The Senolytic Drug JQ1 Removes Senescent Cells via Ferroptosis

Abstract

BACKGROUND

Ferroptosis is an iron-dependent, non-apoptotic programmed cell death. Cellular senescence contributes to aging and various age-related diseases through the expression of a senescence-associated secretory phenotype (SASP). Senescent cells are often resistant to ferroptosis via increased ferritin and impaired ferritinophagy. In this study, we investigated whether treatment with JQ1 could remove senescent cells by inducing ferroptosis.

METHODS

Senescence of human dermal fibroblasts was induced in vitro by treating the cells with bleomycin. The senolytic effects of JQ1 were evaluated using a SA-β gal assay, annexin V analysis, cell counting kit-8 assay, and qRT-PCR. Ferroptosis following JQ1 treatment was evaluated with qRT-PCR and BODIPY staining.

RESULTS

At a certain range of JQ1 concentrations, JQ1 treatment reduced the viability of bleomycin-treated cells (senescent cells) but did not reduce that of untreated cells (non-senescent cells), indicating that JQ1 treatment can selectively eliminate senescent cells. JQ1 treatment also decreased SASP expression only in senescent cells. Subsequently, JQ1 treatment reduced the expression of ferroptosis-resistance genes in senescent cells. JQ1 treatment induced lipid peroxidation in senescent cells but not in non-senescent cells.

CONCLUSION

The data indicate that JQ1 can eliminate senescent cells via ferroptosis. This study suggests ferroptosis as a new mechanism of senolytic therapy.

Introduction

Ferroptosis is a newly discovered iron-dependent form of non-apoptotic cell death that is distinct from other forms of cell death [1, 2]. Ferroptosis is mainly caused by a redox imbalance, which is driven by the depletion of glutathione (GSH) or inactivation of the antioxidant enzyme glutathione peroxidase (GPX4) [3, 4]. Consequently, ferroptosis is accompanied by considerable iron accumulation and lipid peroxidation in cells, ultimately leading to cell death [1, 5, 6]. Recently, many studies have shown excess iron accumulation in aged somatic tissues [7, 8] and pathophysiological evidence linking ferroptosis and age-related diseases, such as neurodegenerative diseases, cardiomyopathy, and tumors [2, 9–11]. Thus, ferroptosis has begun to receive attention in the prognosis and treatment of various degenerative disease [12–14].

Cellular senescence is one of the hallmarks of aging [15], and the accumulation of senescent cells in tissues is affected in many age-related degenerative diseases, including tumors, arthritis [16], Alzheimer’s disease [17], diabetes [18], heart diseases [19, 20], atherosclerosis [21], and various other conditions [22–25]. Cellular senescence occurs when cells are exposed to stress, such as genomic instability, telomere attrition, mitochondrial dysfunction, and oxidative stress caused by reactive oxygen species (ROS) [26]. Senescent cells are characterized by loss of proliferative capacity, resistance to apoptosis, and the transcription and secretion of many proinflammatory cytokines, growth factors, and chemokines, termed the senescence-associated secretory phenotype (SASP) [27]. Due to SASP secretion by senescent cells, the tissue microenvironment becomes inflammatory and is damaged by extracellular matrix (ECM) degradation.

Senolytic drugs, which are agents that selectively induce the death of senescent cells and their SASPs, have become highly potent drugs in treating age-associated diseases [28]. Over the years, animal studies and clinical trials have shown that senolytic therapies reduce the burden of aging, prevent various age-related disease treatments and extend the median lifespan [18, 22, 29–33]. There are many drugs that induce ferroptosis in a wide range of cells, but these drugs are rarely used as senolytic drugs because senescent cells are highly resistant to ferroptosis [34]. Thus, in senescent cells, the amount of intracellular iron stored in ferritin, an iron-storage protein, is increased. Additionally, ferritinophagy, an autophagic degradation of ferritin, is limited in senescent cells. Taken together, due to iron accumulation and impaired ferritinophagy, ferroptosis occurs less frequently in senescent cells than in non-senescent (normal) cells.

