Acid ceramidase promotes senescent cell survival

Cell culture, replicative senescence, and SA-βGal assays

WI-38 human diploid fibroblasts (HDFs) were cultured in Dulbecco's modified Eagle's medium (DMEM, Gibco) supplemented with 20% heat-inactivated fetal bovine serum (FBS, Gibco), antibiotics penicillin and streptomycin (Gibco), and non-essential amino acids (Gibco) in a 5% CO₂ incubator. Primary human coronary artery vascular smooth muscle cells (VSMCs) were obtained as cryopreserved secondary cultures from LifeLine Cell Technology and maintained in VascuLife® SMC Medium Complete Kit from LifeLine Cell Technology following the manufacturer’s protocol.

Proliferating WI-38 cells were used at low population doubling levels (PDL) ranging between PDL20 and PDL24. Replicative senescence was achieved by passaging WI-38 cells in culture until their growth rate decreased indicating that a senescent state had been attained (PDL48 to PDL52). Senescence-associated β-galactosidase (SA-βGal) activity in WI-38 cells was assessed following the manufacturer’s protocol (Cell Signaling Technology). Cellular senescence was quantified using SPiDER-βGal analysis following the manufacturer’s protocol (Dojindo Molecular Technologies). Briefly, cells were seeded overnight at a density of 2.5 × 10⁴ in triplicate in a 96-well plate. The following day, cells were counted using the Cell Count normalization assay following the manufacturer’s protocol (Dojindo Molecular Technologies). Fluorescence was measured using a Perkin Elmer Victor plate reader at excitation 350 nm, emission 461 nm. The same plate was then used for the SPiDER-βGal assay. Cells were lysed at 25° C for 10 min; 50 μl of SPiDER-βGal working solution was added to each well and incubated at 37° C, 5% CO₂ for 60 min; 100 μl of Stop Solution was then added to each well. Fluorescence was measured using a Promega Glomax Explorer at 520 nm excitation, 580-620 nm emission. Normalized SA-βGal activity was calculated by this formula: (Fluorescent intensity of SA-βGal activity) / (Fluorescent intensity of cell number – Fluorescent intensity of blank).

Transfection and treatments

Pre-senescent (PDL38 to PDL45) WI-38 fibroblasts were seeded in triplicate overnight at 2.5×10⁵ cells/well in a 6-well tissue culture plate for ~50% confluence by the next day. Cells were transfected using Lipofectamine 2000 (Thermo Fisher) in 1 ml Optimem (Gibco) following the manufacturer’s protocol. Control siRNA (sc-37007) or ASAH1 siRNA (sc-105032) were used at 35 nM for 48 h. Proliferating (P, PDL20-24) and replicatively senescent (S, PDL48-52) cells were treated with either DMSO (control vehicle) or the inhibitor of RNA polymerase II Actinomycin D (Sigma) (2.5 μg/ml) followed by RNA isolation to test mRNA stability. For ionizing radiation (IR)-induced senescence, proliferating (PDL20-24) WI-38 cells were exposed to 10 Gy and harvested 10 days later. For cycloheximide (CHX) treatments, proliferating (PDL20-24) and replicative senescent (PDL48-52) cells were treated with either DMSO (control vehicle) or 50 μg/ml CHX for 0, 1, 2 or 4 h. For MG132 treatments, proliferating (PDL20-24) and replicative senescent (PDL48-52) cells were treated with either ethanol (control vehicle), or 10 μM MG132 for 0, 2, 4 or 6 h. Cells were treated with (ACi; Calbiochem, ACI IV, BOC, Cat# 533371) at 50 μM for 48 h.

Protein analysis

Cell pellets were washed twice in cold PBS, the supernatant was discarded, and cells were lysed in RIPA lysis buffer (Santa Cruz Biotechnology) containing protease inhibitors (Roche) and incubated on ice for 10 min, followed by sonication for 5 min. The sonicated samples were centrifuged for 10 min at 25° C to remove the insoluble fraction. The remaining supernatant was the whole-cell lysate. Lysates were separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto nitrocellulose membranes (Biorad). Incubations with primary antibodies recognizing ASAH1 (sc-136275), p21 (CDKN1A), p53 (TP53) (from Santa Cruz Biotechnology), as well as an antibody recognizing p16 (CDKN2A, from BD Biosciences), were followed by incubations with the appropriate secondary antibodies conjugated with horseradish peroxidase (GE Healthcare). The levels of proteins GAPDH, β-actin (ACTB), and HSP90 (from Santa Cruz Biotechnology), which displayed unchanged levels, were monitored to control for loading. Signals were developed using Enhanced Chemiluminescence (ECL) and acquired by KwikQuant Imaging System (Kindle Biosciences). Proteomic data are displayed in Supplementary Table 1.

