E3 ligase binding moiety (He et al., 2020). PZ kills senescent cells by eliciting the E3 CRBN-mediated degradation of Bcl-XL. However, in as far as CRBN is poorly expressed in human platelets, PZ-mediated destabilization of Bcl-XL is impaired, resulting in reduced thrombocytopenia.

3. Cell-based therapies.

Senescent cells are cleared by the immune system (senescence immune surveillance) and has been shown to be carried out by natural killer cells (NK), macrophages, and T-cells (Antonangeli et al., 2016; Eggert et al., 2016; lannello et al., 2013; Kang et al., 2011; Krizhanovsky et al., 2008; Lujambio, 2016; Lujambio et al., 2013; Ovadya et al., 2018; Prata et al., 2018; Sagiv et al., 2013, 2016; Sharma et al., 2017; Soriani et al., 2009; Xue et al., 2007). Senescent cells can evade immune surveillance by various mechanisms including up-regulating inhibitory ligands, shedding stimulatory ligands or as a consequence of age-related immunosenescence (Kale et al., 2020; Muñoz et al., 2019; Ovadya et al., 2018; Pereira et al., 2019; Salminen, 2021; Zingoni et al., 2015) (Figure 3). Consequently, boosting senescence immune surveillance is an alternative to senolytic compounds. Additional endogenous mechanisms may be co-opted with the same effect. For example, surface protein DPP4 has been targeted by antibodies to elicit NK-mediated antibody-dependent cellular cytotoxicity (ADCC) (Kim et al., 2017).

Figure 3.

Senescence immune surveillance: SASP-producing senescent cells recruit immune cells that in turn eliminate the senescent cells and terminate the SASP. Loss of senescence immune surveillance could result in the accumulation of senescent cells with age and support of a chronic SASP. Senescent cells are reported to escape from immune surveillance by preventing the activation immune cell receptors that lead to cytotoxic effector functions, recruitment of immunosuppressive cells and by the loss of function of immune cells with aging or immunosenescence.

A recently developed cell-based alternative to senolytic compounds is the use of chimeric antigen receptor T-cells (CAR-T) (Figure 2C). At their core, CAR-T consist of an extracellular antigen recognition domain that targets cell surface ligands (Rafiq et al., 2019). Binding to its target triggers the intracellular T-cell receptor CD3 zeta domain and the cytotoxic activity of the T cell, which kills the cell with the target ligand. Hence, by targeting membrane-bound ligands that are only present on senescent cells, a CAR-T can be programmed to kill senescent cells. Indeed, a CAR-T designed to target urokinase-type plasminogen activator receptor (uPAR) was recently used to eliminate senescent cells and ameliorate senescence-associated pathologies in mouse models of lung adenocarcinoma and liver fibrosis (Amor et al., 2020). Noteworthy, CAR-NK (Kale et al., 2020) or NKs with engineered t-cell receptors (TCR) (Mensali et al., 2019; Parlar et al., 2019) can also be used for this purpose.

4. Evidence for senescent cell accumulation in the brain: effects of genetic ablation or senolytics as therapies for age-related neurodegenerative diseases

Several lines of evidence indicated the accumulation of senescent cells in both aging and diseased brains (Baker and Petersen, 2018; Florence C. C. Tan, Emmette R. Hutchison, 2014; Kritsilis et al., 2018; Martínez-Cué and Rueda, 2020; Saez-Atienzar and Masliah, 2020; Walton et al., 2020; Walton and Andersen, 2019). Many of these studies have reported selective age-related senescence within brain cells as a consequence of stressors including increased oxidative stress, mitochondrial dysfunction, and altered protein homeostasis. One of the most consistent changes observed in the aging brain is increased expression of GFAP (glial fibrillary acid protein), a major intermediate filament protein expressed primarily within astrocytes and neural stem cells (NSCs), which correlates with the increased appearance of cells with a flat senescent morphology (Nichols et al., 1993). This data suggest that astrocytes and NSCs likely senesce with age, thereby impacting on the function of neighboring neurons (Finch et al., 2003). Bhat et al. subsequently demonstrated the increased presence of senescent astrocytes in post-mortem brain tissues during aging which was further elevated in tissues from Alzheimer’s (AD) patients versus age-matched controls (Bhat et al., 2012). Neurons have also been suggested to be capable of developing a senescence-like phenotype. An initial publication by the von Zglinicki group demonstrated a “senescence-like phenotype” in post-mitotic cells including neurons in the aged mouse brain, suggesting that senescence-like neurons are present and may contribute to brain aging (Jurk et al., 2012).

