HDQ-P1 human breast cancer cells carry a homozygous nonsense mutation at codon 213 (CGA to TGA; R213X) in the TP53 gene (32). HDQ-P1 cells (DSMZ, Braunschweig, Germany) were cultured in DMEM low-glucose medium (Hyclone, Logan, UT, USA) supplemented with 10% fetal bovine serum (Hyclone, Logan, UT, USA) and 2.5 µg/ml plasmocin (Invivogen, Toulouse, France). H1299 lung adenocarcinoma cells (ATCC, Wesel, Germany) are p53 null (33). They were grown in RPMI-1640 medium (Hyclone, Logan, UT, USA) supplemented with 10% fetal bovine serum and 2.5 µg/ml plasmocin. For stably transfected H1299 cells, 1 µg/ml puromycin was added in the culture medium. HCT116 wtp53^+/+ colon cancer cells (34) were cultured in McCoy medium (Hyclone, Logan, UT, USA) supplemented with 10% fetal bovine serum and 2.55 µg/ml plasmocin.

The plasmid pCMV-puro-bam with the CMV promoter and the puromycin resistance gene were used as backbone and empty vector. The plasmid pCMV-puro-bam p53 R213X was created by inserting the p53 coding sequence with a nonsense mutation at codon 213 downstream of the CMV promoter in pCMV-puro-bam. The plasmid pCMV-puro-bam R213XΔC-EGFP was generated by inserting the first 213 codons of R213X p53 fused in frame with the EGFP coding sequence downstream of the CMV promoter. All plasmid constructs were made by GenScript (Genscript, Piscataway, NJ, USA).

H1299 cells were plated in six-well plates at a density of 200,000 cells per well. Next day cells were transfected with 1 µg DNA constructs per well by using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol. For selection of stably transfected cells, cells were transfected as indicated above and cultured in the presence of 1 µg/ml puromycin.

Cells were plated in six-well plates at a density of 300,000 cells per well (HDQ-P1) or 150,000 cells per well (H1299 and HCT116 wtp53^+/+). Cells were allowed to adhere overnight and were then treated with 100, 200, and 400 µM of G418 (HyClone Laboratories, Logan, UT, USA) or with 2 mM of gentamicin (Gibco/Life Technologies, Stockholm, Sweden) for 72 h. For combination treatment with either G418 or gentamicin, different concentrations of bortezomib, nutlin-3a, or MI-773 were added for 48 h (bortezomib) or 72 h (nutlin-3a and MI-773). Bortezomib, nutlin-3a, and MI-773 were from Sigma-Aldrich (Stockholm, Sweden). Cells were harvested and lysed with RIPA buffer, 20–50 µg of protein were loaded on 10% SDS/PAGE gels (Life Technologies, Stockholm, Sweden) and run in 1× MOPS (Life Technologies, Stockholm, Sweden). Gels were blotted on nitrocellulose membranes using iBlot (Life Technologies, Stockholm, Sweden). Membranes were blocked in PBS-T with 5% milk and then incubated with antibodies against p53 (DO-1, Santa Cruz, CA, USA; PAb421, Calbiochem Merk, Darmstadt, Germany; HR231, a kind gift from Dr. Thierry Soussi, Department of Oncology-Pathology, Karolinska Institutet, Stockholm, Sweden), p21 (p21 F-5, Santa Cruz, CA, USA; p21 Waf1/Cip1, Cell Signaling Technology, Leiden, Netherlands), or Wig-1 (FJ1, raised against a C-terminal Wig-1 peptide, F. Jerhammar, unpublished). GAPDH (Santa Cruz, CA, USA) and β-actin (Sigma-Aldrich, Stockholm, Sweden) were used as loading controls. Proteins were visualized using HRP-conjugated secondary antibodies and SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Scientific, Stockholm, Sweden) with a CCD camera (Fujifilm, Stockholm, Sweden), analyzed by a Luminescent Image Analyzer LAS-1000 plus, and quantified with Image Quant software (Fujifilm, Stockholm, Sweden).