JQ1, an anticancer drug, is a bromodomain and extraterminal domain (BET) family inhibitor that blocks bromodomain-containing protein 4 (BRD4) and other ferroptosis-resistance proteins, such as GPX4 and SLC7A11 [35, 36]. Thus, JQ1 treatment activates ferritinophagy in cells by inducing ferritin degradation and iron accumulation in the cytosol, followed by the Fenton reaction and subsequent ROS-induced lipid peroxidation. Consequently, treatment with JQ1 activates ferroptosis and induces apoptosis in cells, leading to the removal of cells that do not undergo apoptosis, such as senescent cells.

In this study, we investigated whether JQ1 can remove senescent cells via ferroptosis (Fig. 1). To test this hypothesis, we employed a well-established method to induce senescence, which is use of bleomycin [37]. We investigated whether treatment with JQ1 increases lipid peroxidation, which is a ferroptosis characteristic, and downregulates ferroptosis resistance gene expression in senescent cells. Next, we investigated whether JQ1 treatment effectively removes senescent cells and suppresses SASP expression through the ferroptotic death of senescent cells.

Fig. 1.

Mechanisms of ferroptosis induction in senescent cells by JQ1. Intracellular iron in senescent cells accumulates in the form of ferritin through cellular senescence pathway. JQ1 treatment to senescent cells induces ferritin degradation and subsequent ferroptosis

Materials and methods

Cell culture and senescence induction

Human dermal fibroblasts (HDFs) (PCS-201–012, ATCC, passages 4–6) were cultured at 37 °C and 5% CO₂ in Dulbecco’s Modified Eagle’s Medium (DMEM, Gibco, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (FBS, Gibco) and 1% penicillin/streptomycin (P/S, Gibco). When the cell confluency reached 80–90%, the culture medium was replaced with DMEM with no serum, 1% P/S, and 50 μg/ml bleomycin (#13,877, Cayman Chemical, Ann Arbor, MI, USA) for 2 days to induce senescence.

Senescence-associated β-galactosidase (SA-β gal) assay

SA-β gal assays were performed using a senescence beta-galactosidase staining kit (#9860, Cell Signaling Technology, Danvers, MA, USA) according to the manufacturer’s instructions. Senescent HDFs stained blue were observed by light microscopy, and images were captured. The number of senescent cells/non-senescent HDFs was counted and divided by the area (n = 3).

Cell viability

Non-senescent and senescent HDFs were seeded on 48-well plates. The HDFs were treated with vehicle (DMSO), 35 μM JQ1 (#HY-13030, MedChemExpress, Monmouth Junction, NJ, USA) or 5 μM ABT-263 (#S1001, Selleckchem, Houston, TX, USA) for 48 h (n = 3). In cells co-treated with fer-1, vehicle (DMSO), 35 μM JQ1 and 3.5 μM fer-1 (#HY-100579, MedChemExpress) were administered for 48 h (n = 4). The number of viable cells was assessed using the cell counting kit-8 (CCK-8) assay (DoGenBio, Seoul, Korea) according to the manufacturer’s instructions. The absorbance was measured using a Powerwave X340 microplate reader (BIO-TEK Instruments, Winooski, VT, USA) at 450 nm. The number of viable drug-treated cells was expressed relative to the number of untreated non-senescent and senescent cells.

Annexin V apoptosis analysis by FACS

Non-senescent and senescent HDFs were treated with vehicle (DMSO) or JQ1 (0–70 μM) for 48 h in 24-well plates (n = 3). After 48 h of treatment, the HDFs were detached using trypsin–EDTA and washed twice with phosphate-buffered saline (PBS, pH 7.4). The HDFs were then resuspended in annexin V binding buffer (#422,201, Biolegend, San Diego, CA, USA) and incubated with APC Annexin V (#640,920, Biolegend) and PI (#421,301, Biolegend) for 30 min at room temperature. The mixtures were measured using a Becton Dickinson FACS Canto-II flow cytometer (BD Biosciences, Bedford, MA, USA).

