HDAC2 and 7 down-regulation induces senescence in dermal fibroblasts

HDACs expression is decreased in dermal HDFs during replicative senescence

Except for a decreased expression of HDAC1 [13] and HDAC4 [17] during replicative senescence of fibroblasts, changes in the expression of other canonical HDACs during senescence have not been reported. To investigate whether the expression of other HDACs is also modified during replicative senescence of human dermal fibroblasts (HDFs), we monitored the protein levels of HDACs 1-7 in three HDFs strains: AG04431 human dermal fibroblasts and normal HDFs isolated from 57-yrs-old (NHDF1) and 3-yrs-old (NHDF2) donors. Dermal fibroblasts were passaged in cell culture until replicative senescence (RS), which occured namely when cells reached stable growth-arrest for at least two weeks and when more than 70% of the cell population was SA-βgal positive. HDACs 1-7 proteins levels displayed a general decrease in RS as compared to fibroblasts at early passage (young) (Figure 1A, 1B). This decrease in protein abundance could vary between 29% (HDAC5) and 98% (HDAC2) (Figure 1C, 1D), and seemed progressive as shown in an intermediate passage for NHDFs, corresponding to the middle of their proliferative lifespan (Figure 1B, 1D and Supplementary Figure 1). As a control, we confirmed an increase in the protein abundance of p16^INK4A in RS conditions. The overall decrease in protein abundance of HDACs is corroborated with a significant decrease in the levels of transcripts encoding HDAC1-3, 6, and 7 in RS, compared to early passage (young) AG04431 (Supplementary Figure 2). In order to study whether this difference in protein expression was associated with a difference in enzymatic activity, we carried out a class IIa HDACs activity assay. We observed a decrease in class IIa HDACs activity between young and RS HDFs (Figure 1E, 1F). Together, these results suggest a global decrease in HDACs expression and activity during replicative senescence of dermal fibroblasts.

SAHA treatment in young HDFs promotes the appearance of senescence-associated biomarkers

In order to mimic the impact of this overall decrease in HDACs expression during RS, the inhibition of HDACs was studied using SAHA, a pan-HDACs inhibitor. In cancer cell lines and in normal fibroblasts, SAHA was shown to induce several senescence biomarkers such as growth arrest and an increased proportion of SA-βgal positive cells [18, 19]. Dermal fibroblasts were incubated with SAHA during 24 hr, 48 hr and 72 hr and several biomarkers of senescence were assessed, including the expression of several SASP factors. Subcytotoxic doses of 5 and 10 μM of SAHA were used (Supplementary Figure 3A, 3B). A dose-dependent increase in the acetylation of Histone H3 was observed following SAHA treatment (Supplementary Figure 4A, 4B), as well as a decreased activity of class IIa HDACs (Supplementary Figure 4C). The impact of SAHA treatment on the senescent phenotype of fibroblasts was then analysed. The effect of SAHA on proliferation was determined by the proportion of Ki-67 positive cells [20]. A strong decrease in the proportion of Ki-67 positive cells is observed after SAHA treatment regardless of the dose and duration of the treatment (Figure 2A, 2B). The protein level of p16^INK4-A and p21^WAF-1, two CDKIs that are associated with growth-arrest in senescent cells, were then assessed. The expression of p21^WAF1 protein was increased after 24 hours of treatment with both doses (Figure 2C, 2D), while the protein level of p16^INK4-A remained stable (Figure 2C, 2E). This result is consistent with previous studies that have described increased expression of p21^WAF1 after SAHA treatment [21]. In addition, the proportion of SA-βgal positive cells significantly increased in response to SAHA treatment after 72 hours (Figure 2F). The effect of SAHA on Lamin B1 (LMNB1) mRNA levels was then analysed and the results demonstrated a strong decrease of LMNB1 transcript levels following treatment (Figure 2G). Lamin B1 is a protein of the nuclear lamina, lining the inner surface of the nuclear envelope, and is now considered as a robust biomarker of senescence as its expression is decreased in several models of senescence [22]. Regarding the effect of SAHA treatment on SASP factors expression, SAHA treatment induced an increase in the expression of IL-6, IL-8 (CXCL8), MMP-1, MMP-3, IL-1β and Gro-α (CXCL1) at the mRNA level, whatever the dose and the duration of treatment (Figure 2H). Moreover, the secretion of IL-6 and IL-8 significantly increased from 24 hours for IL-6, and from 48 hours for IL-8 (Figure 2I), confirming a rapid effect of SAHA on SASP induction, as recently reported following the treatment of HDFs with sodium butyrate (NaB) and trichostatin A (TSA), two other HDACi [15].