RNA was extracted with RNeasy mini kit (Qiagen GmbH, Hilden, Germany) according to the manufacturer’s protocol and quantified by a nanodrop-1000 apparatus (Thermo Scientific, Stockholm, Sweden). cDNA was synthesized using the SuperScript II Reverse Transcriptase (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol. Real-time PCR (RT-PCR) was carried out in an Applied Biosystems 7500 RT-PCR System sequence detection system version 1.2 (Applied Biosystems, Foster City, CA, USA) according to the manufacturer’s recommendations. TaqMan Gene Expression Assays (TP53: Hs00153340_m1, MDM2: Hs01066930_m1, BAX: Hs00414514_m1, CDKN1A: Hs00355782_m1, BBC3: Hs00248075_m1, c1/c2: Hs01028907_m1, ACTB: Hs99999903_m1, ZMAT3: Hs00536976_m1, PMAIP1: Hs00560402_m1, FAS: Hs00531110_m1, and GAPDH: Hs99999905_m1) (Applied Biosystems, Foster City, CA, USA). Relative gene expression was calculated by the 2 − ΔΔCt method after normalization to GAPDH levels.

HDQ-P1, H1299-R213X, and empty vector-transfected H1299 cells were seeded in six-well plates with glass slides, at a density of 300,000 cells per well for HDQ-P1 cells and 200,000 cells per well for H1299 cells. The next day, cells were treated with 100 µM G418. After 72 h treatment, cells were washed in PBS, fixed in 4% paraformaldehyde, permeabilized with 0.2% Triton, and washed in PBS. Anti-p53 antibody PAb421 (Merck, Darmstadt, Germany) was diluted 100 times in 2% BSA and anti-p53 antibody FL393 (Santa Cruz, CA, USA) was diluted 400 times in 2% BSA. Both antibodies were mixed and incubated with the cells for 2 h at room temperature. After washing with PBS, cells were incubated with secondary antibodies conjugated with Alexa Fluor dye (Invitrogen/Thermo Fisher Scientific, Sweden), for 1 h at room temperature, washed with PBS and mounted with HardSet mounting medium with DAPI (Vector Laboratories, CA, USA). Images were obtained with a Zeiss Axioplan 2 microscope with an AxioCam HRm Camera.

Stably transfected H1299 cells were plated in 96-well plates at a density of 1,500 cells per well. The next day cells were treated with G418 and nutlin-3a at different concentrations. After 72 h, cells were incubated with cell proliferation reagent WST-1 (Roche Diagnostics GmbH, Mannheim, Germany) for 30 min. Absorbance at 450 nm was then determined in a Tecan reader Infinite M1000 (Tecan Austria, Grödig, Austria).

Nonsense mutant TP53 R213X or empty vector-transfected H1299 cells were seeded in six-well plates at a density of 100,000 cells per well in 2 ml medium. The next day cells were treated with G418 at 200 µM and/or nutlin-3a at 10 µM and incubated for 72 h (all treatment conditions were in duplicates). All cells were collected, washed with PBS, and fixed with ice-cold 100% ethanol overnight at 4°C. The cells were pelleted and the supernatant was discarded. The samples were incubated at 37°C for 30 min in darkness with propidium iodide (0.025 mg/ml) (Sigma-Aldrich, Stockholm, Sweden) and RNase A (0.125 mg/ml) (Sigma-Aldrich, Stockholm, Sweden). 50 µl PBS was added to each sample and the samples were then analyzed with NovoCyte^® flow cytometer (ACEA Biosciences, San Diego, CA, USA). Cells were gated by combining forward scatter-height/side scatter-height and forward scatter-height/forward scatter-area gates.