Studies from our own group interrogating post-mortem tissues from sporadic Parkinson’s disease (PD) patient brains demonstrated an increased expression of several senescent markers, particularly within astrocytes (Chinta et al., 2018). We further show that administration of the environmental toxin paraquat (PQ) increases oxidative stress in the murine brain, resulting in in vivo induction of astrocytic cellular senescence. Genetic elimination of senescent cells utilizing an inducible p16-3MR transgenic mouse model was found to improved age-related Parkinson’s phenotypes including motor deficits, selective dopaminergic midbrain neuron loss, and reduced neurogenesis associated with systemic PQ exposure. This was the first demonstration to our knowledge that senescent cells causatively contribute to neurodegeneration in an in vivo neurodegenerative disease model (Chinta et al., 2018).

In a subsequent study, using a hybrid transgenic mouse model--a cross of INK-ATTAC--apoptosis through targeted activation of caspase 8--with the MAPT P301S PS19 mouse model of tau-dependent neurodegenerative disease, Bussian et al. showed that overexpression of mutated Tau protein induced cellular senescence in conjunction with increased neurodegeneration and loss of cognitive function (Bussian et al., 2018). Increased expression of senescence markers was found to precede the onset of neurofibrillary tangles (NFT) deposition in this mouse model. FACS analysis of brain tissue demonstrated that senescence markers were significantly elevated in astrocytes and microglia, but not in neurons. Notably, the selective inducible genetic ablation of these senescent cells prevented gliosis, neurofibrillary tangle (NFT) deposition, and degeneration of cortical and hippocampal neurons and preserved cognitive function, suggesting that senescence may play a causative role in tau-mediated disease pathology. Use of the senolytic navitoclax showed similar effects.

The potential of accumulation of senescent cells to contribute to AD was also recently explored by Zhang et al. Here they reported increased accumulation of senescent oligodendrocytic precursor cells (OPC) in both the brains of patients with AD and in the amyloid beta APP/PS1 mouse model of AD (Zhang et al., 2019). Senescent OPCs exhibited increased expression of senescence markers in association with amyloid plaques in the brains of both AD patients and mice. In this study, the selective clearance of senescent OPC using combined treatment with DQ significantly attenuated neuroinflammation and cognitive deficits. In this study, the authors reported that they observed no evidence for glial (astrocytic or microglial) senescence.

Utilizing various AD transgenic mouse models, the Orr group reported that accumulation of neurofibrillary tangles (NFTs), but not Aβ plaques, induced a senescence-like phenotype in the aging brain (Musi et al., 2018). Studies with post-mortem brain tissue from progressive supranuclear palsy (PSP) patients also indicated increased expression of the senescent marker p16, directly correlated with brain atrophy and NFT burden. Intermittent treatment of Tau transgenic mice already displaying late-stage pathology with combined DQ significantly attenuated neuronal loss, reduced NFT density, and ventricular enlargement.

Obesity is recognized as a significant public health problem associated with a range of neurodegenerative and psychiatric disorders including depression and anxiety-like behavior. Recent studies from Jurk’s group showed that high-fat diet feeding in mice increased the accumulation of senescent glial cells, coupled with reduced neurogenesis and increased anxiety-like behavior (Ogrodnik et al., 2019). These senescent glial cells were found to accumulate excessive fat deposits; a phenotype termed “accumulation of lipids in senescence” (ALISE). The selective elimination of these senescent cells via both genetic (AP20187) and pharmacological DQ approaches significantly decreased obesity-induced anxiety-like behavior and restored adult neurogenesis.

Whole-brain radiation therapy (WBRT) is a standard treatment for patients with brain metastases and is known to have side-effects including progressive cognitive decline (Yong Woo Lee and Sonntag, 2012). Recent studies by Yabluchanskiy et al. demonstrated that WBRT treatment induces astrocytic cellular senescence in irradiated p16-3MR transgenic mice. To establish a causal relationship between whole-brain radiation (WBI) induced astrocytic cellular senescence and cognitive decline, the authors selectively eliminated senescent cells using navitoclax in whole-brain irradiated mice. The elimination of senescent cells significantly improved cognitive performance in these mice (Yabluchanskiy et al., 2020). In another study, Ogrodnik et al. demonstrated that WBI induced an age-dependent increase in expression of the senescence marker p16Ink4a, specifically in microglia and oligodendrocyte progenitor cells (Ogrodnik et al., 2021). Using both genetic deletion and DQ, the authors further demonstrated that selective clearance of these senescent cells significantly attenuated age-related brain inflammation and cognitive impairment in mice, indicating the therapeutic potential of the senolytic approach for age-associated diseases.