Altogether, these results demonstrate that the treatment of young dermal HDFs with SAHA induces the premature onset of senescence in dermal fibroblasts, including the expression of some SASP factors.

siRNA targeted inhibition of HDAC2 and HDAC7 induces senescent-like phenotype in young HDFs

As SAHA is a pan HDACi, it is not possible to know if its effects are associated with the inhibition of one or more HDACs. In order to decipher whether the premature senescence phenotype associated to HDACs inhibition is related to a specific HDAC, we performed a targeted invalidation by siRNA of two HDACs of different classes whose expression levels were reported to be highly diminished in senescent fibroblasts, i.e. HDAC2 and HDAC7. The invalidation of HDAC2 and HDAC7 has been reported to be associated with proliferation arrest respectively in MCF-7 breast cancer cell line [23] and in mucoepidermoid carcinoma cells [24], but not yet investigated on the potential induction of the senescent phenotype in normal cells. Dermal HDFs were transfected with siRNA targeting either HDAC2 or HDAC7 during 24 hr and maintained in culture for seven days. HDAC2 and HDAC7 transcript levels were decreased from 24 hr post-transfection to at least 168 hr post-transfection (Supplementary Figure 5). Regarding the protein level, HDAC2 and HDAC7 were decreased from 48 hr after siRNA transfection and remained low until 168 hr (Figure 3A). No impact of the invalidation of HDAC2 on the protein abundance of HDAC7 was observed, and vice versa (Figure 3A). This observation confirmed the proper inhibition of HDAC2 and HDAC7 expression by using specific siRNA.

The impact of HDAC2 or HDAC7 invalidation on proliferation was then tested. Results demonstrated that HDAC2 and HDAC7 invalidation strongly reduced the proportion of Ki-67 positive cells, starting 24 hr after transfection (Figure 3B, 3C). This decrease in proliferation was correlated to an increase in the mRNA level of p16^INK4-A and p21^WAF-1 from 48 hr to 168 hr post-transfection (Figure 3D). At the protein level, results displayed a significant increase in the protein abundance of p16^INK4-A at 120 hr post-transfection for siHDAC7, as well as a tendency for an increase in protein expression of p21^WAF-1, for both HDACs siRNA, at 144 hr (Supplementary Figure 6). The proportion of SA-βgal positive cells following HDAC2 and HDAC7 invalidation showed a significant increase in both siHDAC2 and siHDAC7 transfected cells, compared to cells transfected with the non-targeting siRNA (siNT) at 120 hr and 168 hr post transfection (Figure 3E). The transcript levels of LMNB1 was then studied and results demonstrated a decrease in LMNB1 transcript levels at 48 hr and 72 hr in HDFs transfected with siHDAC2 and at 96 hr in HDFs transfected with siHDAC7 (Figure 3F). The effect of siHDAC2 or siHDAC7 transfection on the expression of several SASP factors displayed different results between the two siRNAs. HDFs transfected with siHDAC7 overexpressed significantly IL-6, IL-8, MMP-1 and MMP-3 at some time points (Figure 3G). Conversely, transfection of HDFs with siHDAC2 showed minor effects on the expression of these SASP factors, as only a significant increase in the expression of MMP-3 was detected at 96 hr post-transfection (Figure 3G).