For comparison of dependent variables, normal distribution of the differences between the compared variables was first assessed. If the distribution was normal, ANOVA model for repeated measurements was employed and for multiple comparisons Tukey or Unequal N HSD post hoc tests were applied. If the distribution was not normal, the non-parametric Wilcoxon test was applied. For comparison of independent variables, the Kruskal–Wallis test was applied since the data did not comply with the equal variance requirement for ANOVA model. Statistica 13.0 (Dell Inc., Tulsa, OK, USA) was used for the statistical analysis.

We treated human HDQ-P1 breast carcinoma cells carrying endogenous R213X nonsense mutant TP53 with G418 and examined expression of full-length p53 protein by immunofluorescence staining with FL393 polyclonal p53 antibody and the monoclonal antibody PAb421 that recognizes an epitope in the p53 C-terminus. Treatment with 400 µM G418 strongly induced p53 according to immunostaining with both the FL393 and PAb421 antibodies (Figure 1A), indicating induction of full-length p53. Western blotting confirmed induction of full-length p53 by G418 in a dose-dependent manner (Figure 1B). Gentamicin also induced full-length p53, although much less efficiently (Figure 1B). These data confirm a previous study showing that aminoglycosides G418 and gentamicin induce translational readthrough of premature stop codon R213X and induce production of full-length p53 protein expression in HDQ-P1 cells (29). Induction of full-length p53 by G418 was accompanied by upregulation of p53 targets p21 (encoded by the CDKN1A gene) and Wig-1 (ZMAT3) in a concentration-dependent manner, indicating that the recovered full-length p53 protein is at least partially functional as transcription factor (Figure 1B). Gentamicin did not induce significant levels of p21 and Wig-1.

Next, we examined whether G418 and gentamicin affect p53 mRNA levels. Premature stop codons trigger NMD via remaining exon-junction complexes (11). Our RT-PCR analysis showed that both G418 and gentamicin induced p53 mRNA levels in a time-dependent manner. G418 treatment resulted in an 18-fold induction of p53 mRNA levels while gentamicin caused a more modest 4-fold induction after 3 days treatment. The differences between all G418 and gentamicin-treated samples and the control samples were statistically significant (p < 0.05). For 48 and 72 h treatments, the post hoc Tukey test in the ANOVA model was used for statistical analysis. The data for 24 h treatment did not fit normal distribution, and they were therefore analyzed with the Wilcoxon matched pairs test (Figure 1C). The observed increase in p53 mRNA levels is expected since translational readthrough would remove exon-junction complexes and thus inhibit NMD (35). However, we do not exclude the possibility that p53 mRNA level can be enhanced due to increased transcription.

To test this effect in other cells, we transiently transfected human H1299 lung adenocarcinoma cells with either a p53 R213X construct, or a plasmid containing the first 213 codons of R213X p53 fused in frame with EGFP, or empty vector. Western blot analysis with the DO-1 antibody showed expression of full-length p53 in R213X-transfected cells upon G418 treatment (Figure 1D, upper panel). Similarly, we detected G418-mediated induction of the p53 R213XΔC-EGFP fusion protein in H1299 cells transfected with the p53 R213XΔC-EGFP construct (Figure 1D, upper panel). The apparent molecular weight of the p53 R213XΔC-EGFP fusion protein is larger than that of full-length p53 due to the addition of EGFP. The expression of full-length p53 in this setting was confirmed by the p53 C-terminal antibody HR231 (Figure 1D, lower panel). We also generated stable lines of H1299 cells carrying the R213X construct. G418 treatment induced full-length p53 in the H1299-R213X cells as assessed by immunostaining with the C-terminal antibody PAb421 (Figure 1E).