In addition to genetic and pharmacological approaches, recently several studies have demonstrated a critical role for the immune system in the selective clearance of senescent cells in vivo. For example, immune cells such as macrophages and NKs have been recently implicated in removing senescent cells during embryogenesis and aging (Antonangeli et al., 2019). Jin et al. demonstrated that neuroblasts within the aged dentate gyrus exhibit an increased expression of senescence markers and SASP factors leading to impaired neurogenesis and cognition in conjunction with an increased accumulation of NK cells (Jin et al., 2020). Selective depletion of NK cells via either genetic or antibody-mediated approaches significantly improved neurogenesis and brain function, demonstrating the therapeutic potential of targeting NK cells to enhance age-associated neuropathologies.

5. Potential challenges for the use of senolytics to treat brain disorders

Although senolytics have now been shown to improve brain pathologies in context of several mouse models, their translation for use in the clinic holds several potential challenges. Several of these newly developed senolytics may not easily cross the blood-brain barrier. Because of brain complexity, the compounds may also have differential effects in various brain cell types. Importantly, as the senescence phenomenon acts as a double-edged sword with both beneficial and detrimental effects, continuous elimination of these senescent cells via senolytics may have adverse effects (Palmer et al., 2021; Walton et al., 2020).

5.1. There is no universal senolytic

Cellular senescence and the SASP are highly heterogeneous (Basisty et al., 2020; Gorgoulis et al., 2019). There is no single marker for senescence or the SASP, which makes it necessary to assess multiple markers to identify senescent phenotypes. As described above, the heterogeneity of phenotypes extends to the survival pathways that are critical to sustain the viability of the senescent cells. As of yet, there is no universal target that can be drugged/attacked to kill all senescent cell types. Consequently, senescent cell killing should be considered in the context of the specific mechanism of action of the senolytic compound and the specific cell type. When senolytics are used in CNS disease, it is important to keep this in mind and understand which survival pathways are being disabled by a particular senolytic.

As discussed above, there are to date only a handful of CNS studies targeting Bcl2 antiapoptotic proteins with navitoclax (Bussian et al., 2018; Yabluchanskiy et al., 2020) versus those targeting senescent cells using DQ (Musi et al., 2018; Zhang et al., 2019). There is no consensus based on these (or genetic ablation studies) as to whether senolytics kill neurons (Musi et al., 2018), astrocytes and microglia (Bussian et al., 2018; Yabluchanskiy et al., 2020), or oligodendrocyte precursor cells (OPC) (Zhang et al., 2019). There are two important considerations that may explain these observed differences. First, the studies use different transgenic mouse models of neurodegeneration that can potentially differentially induce senescence dependent on stressor and cell type (Walton et al., 2020). Second, the use of different senolytics may have resulted in killing of different senescent cell types (see Table 1). Specifically, Musi and colleagues (Musi et al., 2018) used DQ to kill senescent-like neurons in a tau model of frontotemporal dementia (FTD), Zhang and colleagues (Zhang et al., 2019) used DQ to kill OPC in an APP model of AD, and Yabluchanskiy and colleagues (Yabluchanskiy et al., 2020) used DQ in a model of radiation-induced senescence. Hence, the differential selectivity of DQ may be explained in part by the mouse model itself (Walton et al., 2020). However, albeit not using the exact same transgenic strain, Musi and colleagues (Musi et al., 2018) and Bussian and colleagues (Bussian et al., 2018) both used tau-based transgenic models of FTD. Nevertheless, the former used DQ to kill senescent-like neurons and the latter used navitoclax as well as genetic deletion to kill senescent glial cells. Rather than representing contradictory results, these studies may imply that senescent-like neurons are more susceptible to inhibition of PI3K/Akt and tyrosine kinases by DQ, whilst senescent astrocytes and microglia could be more vulnerable to the targeting of Bcl-2 anti-apoptotic family proteins. Furthermore, senescence may rely on p16 signaling in astrocytes but not in neurons; Bussian and colleagues (Bussian et al., 2018) as well as Yabluchankiy and colleagues (Yabluchanskiy et al., 2020) use an inducible p16-based mouse model to genetically eliminate senescent astrocytes.