RNA isolation, sequencing, and RT-qPCR analysis

RNA was isolated from WI-38 fibroblasts using Direct-zol RNA kit following the manufacturer’s instructions (Zymo Research). RNA-seq analyses of proliferating and senescent cells (replicative or IR-induced senescence) were performed as described in [28], and the data are available at GEO accession number GSE85771 [28]. Total RNA was used for gene expression analysis by reverse transcription (RT) followed by quantitative (q)PCR analysis. RT was performed by using random hexamers and reverse transcriptase (Maxima Reverse Transcriptase, Thermo Fisher) and qPCR analysis was carried out using gene-specific primers and SYBR green master mix (Kapa Biosystems) using QuantStudio5 (Thermo Fisher). PCR primer pairs (each forward and reverse) were as follows: GGAGTTGCGTCGCCTTAGT and TGGTCCTGAAGGAGGATAGG for ASAH1 mRNA, GTAAAGAAGCTGCCCAGCAC and GCCGGTGAGTGAGCTTGAG for ASAH1 pre-mRNA, CATGTACGTTGCTATCCAGGC and CTCCTTAATGTCACGCACGAT for ACTB mRNA, CTCTGCTCCTCCTGTTCGAC and ACGACCAAATCCGTTGACTC for GAPDH mRNA, CGAACGTCTGCCCTATCAACTT and ACCCGTGGTCACCATGGTA for 18S rRNA, and CCTGCCCAAGCTCTACCTT and AAGGCAGAAGATGTAGAGC for p21 mRNA, GAGCTGGGGAATGGGACT and TGATGGCATGGACTGTGG for GAPDH pre-mRNA, CAAAACTCCCGTGCTGATCA and GGCTGGAGTGCAGTGGCTAT for 7SL RNA, GCTGTGGAGTTCTTAAATATCAACC and TTCTCAATCCTGAAATCCCCTA for MALAT1 RNA, and TTCTCTCCGTCCTCGGATTCTCTG and TCTTCTTGTTCCTCCTCAGAGTCG for MYC mRNA.

Polyribosome fractionation

Cells were incubated with cycloheximide (Calbiochem; 100 μg/ml, 15 min) and the cytoplasmic lysates were prepared in polysome extraction buffer (PEB), fractionated by centrifugation through 10–50% linear sucrose gradients, and divided into 12 fractions for RT-qPCR analysis, as described [29].

Cell cycle analysis

Pre-senescent (PDL41) WI-38 cells were transfected with Control siRNA or ASAH1 siRNA as described previously. After 48 h, cells were harvested, washed with PBS and resuspended in 0.1% Triton-X in PBS for 5 min at 25° C; after brief centrifugation, the supernatant was aspirated. The pellet was incubated with RNase A (10 μg/ml in 20 mM Tris-HCl [pH 7.5]) at 37° C for 15 min; following a brief centrifugation, the supernatant was aspirated and 200 μl of 2 μg/ml propidium iodide was added. After incubation in the dark for 10 min, flow cytometry was performed on a BD FACS Canto for cell cycle analysis.

Caspase 3/7 activity

Apoptosis was monitored by measuring Caspase 3/7 activity using the Caspase-Glo 3/7 Assay System (Promega). Briefly, Caspase-Glo® 3/7 solution at 25° C was added directly to the cells, mixed vigorously for 30 sec, and incubated at 25° C in the dark for 30 to 180 min. Luminescence was then measured using a GloMax plate reader (Promega).

Mass spec and targeted mass spec (metabolomics)