These results demonstrate that the transfection of young HDFs with siHDAC2 or siHDAC7 induces a senescent phenotype highlighted by a growth-arrest, an increase in the proportion of SA-βgal positive cells and a decreased expression of LMNB1. However, only siHDAC7 has an impact on the expression levels of several SASP factors, and is subsequently able to reproduce a senescent phenotype similar to the one obtained upon SAHA treatment.

Re-expression of HDAC7 delays the occurrence of the cell cycle arrest in pre-senescent dermal fibroblasts

Previous results have shown that the inhibition of HDACs with the use of SAHA or siRNA induced the premature appearance of senescence in dermal fibroblasts. The analysis of the re-expression of HDAC2 or HDAC7 expression in pre-senescent cells, namely cells at a few passages to enter in replicative senescence, but still able to proliferate was then addressed. For this purpose, we infected AG04431 HDFs and dermal NHDFs with lentivirus expressing HDAC2 (pLV HDAC2), HDAC7 (pLV HDAC7), or enhanced green fluorescent protein (pLV EGFP) as a control. The increase in the protein levels of HDAC2 or HDAC7 in transduced pre-senescent cells was monitored by Western blot (Figure 4A, 4B and Supplementary Figure 7A, 7B). After transduction, cells were passaged until they reached replicative senescence. Analyses were carried out at passage 4 post transduction for all replicates and for both AG04431 HDFs and NHDFs (corresponding to day 28 for pLV HDAC7 and to day 36 for CTL, pLV EGFP and pLV HDAC2). We first analysed the effect of transduction on cell proliferation. Both HDFs and NHDFs cells transduced with pLV HDAC7 exhibited a resumption of their proliferation as well as an extension of their proliferative lifespan, as shown by an increased cell number (Figure 4C and Supplementary Figure 7C). Moreover, cells transduced with pLV HDAC7displayed an increased proportion of Ki-67 positive signals (Figure 4D, 4E). Re-expression of HDAC7 also results in a significant decrease in the proportion of SA-βgal positive cells (Figure 4F and Supplementary Figure 7D). Moreover, representative cell images demonstrated that HDFs transduced with pLV HDAC7 displayed a more spindle-shaped morphology, resembling young proliferative cells (Supplementary Figure 8). The effect of transduction on the expression of SASP factors was then assessed. Surprisingly, an increased expression of IL-6, IL-8, MMP-1 and MMP-3 was observed in the HDFs transduced with pLV HDAC7 compared with other conditions (Figure 4G). HDAC7 re-expression leads to the extension of the proliferative capacities of HDFs as well as to a decrease in the proportion of SA-βgal positive cells. However, it is not associated with a reduced expression of SASP factors.

NF-ĸB activation is partly involved in IL-6 and IL-8 expression after HDACs inhibition in young HDFs

Our results pointed out that HDACs inhibition in AG04431 dermal fibroblasts, either with SAHA or with siRNA targeting HDAC2 or HDAC7 induced a strong overexpression of some SASP factors, such as IL-6, IL-8, MMP-1 and MMP-3. SASP regulation and particularly IL-6 and IL-8 expression, is largely dependent on NF-κB (nuclear factor-kappa B) activation [11].

NF-κB is a transcription factor composed of several subunits that is sequestered in the cytoplasm by IκBα (NF-κB inhibitor alpha), keeping it in an inactive state and preventing its nuclear translocation. In the presence of an activating signal, the activation of IκBs kinases (IKKs) leads to the phosphorylation and the proteasomal degradation of IκBα, allowing the nuclear translocation of NF-κB and the activation of target genes. In addition to NF-κB, the expression of IL-6 and IL-8 can be dependent on other transcription factors such as C/EBPβ (CCAAT/enhancer binding protein beta) [25]. The protein level of IκBα after SAHA treatment was first assessed. SAHA treatment in AG04331 HDFs induced the degradation of IκBα after 72 hr (Figure 5A, 5B). We confirmed the activation of NF-κB as shown by an increase of p65 nuclear translocation upon SAHA treatment (Figure 5C, 5D). In order to decipher whether the levels of expression of SASP factors after SAHA treatment were dependent on NF-κB activation, we performed NF-κB inhibition using JSH-23, an inhibitor of p65 nuclear translocation. The data highlighted that in cells treated with JSH-23 concomitantly with SAHA treatment, the nuclear signal of p65 is strongly reduced, consolidating the nuclear translocation of NF-κB following SAHA treatment (Figure 5C, 5D). The combined treatment of dermal HDFs with JSH-23 and SAHA significantly reduced the expression of IL-6 and IL-8 (Figure 5E). These results reinforce the link between NF-κB activation and the expression of SASP factors following SAHA treatment.