5.2. Senolytic selectivity depends on the experimental conditions

The survival pathways targeted by senolytics are not exclusive to senescent cells. A paradigmatic example is navitoclax. ON-target Bcl-XL inhibition by navitoclax results in OFF-target killing of platelets (Mason et al., 2007), with thrombocytopenia being a primary dose-limiting toxicity (Wilson et al., 2010). Absolute selectivity can only be possible if there is a survival pathway that is only used by senescent cells. This does not seem to be the case, at least in vitro, as senolytics seem to invariably kill non-senescent cells at higher doses (Table 1). At doses innocuous to non-senescent cells, the killing of senescent cells is often dramatically reduced. In this regard, senolytics are akin to chemotherapeutic agents; this should come as no surprise that several compounds used as senolytics were originally studied as oncolytics (Murakami et al., 2008; Trepel et al., 2010; Wilson et al., 2010). However, relative to cancer cells, the slow accumulation of senescent cells can be capitalized to devise senolytic intervention schedules that, in time, can result in high efficacy and selectivity, foror example, by local and repeated acute senolytic administration (Jeon et al., 2017). The importance of a scheduled intervention with minimal OFF-target killing is particularly important for organs that cannot replenish their cells. That is, organs in which the killing of non-senescent cells progressively and permanently depletes the organ of healthy cells.

5.3. Postmitotic neurons may constitute a dangerous target

Adult neurogenesis is not a CNS-wide regenerative mechanism, but one that replenishes cells in the olfactory bulb and the dentate gyrus of the hippocampus (Kempermann et al., 2018). Of note, it is neural precursor cells (NPC), not neurons, that proliferate and later differentiate into neurons and their ability to restore overall brain function has been shown to be limited. As neurons do not proliferate, if killed by senolytics they are permanently lost. There is currently a slowly increasing consensus that neurons can indeed become senescent or ‘senescent-like’, although data that originally supported this concept was published nearly 10 years ago (Jurk et al., 2012). Regardless of whether neurons senesce or not, determining whether neurons are killed by senolytics may be important in terms of their potential use as a CNS intervention. That is not to say, senescent-like neuron killing may be therapeutic in some conditions (Musi et al., 2018), however it is important to consider the impact of removing these largely irreplaceable post-mitotic cells, including the stage at which they may be removed without having a major impact on overall brain function as well as possible drug side-effects (Walton et al., 2020; Walton and Andersen, 2019).

5.4. Boosting senescence immune surveillance as an alternative to senolytic therapies

The brain is under immune privilege (Benakis et al., 2018; Galea et al., 2007; Korin et al., 2017). Immune cells such as monocytes, peripheral macrophages, CD8⁺ cytotoxic T-cells, CD4⁺ T-cells and NK cannot access the parenchyma of the healthy unaged brain. Nevertheless, there is a progressive infiltration of immune cells with age and which is further elevated in certain CNS pathologies including AD (Gorlé et al., 2016; Nation et al., 2019; Poli et al., 2013; Sweeney et al., 2018). This is potentially owed to increasing BBB permeability and or brain “inflammaging” which may be partly driven by pro-inflammatory factors of the SASP. Brain inflammation can act as a chemo-attractant that may make it possible to use senolytic NK, macrophages, or CAR-T as brain therapeutics, although one must always consider the potential downside of killing neurons or oligodendrocytes. NK-mediated killing of neurons after ischemia or oligodendrocyte killing by the immune system in multiple sclerosis (MS) are clear cautionary examples. Importantly, NKs and especially CAR-T are not only lytic cells, but also pro-inflammatory. Hence, these may elicit unwanted increases in neuroinflammation and the immune response. Along these lines, CAR-T therapies often result in cytokine release syndrome (CRS) and/or neurotoxicity (Bonifant et al., 2016), a trade-off that may be acceptable in cancer, but severely limits their use in the brain.

5.5. CNS-targeting therapies need not target the CNS

Another aspect of senolytic intervention in the context of CNS pathology is the possibility that senolytics are achieving their effects by killing senescent cells outside of the CNS. As noted, senolytics are known to ameliorate disease progression in normal and accelerated aging mouse models (Baker et al., 2016, 2011). That is, senolytics already improve overall health. It has yet to be addressed whether senolytics are acting by countering neurodegeneration in transgenic mouse models in part or exclusively by eliminating non-CNS cells. Identifying potential targets outside of the CNS may afford new insights into CNS diseases. This could make BBB-impermeable drugs invaluable in future research in CNS pathology as they may constitute a safer approach than targeting senescent CNS cells directly.