Cells were harvested from 10-cm dishes individually from each of these groups: proliferating, replicative senescent, senescent after IR, Control siRNA, and ASAH1 siRNA. A minimum of 7 replicates were collected for each group with a minimum of 3×10⁶ cells collected per replicate. The cells were centrifuged and washed twice in cold PBS. The supernatant was discarded, and the cell pellet was frozen at -80° C. These frozen pellets were then used for targeted metabolomics using the BIOCRATES MxP Quant 500 Kit to assay approximately 40 metabolites within the following classes: (1) Ceramides, (2) Fatty acids, and (3) Sphingomyelins. Metabolites were extracted and concentrations were measured at Biocrates Life Sciences AG (Innsbruck, Austria) in pmol/106 cells. Targeted metabolomics was performed using tandem mass spectrometry with the Biocrates MxP® Quant 500 Kit (Biocrates, Innsbruck, Austria). Briefly, metabolites were measured using liquid chromatography tandem mass spectrometry (LC-MS/MS). Lipids specifically were measured by flow injection analysis-tandem mass spectrometry (FIA-MS/MS) using a 5500 QTRAP® instrument (AB Sciex, Darmstadt, Germany) with an electrospray ionization source, and small molecules were measured by LC-MS/MS using the same 5500 QTRAP® instrument. Measurement techniques are described at https://patents.google.com/patent/EP1897014B1 and https://patents.google.com/patent/EP1875401B1. A 96-well-based sample preparation device with internal standards was used to measure metabolite concentration in each sample (predefined sample amount). Next, a phenyl isothiocyanate solution was used to derivatize select classes of metabolites, and the target analytes were extracted with an organic solvent, followed by a dilution step. The extracts were then analyzed by FIA-MS/MS and LC-MS/MS methods using multiple reaction monitoring to detect the analytes. Data were quantified using mass spectrometry software (Sciex Analyst®) and imported into Biocrates MetIDQ™ software for further analysis. Metabolite values (pmol/10⁶ cells) below the limit of detection (LOD) or above the limit of quantification (ULOQ) were excluded from analyses. A proportional odds model [30], ordered logistic model and generalization of the non-parametric Wilcoxon and Kruskal-Wallis tests, was used to compare metabolite concentrations across three comparisons: Replicative senescence vs Proliferating cell control, IR-induced senescence vs Proliferating cell control, and ASAH1 siRNA vs Control siRNA. Estimates for comparisons with minimal sample size (i.e., only 1 or 2 metabolite measurements per group) could not be accurately estimated by the model and were excluded. Log odds ratios (OR; e.g., the increased or decreased odds of a higher metabolite concentration in one group compared to the other) and results significant at p < 0.05 were indicated. Complete analysis results are included in Supplementary Table 2.

ASAH1 levels increase in senescent cells

We used proliferating (‘P’, early population doubling level, PDL22) primary human WI-38 fibroblasts to obtain senescent cells either by replicative exhaustion (‘S’, PDL52) or by treatment with ionizing radiation (‘IR’, 10 Gy followed by culture for 10 additional days). Senescence was assessed by monitoring SA-βGal activity and the levels of senescence-associated markers p16 and p53 in both models (Figure 1A, 1B; Supplementary Figure 1A). To assess the changes in protein levels, whole-cell protein lysates prepared from proliferating and senescent cells were subjected to mass spectrometry (MS) proteomic analysis (Supplementary Table 1). We also extracted whole-cell RNA from proliferating and senescent cells and performed RNA-seq analysis using an Illumina HiSeq 4000 instrument at a depth of 75 million unidirectional reads for a total of 150 million paired-end reads per sample (GEO number GSE85771) [28]. In this analysis, we compared upregulated proteins and RNAs in replicative and IR-induced senescence (Figure 1C; Supplementary Figure 1B). We observed that both DPP4 mRNA and DPP4 protein were highly expressed, as we previously reported [31]. In addition, we observed that acid ceramidase (ASAH1) was upregulated in this comparison at both the protein and mRNA levels (Figure 1C). We also detected increased expression of ASAH1 upon doxorubicin-induced senescence in vascular smooth muscle cells (VSMCs) (Supplementary Figure 1C). The increase in ASAH1 mRNA levels in replicative (S) and IR-induced senescence was validated by RT-qPCR analysis. These senescent cells also displayed elevated levels of p21 mRNA (Figure 1D). Western blot analysis showed a robust increase in ASAH1 expression in both models of cell senescence (Figure 1E). In sum, ASAH1 protein levels increase strongly in senescent fibroblasts.