In order to further study if the full-length p53 protein induced by aminoglycosides in HDQ-P1 cells is transcriptionally active, we examined expression of p53 target genes by both RT-PCR and Western blotting. G418 induced p21, Bax, Puma, Fas, Mdm2, Noxa, and Wig-1 at the mRNA level, while gentamicin was less efficient. The differences between G418-treated and control samples were statistically significant for Wig-1, p21, Fas, Mdm2, and Noxa, (p < 0.05) but not for Bax and Puma. For gentamicin-treated samples, the differences were statistically significant (p < 0.05) for all tested targets except Puma. The Wilcoxon matched pairs test was applied here because the distributions in several groups differed significantly from the normal distribution (Figure 2A). Interestingly, mRNA induction for the studied genes correlated poorly between G418 and gentamicin (r = −0.32), suggesting not only differences in the potency between these two antibiotics but possibly also in the activity of the induced full-length p53 protein (see Discussion). We observed the most prominent upregulation of p21 and Wig-1 mRNA by G418, whereas Puma mRNA showed the strongest response after treatment with gentamicin. Western blotting demonstrated that both p21 and Wig-1 were strongly induced by G418, whereas upregulation after gentamicin treatment was relatively weak (Figures 1B and 2B). p21 induction by G418 and gentamicin is likely to be p53-dependent, since the induction is only observed in H1299-R213X cells and not in H1299 cells transfected with empty vector (Figure 2B). These results indicate that aminoglycosides induce full-length p53 that is at least partially functional as transcription factor in HDQ-P1 and H1299-R213X cells.

Next we asked if levels of recovered full-length p53 could be further enhanced by combination treatment with agents that inhibit p53 proteasomal degradation. Combination of G418 with the proteasome inhibitor bortezomib at 10 nM increased full-length p53 levels upon treatment with G418 or gentamicin according to Western blotting (Figure 3A, upper panel). Bortezomib itself did not enhance full-length p53 expression in these cells, but caused a detectable increase in the levels of truncated p53. Combination of aminoglycosides and bortezomib also enhanced expression of the p53 target p21 as shown by Western blotting (Figure 3A, upper panel). We reasoned that specific inhibition of Mdm2-mediated proteasomal degradation of p53 might further enhance p53 levels. Thus, we treated the cells with a combination of aminoglycosides and the Mdm2 inhibitor nutlin-3a (36). This combination showed strong synergy and increased the levels of full-length p53 as well as p21 in a dose-dependent manner (Figure 3A, middle panel). Nutlin-3a itself did not have any effect on the expression of p53 and p21. We also tested MI-773, another inhibitor of p53-Mdm2 binding (37). MI-773 was at least as potent as nutlin-3a in enhancing full-length p53 protein levels in combination with G418 or gentamicin (Figure 3A, lower panel).

We subsequently examined expression of p53 and p53 target genes at the mRNA level upon combination treatment with aminoglycosides and nutlin-3a. As shown in Figure 3B, the combination treatment resulted in a pronounced synergistic increase in expression of p53 as well as p53 targets p21, Fas, Mdm2, and Wig-1 at the mRNA level. Statistically significant differences (p < 0.001) between combination treatment and G418 alone were observed for p53, p21, Fas, Mdm2, and Wig-1 according to the post hoc Tukey test in the ANOVA model. Combination treatment with gentamicin and nutlin-3a caused a statistically significant increase in p53 mRNA (p < 0.001) compared with gentamicin alone according to the post hoc Tukey test in the ANOVA model. Expression of p53 targets Wig-1, p21, and Mdm2 also showed a significant increase (p < 0.05) in mRNA levels following combination treatment compared with single treatment with gentamicin. Statistical analysis was done using the Wilcoxon matched pairs test since the distributions differed significantly from the normal one. Nutlin-3a itself did not induce any substantial increase in mRNA levels of p53 nor p53 target genes. Since G418 shows stronger readthrough effect than gentamicin and since bortezomib has higher general toxicity than nutlin-3a, we focused on the combination of G418 and nutlin-3a for the following studies. According to Western blotting (Figure 3C), the effect of the combined treatment with nutlin-3a and G418 on p21 expression is dependent on p53, since H1299 cells carrying empty vector did not show any increase in p21.