The analysis of IκBα protein abundance after HDAC2 or HDAC7 siRNA transfection displayed a slight decrease at 96 hr following transfection with HDAC7 siRNA, suggesting a possible NF-κB activation in this condition (Supplementary Figure 9). These results confirm the possible involvement of NF-κB activation in the regulation of the expression of SASP factors upon HDACs inhibition either through treatment using SAHA or through targeted siRNA invalidation.

Cell culture and treatments

AG04431 HDFs (Coriell Institute) were cultured in BME, supplemented with 10% FBS and 2 mM L-glutamine (Life Technologies) until replicative senescence.

Primary normal human dermal fibroblasts (NHDFs) were isolated from adult abdominal (NHDF1) and young foreskin (NHDF2) samples as previously described [47]. Samples were obtained after plastic surgery (Dr. B. Bienfait, Clinique St-Luc, Bouge Belgium) or circumcision (Dr L. de Visscher, Clinique St-Luc, Bouge, Belgium) following approval by the ethic committee of the Clinique St-Luc (Bouge, Belgium). NHDFs were cultured in BME, supplemented with 10% FBS and 2 mM L-glutamine (Life Technologies) until intermediate replicative lifespan (corresponding to 50% of their proliferative lifespan) and until replicative senescence. All cells were cultured at 37° C in a humidified atmosphere containing 5% CO₂.

HDFs were studied at early passage (young), i.e. at passages prior the middle of their proliferative lifespan, or when pre-senescent, i.e. at a few passages to entering in replicative senescence, but still proliferating.

Since NHDFs have a longer proliferative lifespan than AG04431, we used an additional passage, called “intermediate”, for these cells corresponding approximately to the middle of their proliferative lifespan.

For SAHA treatment, AG04431 HDFs were treated with 0, 5 or 10 μM SAHA (Selleckchem) during 24, 48 or 72 hr. Fresh medium containing SAHA was renewed every day.

For NF-κB pathway study, AG04431 HDFS were treated with 0 or 10 μM of SAHA during 24, 48 or 72 hr and with 0 or 100 μM of JSH-23 (Sigma-Aldrich) concomitantly during the first 24 hours. Fresh medium containing SAHA was renewed every day. Positive control for NF-ĸB activation was generated by treating HDFs with 20 ng/ml of TNF-α (R&D Systems) during 20 minutes.

Cell transfection and RNA interference

AG04431 HDFs transfections were performed using siGENOME SMART pool human HDAC2 or HDAC7 (Dharmacon). Non-target siRNA was used as a control for non-specific effect (Dharmacon). Cells were transfected 24 hr under standard culture conditions with 25 nM siRNA using Dharmafect (Dharmacon) transfection reagent according to the manufacturer’s instructions. Cells were cultured under standard culture conditions during 7 days until analysis.

Lentiviral transduction

Mammalian gene expression lentiviral vectors pLV EF1A EGFP, pLV EF1A hHDAC2, pLV EF1A hHDAC7 were purchased from Vector Builder and were transformed, amplified and purified by GIGA-Viral vectors platform (ULiège). AG04431 HDFs and NHDFs were transduced with a multiplicity of infection (MOI) of 50 in fresh medium containing protamine sulfate (8 μg/ml, MP biochemicals). At 24 hr after the transduction, the medium was replaced with fresh medium. At confluency, cells were subcultured until they reached replicative senescence.

