NF90 coordinately represses the senescence-associated secretory phenotype

Changes in levels of SASP mRNAs with replicative senescence

We investigated the post-transcriptional regulation of the expression of SASP factors in two human fibroblast models: WI-38 and IDH4. We compared ‘young’, proliferating (P), early-passage (population doubling, pdl ~20) WI-38 fibroblasts with late-passage, senescent (S, pdl ~54) WI-38 fibroblasts. We also compared IDH4 fibroblasts, which resembling ‘young’ cells and proliferate rapidly (P) when cultured with dexamethasone (Dex) for stable expression of SV40 large-T antigen, and become senescent (S) by ~9 days after Dex is removed from the culture media. As shown in Fig. 1A, a number of RBPs showed lower expression levels in senescent cells (S, -Dex) than in proliferating cells (P, +Dex). The diminished levels of HuR, AUF1, and nucleolin in senescent cells was previously reported [30-32]. However, the lower levels of NF90 in senescent fibroblasts was a novel finding. The senescence marker p21 was upregulated in senescent WI-38 and IDH4 populations (Fig. 1B).

In parallel, we surveyed a subset of mRNAs expressed in WI-38 and IDH4 cells (each proliferating and senescent, P and S). Reverse transcription (RT) followed by real-time-quantitative (q)PCR amplification was performed using a battery of primers that detected mRNAs encoding SASP products (as described by Coppé et al [4]); the levels of 18S rRNA were also measured and used for normalization. Quantification of the signals revealed that a handful of SASP mRNAs were lower in senescent cells of both cell types, as compared with proliferating cells, but generally SASP mRNAs were either unchanged or higher in senescent cells (Fig. 1C,D), supporting the view that besides higher secretion, SASP factor biosynthesis was also upregulated in senescent cells [4]. The relative fold changes were not identical in both sets of P and S cell types, but both cell groups shared the most highly upregulated transcripts, including IL8, CCL2, CXCL1, CXCL2, TNFRSF11B,IL6, and ICAM1 mRNAs, encoding IL-8, MCP-1, GROa, GRO(a,b,g), osteoprotegerin, IL-6, and ICAM1, respectively.

NF90 associates with numerous SASP mRNAs

As most SASP mRNAs (Fig. 1C,D) are known targets of various RBPs, we hypothesized that a given RBP might coordinately regulate the fate of several SASP mRNAs. Due to a number of technical and biological limitations we did not pursue analysis of HuR, AUF1, nucleolin or TTP (not shown); instead, we focused on NF90. By ribonucleoprotein (RNP) immunoprecipitation (RIP) analysis using an anti-NF90 antibody under conditions that preserved the integrity of RNPs, NF90 was effectively enriched after IP of both WI-38 and IDH4 lysates (Fig. 2A). The RNA present in the IP was isolated and subjected to RT-qPCR [reverse transcription (RT) followed by real-time, quantitative (q)PCR] analysis to monitor the levels of SASP mRNAs in NF90 IP relative to the control (IgG) IP. The same catalog of transcripts as in Fig. 1C was assayed, but only those mRNAs showing enrichments ≥twofold in the NF90 IP from WI-38 or IDH4 cells are shown (Fig. 2B,C). The most highly enriched NF90 target in both cell lines was IL8 mRNA, which showed a striking level of enrichment of >100-fold in WI-38 cells (pdl 22) and ~40-fold in IDH4 (+Dex) cells. Other transcripts such as CXCL1, IL6, FAS, and TNFRSF11B mRNAs were also enriched in NF90 IP materials obtained from WI-38 and IDH4 cells. These findings indicate that NF90 collectively associates with many mRNAs encoding SASP factors in human fibroblasts.

To investigate if NF90 was relevant to fibroblast senescence, we silenced NF90 by using small interfering (si)RNA. As shown in Fig. 3A, 48 h after transfection with NF90 siRNA, NF90 levels were markedly reduced, while the levels of senescence markers p21, p16, p53 were higher, suggesting that NF90 might contribute to repressing or delaying cellular senescence; nonetheless, by 48 h after siRNA transfection, all of the cells were negative for the activity of the senescence marker ß-galactosidase (not shown). Silencing NF90 was growth inhibitory for both WI-38 and IDH4 fibroblasts, as determined by measuring changes in cell number two weeks after siRNA transfection (Fig. 3B).

NF90 represses the translation of several SASP mRNAs

Silencing NF90 in WI-38 cells lowered the steady-state abundance of several SASP mRNAs. Reductions to one-half of their concentration in the NF90 siRNA group relative to the Ctrl siRNA group were seen for CSF2, LEP, and IL8 mRNAs in WI-38 cells (Fig. 4A), and for PLAU, CSF2, and TNFRSF1B mRNAs in IDH4 cells (Fig. 4B). The only transcript reduced in both groups was CSF2 mRNA (encoding GM-CSF), an mRNA that did not appear to interact with NF90 (Fig. 2B,C). Regarding transcripts upregulated after NF90 was silenced, only HGF mRNA (in WI-38 cells) and IGFBP4 mRNA (in IDH4 cells), two NF90 targets, showed twofold enrichment. These observations suggested that NF90 did not globally influence the levels of target SASP mRNAs.