6. Conclusions:

There is a rapidly growing body of evidence for an involvement of senescent cells in CNS pathology (Florence C. C. Tan, Emmette R. Hutchison, 2014; Saez-Atienzar and Masliah, 2020; Walton et al., 2020; Walton and Andersen, 2019), but also a lack of agreement on whether one particular cell type or several cell types that are the culprit. This may be explained by the use of different transgenic models which use different stressors to elicit neuropathology as well as the use of different senolytic compounds which are known to be variable in their effectiveness in killing senescent cells. Many additional senolytic strategies are becoming available, including those that have not yet been tested in the context of CNS pathology, as well as alternative avenues capitalizing on immune cells. In moving forward, it will be important to accurately ascertain which cells undergo senescence in the CNS and contribute to a given disease-related stress and which senolytic strategy is the most effective in killing culprit senescent cells whilst sparing non-senescent cells. In case there is a therapeutic advantage to their killing, the post-mitotic status of neurons should be carefully considered. If not, the emphasis should arguably be placed on determining whether there is OFF-target killing of neurons. In this regard, the lack of a ‘universal’ senolytic is likely beneficial--our inability to kill all senescent cells with a single senolytic compound may allow us to selectively kill harmful senescent cells whilst sparing non-harmful senescent cells in a disease-dependent fashion. If a therapeutic effect in the CNS can be derived from the elimination of senescent cells in the periphery, senolytics that cannot cross the BBB can also prove highly beneficial.

Highlights.

• Brain senescent cells are involved in CNS pathology associated with both aging and neurodegenerative diseases, although the specific type of cells that undergo senescence in these conditions is not yet well-defined.

• The killing effect of senolytics is dependent on the type of senolytic and the target cell.

• Senolytics do not selectively but preferentially kill senescent cells over healthy cells.

• To identify the most clinically useful senolytics in the context of aging and/or neurodegenerative disease, a better understanding of optimal dosage and the senolytic pathway and specific cell target involved will be essential.

Abbreviations

ALISE

Accumulation of lipids in senescence

ADEPT

antibody-directed prodrug enzyme therapy

AD

Alzheimer’s disease

Apaf-1

apoptotic protease activating factor 1

Bcl-2

B-cell lymphoma 2

Bcl-XL

B-cell lymphoma extra-large

Bleo

Bleomycin

Bleo (IV)

Bleomycin administered in vivo

BMDMC

Bone Marrow Derived Mesenchymal Cells

CAR-T

chimeric antigen receptor T-cells

CPC

Cardiac Progenitor Cells

CRBN

cereblon

CRS

Cytokine release syndrome

Doxo

Doxorubicin

DQ

Dasatinib and Quercetin

DTR

diphtheria toxin receptor

Eto

Etoposide

Ercc1^−/Δ”

mouse model of human progeroid syndrome

FDA

Food and Drug Administration

FKB-Casp8

FK506 binding protein-caspase 8

FT2AEC

Fibrotic Type II Alveolar Epithelial Cells

FTD

Frontotemporal dementia

GCV

ganciclovir

GMD

galactose-modified duocarmycin

HSV-TK

truncated herpes simplex virus thymidine kinase

HUVEC

human umbilical endothelial cell

IR

Ionizing radiation

Mcl1

Myeloid Leukemia 1

MS

Multiple sclerosis

MEF

mouse embryonic fibroblast

MSC

Mesenchymal Stem Cells

NFT

Neurofibrillary tangles

NK

Natural Killer Cell

NPC

Neural precursor cells

NSC

Neural stem cell

OPC

Oligodendrocyte Precursor Cells

OS

Oxidative Stress

OXR1

oxidation resistance 1

PA

Preadipocytes

PD

Parkinson’s disease

PL

piperlongumine

PQ

Paraquat

PROTAC

bispecific proteolysis targeting chimera

PSP

Progressive supranuclear palsy

Raf-1

Raf1 proto-oncogene, serine/threonine kinase

REC

Renal Epithelial Cells

ROS

reactive oxygen species

RS

Replicative Senescence

SA-β-Gal

senescence associated β galactosidase

SASP

Senescence-associated secretory phenotype

T2AEC

Type II Alveolar Epithelial Cells

uPAR

urokinase-type plasminogen activator receptor

WBRT

Whole-brain radiation therapy