RNA extraction and quantitative real-time PCR

To analyze the mRNA levels of ferritin heavy chain (FTH) and ferritin light chain (FTL), HDFs were treated with bleomycin (50 μg/ml) in serum-free media for 0, 24, 48, and 72 h (n = 4) or HDFs were treated with bleomycin for 48 h and followed by treatment with or without JQ1 for 48 h (n = 3). For BRD4, SLC7A11, GPX4, Nrf2, and p53 mRNA expression, non-senescent and senescent HDFs were treated with JQ1 (35 μM) for 24 h or vehicle (DMSO) (n = 4). For p21, IL1β, IL6, TNF alpha, and MMP9 expression, non-senescent and senescent HDFs were treated with JQ1 (35 μM) for 48 h or vehicle (DMSO) (n = 4). Total RNA was prepared using QIAzol Lysis Reagent (Qiagen, Valencia, CA, USA) according to the manufacturer’s protocol. RNA concentrations were determined by a NanoDrop spectrometer (ND-2000, NanoDrop Technologies, Wilmington, DE, USA). Five hundred nanograms of total RNA was reverse transcribed into cDNA using PCR PreMix (Bioneer, Daejeon, Korea) according to the manufacturer’s protocol. Quantitative real-time PCR (qRT-PCR) reactions were performed using SYBR green-based reagent with TOPreal™ qPCR 2X PreMIX (Enzynomics, Daejeon, Korea). Cycling conditions consisted of initial denaturation at 95 °C for 15 min, followed by 55 cycles at 95 °C for 10 s, 60 °C for 15 s, and 72 °C for 30 s. Gene expression was normalized by GAPDH expression. The primers used in this study are listed in (Table 1).

Table 1.

Primers sequences used for qRT-PCR

Gene	Forward (5′-3′)	Backward (5′-3′)
FTH	TCC TAC GTT TAC CTG TCC ATG T	GTT TGT GCA GTT CCA GTA GTG A
FTL	CAG CCT GGT CAA TTT CTA CCT	GCC AAT TGC CGG AAG AAG TG
BRD4	AAC CTG GCG TTT CCA CGG TA	GCC TGC ACA GGA GGA GGA TT
SLC7A11	ATG CAG TGG CAG TGA CCT TT	GGC AAC AAA GAT CGG AAC TG
GPX4	AGA GAT CAA AGA GTT CGC CGC	TCT TCA TCC ACT TCC ACA GCG
Nrf2	CAC ATC CAG TCA GAA ACC AGT GG	GGA ATG TCT GCG CCA AAA GCT G
p53	CCT CAG CAT CTT ATC CGA GTG G	GG ATG GTG GTA CAG TCA GAG C
p21	CCT CAT CCC GTG TTC TCC TTT TCC T	GTA CCA CCC AGC GGA CAA GTG GG
IL1β	GGA CAA GCT GAG GAA GAT GCT GGT T	TCG TTA TCC CAT GTG TCG AAG AAG A
IL6	CTC AAT ATT AGA GTC TCA ACG CCC A	GAG AAG GCA ACT GGA CCG AA
MMP9	CGT GAT TGA CGA CGC CTT TG	CTG TAC ACG CGA GTG AAG GT
TNF-alpha	TGC ACT TTG GAG TGA TCG GC	ACT CGG GGT TCG AGA AGA TG
GAPDH	TGC ACC ACC AAC TGC TTA GC	GGC ATG GAC TGT GGT CAT GAG

Lipid peroxidation

A fluorescent dye, BODIPY 581/891 (#27,086, Cayman Chemical, MI, USA), was used to detect the levels of lipid peroxidation induced by ROS. Both non-senescent and senescent HDFs were treated with vehicle (DMSO) or JQ1 (35 μM) for 12 h to induce lipid peroxidation (n = 4). After washing 2 times with PBS, the samples were treated with BODIPY 581/891 (2 μM) for 30 min. After washing 2 times, fluorescence images were captured by fluorescence microscopy (Olympus, Tokyo, Japan). Mean fluorescence intensities from BODIPY-treated samples were measured using a Becton Dickinson FACS Canto-II flow cytometer (BD Biosciences) after detaching by trypsin–EDTA and resuspending in PBS.