Senescence-associated beta-galactosidase (SA-βgal) activity

Cells were seeded in 6-well plates (Corning) and the next day, SA-βgal staining was performed as described in [48, 49] during 16 hr. The proportion of SA-βgal positive cells was quantified by counting at least 300 cells per well.

RNA isolation and real time PCR

Total RNA was isolated using the ReliaPrep cell and tissue miniprep kit (Promega) and reverse-transcribed using GoScript Reverse Transcriptase Kit (Promega) following manufacturer’s instructions. Real-time polymerase chain reaction (PCR) was performed using GoTaq qPCR Master Mix (Promega), primers and the Viia7 Real-Time PCR system (Applied Biosystems). Primers used are referenced in Supplementary Table 1. Relative abundance was determined with the ΔΔCq method [50] normalized to the mRNA abundance of GAPDH and expressed relative to the stated control.

Western blot analysis

Total cell protein lysates were obtained using a buffer containing 40 mM TrisHydroxyMethyl (Tris, Merck Millipore), 150 mM KCl (Merck Millipore), 1 mM EDTA (Merck Millipore), pH 7.5, 1% Triton X-100 (Sigma-Aldrich), supplemented with Complete Protease Inhibitor Cocktail (Roche) and 4% phosphatase inhibitor buffer (25 mM Na₃VO₄ (Sigma-Aldrich), 250 mM 4-nitrophenyl phosphate (Sigma-Aldrich), 250 mM β-glycerophosphate (VWR), 125 mM NaF (Merck Millipore)). Protein concentration was determined by Pierce 660 nm Protein Assay (Thermo Scientific). 5 to 10 μg of proteins were separated by SDS-PAGE on 10% SDS-PAGE (NuPAGE, Thermo Fisher) gels and electrotransferred on a polyvinylidene fluoride membrane (Merck Millipore) or on a nitrocellulose membrane (Merck Millipore). Molecular weights were determined with colour pre-stained protein standard (New England Biolabs). Membranes were blocked 1 hr at room temperature in Odyssey blocking buffer (LI-COR) and then overnight at 4° C in Odyssey blocking buffer with 0.1% Tween 20 (Bio-Rad) containing a primary antibody. Membranes were incubated with the secondary antibody for 1 hr in Odyssey blocking buffer with 0.1% Tween 20. Then the membranes were dried 1 hr at 37° C and scanned using Image Studio Lite software V3.1.4 (LI-COR) for quantification. Antibodies used in this study are referenced in Supplementary Table 2.

Immunofluorescence

Cells were fixed with paraformaldehyde 4% (PFA, Merck Millipore) during 10 min. Cells were then permeabilized with PBS + 1% Triton-X100 (Sigma-Aldrich) during 5 min. Cells were washed in PBS containing 2% Bovine Serum Albumin (BSA, Santa Cruz Biotechnology) 3 times during 10 minutes. Glass coverslips were then incubated in PBS + BSA 2% containing a primary antibody overnight at 4° C. Cells were incubated with PBS + BSA 2% containing a secondary antibody at room temperature in a wet room. Cells were then stained with DAPI (1 ng/ml in PBS, Sigma-Aldrich) for 5 minutes in room temperature. Coverslips were then mounted on superfrost microscope slides using Mowiol (Sigma-Aldrich) and analysed by confocal microscopy (Leica). Antibodies used are referenced in Supplementary Table 2.

Enzyme-linked immunosorbent analysis (ELISA)

Supernatants were collected after SAHA treatment and kept in -80° C until the test. IL-6 and IL-8 secreted levels were analysed using Human IL-6 Quantikine® Elisa (R&D Systems) and Human IL-8 Quantikine® Elisa (R&D Systems) following manufacturer’s instructions. IL-6 and IL-8 secreted levels were quantified relative to total protein content.

Statistical analysis

Experiments were performed at least 3 times independently except when noted otherwise. Results are expressed as mean +/- SD (standard deviation). Statistical analyses were performed by using t-test, ANOVA I or ANOVA II (GraphPad Prism® Software) (NS: non significant; *: p<0.05; **: p<0.01; ***: p<0.001).