Altered ceramide metabolites in senescent cells

We then explored whether metabolites relevant to ASAH1 function – that is, ceramides, fatty acids and sphingomyelins – were altered in senescence (Figure 2A; results across all metabolites are included in Supplementary Table 2; ‘Replicatively Senescent vs Proliferating’ and ‘IR Senescent vs Proliferating’). When comparing replicatively senescent to proliferating control cells (Figure 2B) we found that one ceramide [Cer(d18:2/22:0)] and eight sphingomyelins [SM (OH) C16:1, SM C24:0, SM (OH) C24:1, SM C18:0, SM C24:1, SM (OH) C22:1, SM C26:1, SM (OH) C22:2] were significantly higher (log OR > 1; p < 0.05) in replicative senescence. When comparing IR-induced senescence to proliferating control cells (Figure 2C) we found that one ceramide [Cer(d18:1/24:1)] and two sphingomyelins [SM C16:0, SM C16:1] were significantly lower (log OR < 1; p < 0.05) in IR-induced senescence. Against our prediction, we did not observe an overlap or common changes in ASAH1-related metabolites between replicative and IR-induced senescence. The discrepancies could be dependent on the specific induction pathways, in this case replicative or IR-induced senescence, or heterogeneity of the senescent cells. Thus, these data suggest that while ASAH1-related metabolites may be altered in senescence, these two models of senescence do not share the metabolomic alterations.

The rise in ASAH1 levels occurs largely independently of increased ASAH1 mRNA stability or translation

To understand whether the twofold increase in ASAH1 mRNA levels in senescent cells is due to a rise in transcription, we measured the levels of ASAH1 pre-mRNA (a surrogate of ASAH1 gene transcription) by RT-qPCR analysis using primers that amplify exon-intron junctions. As shown in Figure 3A, there were no significant changes in ASAH1 pre-mRNA levels in senescent cells, suggesting that the increase in ASAH1 mRNA levels in senescent cells was not due to elevated transcription (Figure 3A). To study if ASAH1 mRNA was more stable in senescent cells upon replicative exhaustion, we treated fibroblasts with the RNA polymerase II inhibitor Actinomycin D in order to block transcription and measure by RT-qPCR analysis the rate of mRNA clearance (the half-life) in proliferating and senescent cells. As shown in Figure 3B, the ASAH1 mRNA was very stable in both senescent and proliferating cells; the half-life appeared to be longer than 8 hours, and thus could not be measured by treatment with Actinomycin D, as this drug is quite toxic to cells beyond 8 hours of continuous exposure. As controls, we measured the decay of the labile MYC mRNA [32], and the stable GAPDH mRNA (Figure 3B). The relative levels of nuclear and cytoplasmic ASAH1 mRNA were also unchanged with senescence (Supplementary Figure 2). We assayed the levels of noncoding RNAs 7SL and MALAT1 as cytoplasmic and nuclear controls, respectively (Supplementary Figure 2) [33, 34]. In sum, while ASAH1 mRNA may be slightly more stable in senescent cells to account for the twofold rise in levels in senescent cultures, we were not able to accurately calculate the half-lives.

Next, we investigated the marked rise in ASAH1 protein levels. First, we tested whether ASAH1 translation may be elevated in senescent cells by fractionating the cytoplasmic compartment of the cell on sucrose gradients using standard procedures, as previously described [35]. We collected fractions that were devoid of ribosome particles (unbound RNA, fractions 1 and 2), fractions containing ribosomal subunits and single ribosomes (fractions 3-5), low-molecular-weight polysomes (LMWP, fractions 6-8), and high-molecular-weight polysomes (HMWP, fractions 9-12). RNA was prepared from each fraction, and the levels of ASAH1 mRNA in each fraction were calculated as a percentage of total ASAH1 mRNA in the gradient. This analysis revealed only a slight shift in the distribution of ASAH1 mRNA across polysome fractions, suggesting that increased translation does not contribute strongly to the rise in ASAH1 protein levels in senescent cells (Figure 3C).

In the absence of clear evidence from polysome analysis that ASAH1 translation was elevated, we examined whether ASAH1 protein stability was altered between proliferating and senescent cells. To assess the changes in ASAH1 protein stability, we treated proliferating and senescent cells with cycloheximide (CHX), an inhibitor of de novo translation, and monitored the clearance of ASAH1 by Western blot analysis. Interestingly, ASAH1 was markedly stable in senescent cells, while it was undetected in proliferating cells; the control protein GAPDH was stable in both populations (Figure 4A). To look at ASAH1 protein stability in a different way, we treated cells with the proteasome inhibitor MG132. As shown in Figure 4B, ASAH1 protein was found to accumulate slightly in proliferating cells after four hours, but it was strongly induced after two hours of treatment with MG132 in senescent cells, as determined by Western blot analysis. We assessed the levels of the labile protein p21 to monitor the efficiency of the CHX and MG132 treatments (Figure 4A, 4B). Recently, it has been shown that ASAH1 is ubiquitinated and degraded by the proteasome through UBTD1 (ubiquitin domain-containing protein 1) [36]. These findings indicate that ASAH1 levels increase in senescent WI-38 fibroblasts largely due to enhanced protein stability.