To investigate if NF90 affected the translation of target SASP mRNAs, as was shown for a different subset of mRNAs expressed in cervical carcinoma cells [33], we prepared reporter constructs bearing the 3'UTR of target mRNAs, the mRNA region where NF90 preferentially interacts [33]. Constructs were prepared to test the full-length 3'UTRs of CCL2, CCL16, TNFRSF11B, CSF2,CXCL1, IL8, and IL6 mRNAs using the reporter plasmid psiCHECK2 as backbone (Fig. 5A). Confirmation that NF90 was capable of interacting with these 3'UTR segments was obtained by preparing biotinylated RNA spanning the 3'UTRs of these mRNAs (Methods). After incubation of the biotinylated transcripts with WI-38 (early-passage) cell lysates, the resulting RNP complexes were pulled down using streptavidin-coated beads, and the presence of NF90 in the pulldown RNP complexes was detected by Western blot analysis. As shown in Fig. 5B, NF90 associated with each of the full-length biotinylated 3'UTRs (Fig. 5A), but not with negative control RNAs spanning the nucleolin (NCL) mRNA 5'UTR and coding region. These results indicate that NF90 associates in vitro with the 3'UTRs of SASP mRNAs.

The influence of NF90 upon the 3'UTRs of SASP mRNAs was assessed using the luciferase reporter constructs (Fig. 5A). The ratio of renilla luciferase (RL, encoded by a reporter transcript bearing each 3'UTR of interest) to firefly luciferase (FL, encoded by an internal control reporter transcript) was set as 1 for the parent vector (psiCHECK2). RL/FL ratios in NF90 siRNA-transfected cells relative to Ctrl siRNA-transfected cells, shown in Fig. 5C (WI-38 cells, left; IDH4 cells right) indicated substantial increases in reporter expression (except in the psiCHECK-CCL2 and psiCHECK-TNFRSF11B groups), indicating that NF90 repressed the production of the reporter proteins and this repression was alleviated after silencing NF90.

The translation index for each reporter was calculated as the ratio of changes in luciferase activity (RL/FL) relative to the ratio of changes in luciferase mRNA levels (RL mRNA/FL mRNA). In WI-38 cells, high translation indexes were seen for reporters expressing the 3'UTRs of CXCL1, IL8,IL6, and CCL16 mRNAs, further suggesting that NF90 repressed their translation. The reporter bearing CSF2 3'UTR also showed an elevated translation index, despite not being a direct target of NF90 (Fig. 2C,D), suggesting that the effect of NF90 on the CSF2 3'UTR was indirect.

Production of SASP factors is repressed by NF90

Finally, we examined if the NF90-dependent differen-ces in post-transcriptional regulation of SASP mRNAs were reflected as changes in SASP factor abundance. By multiplex analysis, we first measured the levels of GM-CSF, MCP-1, GRO1, IL-6, and IL-8, encoded by NF90 target transcripts CSF2 mRNA, CCL2 mRNA, CXCL1 mRNA, IL6 mRNA and IL8 mRNA, respectively. The levels of these SASP factors were markedly elevated in senescent WI-38 and IDH4 cells relative to early-passage, proliferating cells (Fig. 6A). Importantly, by 1–3 weeks after silencing NF90 through sequential (weekly) siRNA transfections, MCP-1, GROa, IL-6, and IL-8 levels were markedly elevated in proliferating WI-38 and IDH4 populations (Fig. 6B), in line with the finding that NF90 silencing increased the translation of mRNAs bearing the CCL2,CXCL1, IL6 and IL8 3'UTRs (Fig. 5). By contrast, GM-CSF, encoded by a transcript that was not a target of NF90 (CSF2 mRNA) was actually lower in the NF90-silenced groups than in Ctrl siRNA cells (Fig. 6B). Taken together, our findings support the notion that NF90 represses the translation of SASP factors GROa, MCP-1, IL-6, and IL-8 through the 3'UTRs of CCL2, CXCL1, IL6 and IL8 mRNAs in proliferating cells, which express high levels of NF90. In senescent cells, NF90 levels decline, which in turn de-represses the biosynthesis of the major SASP factors MCP-1, GROa, IL-6, and IL-8.