Statistical analysis

All statistical analyses were performed using Prism 8.0.2 (GraphPad Software Inc., La Jolla, CA, USA). All quantitative data are expressed as the means ± standard deviation. For the comparison of two groups, a two-tailed Student’s t-test was used. For comparisons among more than three groups, one-way analysis of variance (ANOVA) followed by Tukey’s test was used. A difference with a p < 0.05 was considered statistically significant.

Results

Ferritin upregulation in bleomycin-induced senescent cells

We studied a well-established model of cellular senescence HDFs induced by bleomycin (which induced free radical production in contact with oxygen, leading to double-strand DNA breaks). When cultured for 2 days, approximately 50% of bleomycin-treated HDFs stained positive for SA-β gal activity, which also had a flat-cell phenotype, compared with DMSO-treated controls (Fig. 2A). Bleomycin-induced senescent cells showed elevated levels of ferritin heavy chain (FTH) and ferritin light chain (FTL), a component of the ferritin iron storage complex, at time points later than 24 h (Fig. 2B). In addition, HDFs treated with bleomycin for 48 h and followed by treatment with JQ1 for 48 h showed decreased mRNA expressions of FTH and FTL (Fig. 2C). Thus, bleomycin-induced cellular senescence upregulated ferritins, similar to other senescence types. JQ1 treatment downregulated expressions of ferritins in senescent HDFs.

Fig. 2.

Cellular senescence upregulates ferritin expression. A SA-β gal staining of HDFs cultured DMSO or bleomycin (50 μg/ml) treatment. Scale bars 10 μm. mRNA expressions of ferritin heavy chain (FTH) and ferritin light chain (FTL) in B HDFs treated with bleomycin for 0, 24, 48, and 72 h and C HDFs treated with bleomycin for 48 h and followed by treatment with or without JQ1 (35 μM) for 48 h. Statistical significance was calculated by Student’s t-tests A or one-way analysis of variance (ANOVA) with Tukey’s significant difference multiple comparisons B. *p < 0.05, **p < 0.01, ns means no significant difference. All values are mean ± SD

Senolytic effect of JQ1

To evaluate the effect of senescent cell clearance on ferritin accumulation, we first tested senolytic JQ1, which was recently found to specifically induce cell death in senescent cells. First, we optimized the concentration of JQ1 by treating non-senescent cells and bleomycin-induced senescent cells with different amounts of JQ1 (0 μM (untreated) 17.5 μM, 35 μM, 52.5 μM, 70 μM) for 48 h. Preliminary screening by annexin V-propidium iodide (PI) apoptosis staining showed that treatment with JQ1 decreased the viable cell number and increased the annexin V-positive dead cell percentage (Fig. 3A). However, after treatment with more than 35 μM JQ1, the non-senescent cells were also affected by the drug, so we used 35 μM JQ1 for further study. To further confirm the senolytic effect of JQ1 on bleomycin-induced senescent cells, JQ1′s senolytic activities were compared to those of the previously reported senolytic drug ABT-263 (an inhibitor of the anti-apoptotic protein Bcl2) [38, 39]. JQ1 showed senolytic activity similar to that of ABT-263, killing more than 50% of bleomycin-induced senescent cells (Fig. 3B). JQ1 treatment also reduced SA-β gal-positive senescent cells by approximately 50% compared with untreated senescent cells, suggesting that JQ1 selectively eliminated senescent cells (Fig. 3C).

Fig. 3.

JQ1 selectively removes senescent cells (senolytic effects). A JQ1-mediated induction of apoptosis of senescent HDFs treated with bleomycin, as revealed by staining with Annexin V and PI by FACS analysis. B Senolytic effects of JQ1 and ABT263 on HDFs, as indicated by CCK-8 assay. C SA-β gal staining showing that JQ1 removes senescent cells. Scale bars 10 μm. A-C Statistical significance was calculated by one-way analysis of variance (ANOVA) with Tukey’s significant difference multiple comparisons. **p < 0.01, ns means no significant difference. All values are mean ± SD

JQ1 reduced SASP gene expression in senescent cells

To determine whether JQ1 has sufficient therapeutic efficacy, the mRNA levels of SASP factors in JQ1-treated non-senescent cells and senescent cells were determined by qRT-PCR. Notably, bleomycin-induced senescent HDFs treated with JQ1 significantly had reduced levels of p21, a frequently used senescence marker, and SASP factors IL1β, IL6, TNF alpha, and MMP9 relative to DMSO-treated bleomycin-induced senescent HDFs. JQ1 itself did not affect the mRNA levels of the SASP factors of non-senescent HDF gene expression. The data suggested that JQ1 downregulated SASP by removing senescent cells, not non-senescent cells (Fig. 4).