In addition, we tested if G418, nutlin-3a, and G418 in combination with nutlin-3a would affect the expression of p53 and p53 target genes in wild-type p53-carrying HCT116 cells. G418 itself did not cause any substantial induction of p53 or p53 target genes in the HCT116 wtp53^+/+ cells, except at the highest concentration (400 µM) where a modest induction of p53 and p21 was observed (Figure S1A in Supplementary Material). This could be due to cellular toxicity since we observed increased cell death at this concentration. As expected, nutlin-3a alone induced p53 protein but not mRNA in the HCT116 cells, and induced p53 target genes at both protein and mRNA levels (Figures S1B,C in Supplementary Material). G418 did not further enhance induction of p53 protein by nutlin-3a, although we observed increased levels of p21 and Fas mRNA upon combination treatment with nutlin-3a and G418. The differences between nutlin-3a-treated and control samples were statistically significant (p < 0.001) for Mdm2, p21, and Fas mRNA according to the post hoc Tukey multiple comparison in the ANOVA model (Figure S1C in Supplementary Material).

We further studied the ability of full-length p53 induced by aminoglycosides to inhibit cell growth and trigger cell death in H1299 cells carrying the R213X p53 construct. First, we assessed the effect on cell growth using the WST-1 proliferation assay. Induction of full-length p53 by G418 resulted in a marked inhibition of cell growth (Figure 4A). Cells carrying the R213XΔC-EGFP fusion construct or empty vector were less sensitive to G418, confirming that the observed effect is dependent on full-length p53. We observed statistically significant differences (p < 0.05) in G418-induced growth inhibition between H1299 cells carrying the R213X p53 construct and H1299 cells carrying empty vector according to the Kruskal–Wallis test. We also examined the cell cycle distribution after treatment with G418. As shown in Figure 4B, G418 induced G1 cell cycle arrest in the R213X-carrying H1299 cells but had no major effect in the H1299 empty vector control cells, indicating that the induction of G1 arrest by G418 is p53-dependent. The difference in G1 population between G418-treated and control R213X-carrying H1299 cells was statistically significant (p < 0.05), while the difference between G418-treated and control H1299 empty vector cells was not significant according to the Unequal N HSD post hoc test in the ANOVA model test.

We finally tested the effect of the combination of aminoglycosides and nutlin-3a on cell growth and survival in H1299 cells carrying R213X mutant p53. Analysis of cell growth by the WST-1 assay demonstrated a significant increase in growth suppression upon treatment with G418 in combination with nutlin-3a in H1299-R213X cells, when compared with treatment with G418 alone. The difference between combination treatment and G418 alone at 50 µM was statistically significant in H1299-R213X cells (p < 0.05) but not in H1299 empty vector cells according to the post hoc Tukey analysis in the ANOVA model. This indicates significant synergy (Figure 4A). Importantly, nutlin-3a did not induce any further growth suppression in the H1299 R213XΔC-EGFP and empty vector control cells, confirming that this synergistic growth-inhibitory effect is p53-dependent.

Moreover, we analyzed the effect of the combined treatment with G418 and nutlin-3a on cell cycle distribution by FACS-PI. As shown in Figure 4B, nutlin-3a did not enhance G1 arrest induced by G418. However, nutlin-3a was potent in enhancing the induction of a sub-G1 cell population in combination with G418 in H1299-R213X cells when compared with control or nutlin-3a treated cells. The difference reached statistical significance (p < 0.001) according to the Unequal N HSD post hoc test in the ANOVA model (Figure 4C). This synergistic effect on sub-G1 population induction is p53-dependent, since we did not observe any synergistic increase in H1299 empty vector control cells. Thus, we observed a statistically significant increase in sub-G1 population for both G418 and G418 in combination with nutlin-3a in H1299-R213X cells (p < 0.05, unequal N HSD post hoc test in ANOVA model).