Cell culture, treatments, transfection, and plasmids

Human WI-38 and IDH4 fibroblasts were cultured in Dulbecco's modified essential medium (DMEM, Invitrogen) supplemented with 10% fetal bovine serum and antibiotics. IDH4 cell culture medium was further supplemented with 1 mM dexamethasone (Dex) for constitutive expression of SV40 large T antigen to suppress senescence and induce proliferation [47]. To induce senescence of IDH4 cells, Dex was removed from the culture media and DMEM was supplemented with charcoal-stripped serum. A control small interfering RNA (Ctrl siRNA) was from Qiagen, and the siRNA directed to NF90 was from Santa Cruz Biotechnologies; the small RNAs were used at 50 nM for WI-38 cells and 10 nM for IDH4 cells. The reporter plasmids psiCHECK2-IL6, -IL8, -CXCL1, -CSF2, -TNFRSF11B, -CCL16, and -CCL2, which were constructed by cloning the corresponding PCR-amplified 3'UTR segments after the Renilla (luciferase) RL coding region. Small RNAs and plasmids were transfected with Lipofectamine RNA iMAX and Lipofectamine-2000 (Invitrogen) respectively. Luciferase activity assays were performed using the Dual-Luciferase® Reporter Assay System (Promega).

Western blot analysis

Whole-cell lysates were prepared using radioimmunoprecipitation assay (RIPA) buffer (10 mM Tris-HCl [pH 7.4], 150 mM NaCl, 1% NP-40, 1 mM EDTA, 0.1% SDS, 1 mM dithiothreitol), separated by electrophoresis in SDS-containing polyacrylamide gels, and transferred onto polyvinylidene difluoride (PVDF) membranes (Millipore). Incubations with primary antibodies recognizing NF90 (mouse monoclonal anti-DRBP76, BD Transduction Laboratories), p21 and AUF1 (Millipore), Nucleolin, HuR (Santa Cruz Biotech) or GAPDH (Abcam) were followed by incubations with the appropriate secondary antibodies conjugated with horseradish peroxidase (HRP) (GE Healthcare) and by detection using enhanced luminescence (GE Healthcare).

RNA analysis

Total RNA was prepared from whole cells or from RNP immunoprecipitation (IP) samples using TRIzol (Invitrogen). After reverse transcription (RT) using random hexamers and Maxima® Reverse Transcriptase (Thermo), the abundance of transcripts was assessed by real-time, quantitative PCR (qPCR) analysis using the SYBR green PCR master mix (Kapa Biosystems) and gene-specific primer sets (Table 1). RT-qPCR analysis was performed on Applied Biosystems model 7300 and 7900 instruments.

Immunoprecipitation of ribonucleoprotein (RNP) complexes

Endogenous mRNA-protein RNPs were precipitated from cytoplasmic lysates prepared from early-passage WI-38 or proliferating IDH4 cells. Lysates were prepared in a buffer containing 10 mM Hepes, 100 mM KCl, 5 mM MgCl₂, 25 mM EDTA, 0.5% IGEPAL, 2 mM DTT, 50 U/ml RNase out and protease inhibitors. Lysates were incubated (4 h, 4°C) with a suspension of protein-A Sepharose beads precoated with anti-NF90 or mouse IgG (BD Transduction Laboratories). The beads were washed with NT2 buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM MgCl₂?, 0.05% IGEPAL) and then incubated with NT2 buffer containing RNase-free DNase I (15 min, 30°C), washed with NT2 buffer and further incubated in NT2 buffer containing SDS and Proteinase K to digest proteins bound to the beads. RNA was extracted using phenol and chloroform, precipitated in the presence of glycoblue, and further analyzed by RT-qPCR using the primers listed in Table 1.

Biotin pulldown analysis

Primers used to prepare biotinylated transcripts spanning SASP mRNAs are listed below. After purification of the template PCR products, biotinylated transcripts were synthesized using MaxiScript T7 kit (Ambion) and whole-cell lysates (200 μg per sample) were incubated with 3 μg of purified biotinylated transcripts for 30 min at room temperature, and then complexes were isolated with Streptavidin-coupled Dynabeads (Invitrogen). Proteins present in the pulldown material were studied by Western blot analysis as described [33].

To synthesize biotinylated 3'UTR RNAs corresponding to the SASP mRNAs, PCR fragments were prepared using forward primers as in that contained the T7 RNA polymerase promoter sequence [(T7), CCAAGCTTCTAATACGACTCACTATAGGGAGA] and the primer sequences in Table 2. The sizes of the resulting RNAs were 427 b for IL6, 1250 b for IL8, 700 b for CXCL1, 314 b for CSF2, 816 b for TNFRSF11B, 1067 b for CCL16, and 372 b for CCL2 3'UTR. Primer pairs to amplify PCR templates for the synthesis of negative RNA controls in biotin pulldown experiments (NCL 5'UTR and two CR segments) are indicated in Table 2 and reported previously [48]; the sizes of the resulting RNAs were 142 b for NCL(5'), 550 b for NCL(CR1), and 550 b for NCL(CR2).

Measurement of SASP factors

The concentration of SASP factors (GM-CSF, MCP-1, GROa, IL-6, and IL-8) in the culture supernatant was analyzed by using multiplex assay, following the manufacturer's instructions (Biorad Laboratories, Hercules, CA). Data were normalized and expressed as pg per ml per cell per 24 h.