Loss of ASAH1 suppresses cellular senescence and alters associated metabolites

To understand the role of ASAH1 in senescence, we transfected pre-senescent WI-38 cells with ASAH1-directed siRNA to reduce ASAH1 levels or a control siRNA. Silencing ASAH1 was confirmed by Western blot analysis, showing a robust reduction in cells transfected with ASAH1 siRNA compared to control cells (Figure 5A). ASAH1-depleted cells showed reduced levels of the senescence markers p16, p21, and p53 (Figure 5A). Furthermore, SA-βGal analysis revealed a decrease in positive (blue) senescent cells following ASAH1 depletion; this finding was confirmed by the SPiDER-βGal assay (Figure 5B, 5C), supporting the notion that lowering ASAH1 levels suppressed cell senescence. Cell cycle analysis indicated that depletion of ASAH1 in pre-senescent cells (PDL41) increased S-phase cells compared to control cells (Supplementary Figure 3) suggesting that ASAH1-depleted cells are more proliferative. When comparing ASAH1 siRNA to control siRNA populations (Figure 5D; results for all metabolites are in Supplementary Table 2 – ‘ASAH1 vs Control siRNA’) we found that sphingomyelins [SM C16:0, SM C16:1] were significantly decreased (log OR < 1; p < 0.05) in ASAH1-depleted cells. These changes may also be directly associated with ASAH1 silencing rather than senescence. Together, these data indicate that ASAH1 depletion reduces or delays cellular senescence and related metabolites, supporting a role for ASAH1 in promoting senescence.

ASAH1 inhibition induces senescent cell death while ASAH1 depletion sensitizes cells to senolysis

Senescence is detrimental for organ homeostasis with advancing age [37]. As senescent cells accumulate in aged tissue, they promote a range of age-related diseases and declines [38–40]; therefore, targeting and clearing senescent cells is clinically important. Senescent cell death has been achieved by a combination of Dasatinib plus Quercetin (D+Q), both approved by the US Food and Drug Administration (FDA) [41]. Thus, we examined whether ASAH1 might be implicated in senolysis and senescent-cell viability. To test these possibilities, we used two approaches. First, we treated proliferating or senescent cells with the ASAH1 inhibitor IV (ACi) [42] and monitored cell viability. As shown in Figure 6A, treatments with ACi alone did not have a significant impact on the viability of proliferating cells, but reduced by ~20% the viability of senescent cells. Importantly, we also assessed the effect of ACi in combination with D+Q, to evaluate whether ACi could enhance senolysis. Indeed, although the combination of ACi and D+Q was more toxic than D+Q alone for proliferating cells, it strikingly enhanced the effectiveness of D+Q in reducing senescent-cell viability (Figure 6A). To better characterize the increased death of senescent cells caused by ACi, we assessed apoptosis by monitoring caspase 3/7 activity. Inhibition of ASAH1 significantly enhanced caspase 3/7 activity in senescent cells treated with the senolytics D+Q (Figure 6B) in agreement with the reduction in viability observed in Figure 6A, while the different treatments did not influence caspase 3/7 activity in proliferating cells. To test if the reduction in viability caused by these drugs was due to apoptosis, we studied if the inhibitor of caspases ZVAD-FMK rescued cell death. As shown in Figure 6C, treatment with ZVAD-FMK significantly reduced senescent cell death by either by either D+Q alone or D+Q in combination with ACi (Figure 6C). The reduction of caspase activity in these conditions was confirmed by performing a caspase 3/7 activity assay (Figure 6D). These data suggest that pharmacological inhibition of ASAH1 potentiates cell death caused by senolytics at least in part by increasing caspase activity.

We then silenced ASAH1 in pre-senescent cells (PDL38-45) by transfection with ASAH1 siRNA and treated the cells with D+Q (see Materials and Methods). As shown, D+Q treatments decreased the number of viable proliferating and pre-senescent cells; however, pre-senescent cells were further sensitized to the toxicity of D+Q treatments if ASAH1 was inhibited, exhibiting more cell death than the untreated pre-senescent population (Figure 6E). Together, these data indicate that ASAH1 promotes senescent cell survival and resistance towards the senolytic combination D+Q (Figure 6F).