Fig. 4.

JQ1 reduces SASP gene expressions in senescent cells. SASP-associated gene expression by qRT-PCR analysis. Statistical significance was calculated by one-way analysis of variance (ANOVA) with Tukey’s significant difference multiple comparisons. **p < 0.01, ns means no significant difference. All values are mean ± SD

JQ1 regulated ferroptosis-related gene expression in senescent cells

To substantiate whether JQ1 kills senescent cells through ferroptosis, ferroptosis-related gene expression was measured by qRT-PCR in non-senescent cells and bleomycin-induced senescent cells treated with JQ1. The ferroptosis resistance genes BRD4, GPX4, SLC7A11, and Nrf2 were downregulated in senescent cells treated with JQ1 but not in non-senescent cells (Fig. 5A, B). Interestingly, p53, known to enhance ferroptosis by targeting GPX4, SLC7A11, and Nrf2 [40], was upregulated by JQ1 treatment in senescent cells only. Our results suggest that JQ1 can clear senescent cells by activating the ferroptotic pathway.

Fig. 5.

Changes in ferroptosis-related gene expression following JQ1 treatment by qRT-PCR analysis. A JQ1 treatment to bleomycin-treated HDFs (senescent cells) downregulated ferroptosis-resistance gene expressions and upregulated p53 expression. B JQ1 treatment to non-senescent cells induced increase or no change in the same gene expressions. A, B Statistical significance was calculated by Student’s t-tests. *p < 0.05, **p < 0.01, ns means no significant difference. All values are mean ± SD

JQ1 induced lipid peroxidation in senescent cells

Then, we investigated one of the hallmarks of ferroptosis, the levels of lipid peroxidation induced by ROS in bleomycin-induced senescent cells, by BODIPY 581/591 dye. Green fluorescence demonstrated that lipid peroxidation occurred more frequently in senescent cells treated with JQ1 than in non-senescent cells (Fig. 6A). Mean fluorescence intensities of BODIPY staining with FITC light showed the same results quantitatively (Fig. 6B).

Fig. 6.

JQ1 induces lipid peroxidation in senescent cells. A BODIPY staining images and B FACS mean fluorescence intensity showing that JQ1 treatment for 12 h increased lipid peroxidation only in bleomycin-treated HDFs (senescent cells). Green indicates lipid peroxidation. Scale bars 100 μm. Statistical significance was calculated by one-way analysis of variance (ANOVA) with Tukey’s significant difference multiple comparisons. **p < 0.01, ns means no significant difference. All values are mean ± SD

JQ1 eradicated senescent cells through ferroptosis

Because JQ1 eliminated senescent cells by inducing cell death in senescent cells (Fig. 3), we examined whether this removal of senescent cells with JQ1 treatment occurred through ferroptosis by performing a CCK-8 assay. Co-treatment with ferrostatin-1 (fer-1), a ferroptosis inhibitor, and JQ1 showed a lower senescent cell killing effect than JQ1 treatment alone (Fig. 7). The non-senescent cells were not affected by the same concentration of JQ1. This result suggested that the ferroptotic pathway of JQ1 in senescent cells was blocked by fer-1 treatment and that JQ1 killed senescent cells via ferroptosis.

Fig. 7.

JQ1 removes senescent cells via ferroptosis. CCK-8 assay showing that JQ1′s seonlytic effects are suppressed by fer-1, the ferrotosis inhibitor. Statistical significance was calculated by one-way analysis of variance (ANOVA) with Tukey’s significant difference multiple comparisons. *p < 0.05, **p < 0.01, ns means no significant difference. All values are mean ± SD

Discussion

Senolytic treatment has been proposed as a way to prevent or delay age-related diseases [18, 22, 28–33]. Similar to other senolytic cell death types, such as apoptosis, ferroptosis is thought to be a new type of senolytic cell death. However, conventional ferroptosis inducers, such as erastin and ferric ammonium citrate (FAC), do not induce ferroptosis in senescent cells; however, they killed many normal cells [34]. Thus, it is necessary to identify other approaches to induce ferroptosis in senescent cells. Senescent cells have elevated iron levels (tenfold higher than non-senescent cells) [41] and are often resistant to ferroptosis due to interrupted ferritinophagy. To utilize these features of senescent cells, we used JQ1 as a ferroptosis inducer for senolytic treatment.

Senescent cells are resistant to certain types of death, such as apoptosis, and can avoid phagocytosis by macrophages. Therefore, to induce death in senescent cells effectively, several senolytic drugs have been suggested. JQ1 was previously considered an anticancer drug, but similar to other BET family degraders, such as OTX015 and ARV825, JQ1 is known to have a senolytic effect on senescent cells [36, 42]. However, there has been no research on whether ferroptosis occurs in the pathway of BET family degrader senolytic effects. Similarly, no previous study has shown that JQ1′s senolytic pathway is involved in ferroptosis. In this study, we found that JQ1′s senolytic effect occurred through ferroptosis and suggested that other BET family degrader cases would be similar.

We found ferroptosis-associated senolytic effects through several experiments. BODIPY staining and CCK assays using a ferroptosis inhibitor (fer-1) demonstrated that ferroptosis occurred with JQ1 treatment. In RT-PCR analysis, however, the expression of SLC7A11 and p53 had controversial effects on senescence and ferroptosis. In this study, SLC7A11 downregulation and p53 upregulation showed ferroptotic effects on senescent cells, but the effects of SLC7A11 and p53 expression need further investigation.

In the progression of senescence and ferroptosis, abundant phosphatidylserine (PS) becomes inverted into the cell membranes, and the intracellular iron level gradually increases [34, 35]. Therefore, dead cells that underwent senescence and ferroptosis would have increased PS and iron inside. Increased PS in the cell membrane is a similar feature of cells killed via apoptosis [43]. This fact suggests that ferroptotic cells can actively be taken up by immune cells, such as macrophages and dendritic cells, by phagocytosis, leading to regenerative phenotypes, such as M2 macrophages and tolerogenic dendritic cells [44, 45]. This phenomenon would change the surrounding environment to be anti-inflammatory, and damaged tissues could heal, leading to disease treatment. When phagocytosis occurs by immune cells, increased iron is also taken up by immune cells. The effects of iron uptake on immune cells are controversial. Studies have shown that iron uptake by macrophages changes the macrophage phenotype to M1 [46] or M2 [47, 48]. Nevertheless, there have been many studies using intracellular iron for disease treatment [49–51]. Therefore, increased intracellular iron in phagocytic cells that engulf ferroptotic dead bodies of senescent cells may be useful for various therapies.

Cellular senescence and SASP secretion influence adjacent tissues, but these effects are both beneficial and harmful. Among the benefits, growth arrest suppresses cancer, while SASPs fine-tune certain embryonic structures and optimize wound healing in adults [52, 53]. On the negative side, SASPs can cause chronic, low-grade inflammation and age-related pathologies [18–21]. In this context, surrounding cells become inflammatory, and the matrix is degraded via secretion of inflammatory cytokines and matrix metalloproteinases. In this study, we showed that JQ1-treated senescent cells have decreased levels of inflammatory cytokines and matrix metalloproteinases. Therefore, JQ1 can be used to treat age-related diseases.

In summary, we used JQ1 as both a senolytic drug and ferroptosis inducer. In our study, JQ1 potently removed senescent cells via ferroptosis, suggesting that JQ1 may be useful for senolytic therapy.

Compliance with ethical standards

Conflict of interest

The authors declared that they have no conflict of interest.

Ethical statement

There are no animal experiments carried out for this article.