Dysregulated ΔNp63α inhibits expression of Ink4a/arf, blocks senescence, and promotes malignant conversion of keratinocytes

Abstract

p63 is critical for squamous epithelial development, and elevated levels of the ΔNp63α isoform are seen in squamous cell cancers of various organ sites. However, significant controversy exists regarding the role of p63 isoforms as oncoproteins or tumor suppressors. Here, lentiviruses were developed to drive long-term overexpression of ΔNp63α in primary keratinocytes. Elevated levels of ΔNp63α in vitro promote long-term survival and block both replicative and oncogene-induced senescence in primary keratinocytes, as evidenced by the expression of SA-β-gal and the presence of nuclear foci of heterochromatin protein 1γ. The contribution of ΔNp63α to cancer development was assessed using an in vivo grafting model of experimental skin tumorigenesis that allows distinction between benign and malignant tumors. Grafted lenti-ΔNp63α keratinocytes do not form tumors, whereas lenti-GFP/v-ras(Ha) keratinocytes develop well-differentiated papillomas. Lenti-ΔNp63α/v-ras(Ha) keratinocytes form undifferentiated carcinomas. The average volume of lenti-ΔNp63α/v-ras(Ha) tumors was significantly higher than those in the lenti-GFP/v-ras(Ha) group, consistent with increased BrdU incorporation detected by immunohistochemistry. The block in oncogene-induced senescence corresponds to sustained levels of E2F1 and phosphorylated AKT, and is associated with loss of induction of p16(ink4a)/p19(arf). The relevance of p16(ink4a)/p19(arf) loss was demonstrated in grafting studies of p19(arf)-null keratinocytes, which develop malignant carcinomas in the presence of v-ras(Ha) similar to those arising in wildtype keratinocytes that express lenti-ΔNp63α and v-ras(Ha). Our findings establish that ΔNp63α has oncogenic activity and its overexpression in human squamous cell carcinomas contributes to the malignant phenotype, and implicate its ability to regulate p16(ink4a)/p19(arf) in the process.

Figure 1. Lentivirus infection drives long term and stable gene expression in primary keratinocytes.

A, Primary mouse keratinocytes were infected with lentivirus encoding GFP, and infection efficiency was assessed by FACS analysis at day 5 post-infection. Y-axis indicates the intensity of GFP, X-axis represents forward angle light scatter. B, Primary keratinocytes were infected with lentivirus encoding GFP or ΔNp63α and whole cell protein was collected at day 5 and 14 and analyzed for p63 expression by western blot. C, Primary keratinocyte cultures expressing lenti-GFP or lenti-ΔNp63α were fixed at 14 days post-lentiviral transduction and incubated with anti-p63 antibody, followed by a secondary antibody conjugated with Alexa-488. The cells were then stained with DAPI to visualize the cell nuclei.

Figure 2. Dysregulated expression of ΔNp63α promotes survival and blocks replicative senescence of primary keratinocytes.

A, Phase morphology of lenti-GFP- and lenti-ΔNp63α expressing keratinocytes cultured for 1 and 14 days post-infection. Results shown are representative of three independent experiments. B, Quantification of BrdU positive cells in lenti-GFP or lenti-ΔNp63α cultures at timepoints noted following lentiviral gene transduction. Data shown represent the means ± S.E. of three independent experiments. * indicates a statistically significant difference between lenti-GFP and lenti-ΔNp63α infected cells at p<0.05. C, Primary keratinocytes expressing lenti-GFP (left panel) or lenti-ΔNp63α (right panel) were fixed at 14 days post-lentiviral transduction. The cell senescence status was assessed by the presence of HP-1γ nuclear foci (red). The cells were double stained with p63 antibodies (green). Matched arrows indicate the same cell stained with both different antibodies. The image presented is representative of three independent experiments.

Figure 3. Overexpression of ΔNp63α inhibits the up-regulation of p16^ink4a and p19^arf associated with cell senescence.

A, Whole cell protein was collected from primary keratinocytes expressing lenti-GFP or lenti-ΔNp63α at timepoints indicated following lentiviral gene transduction. Expression levels of p16^ink4a, p19^arf, E2F1 and p63 were detected by western blot. Equal protein loading was confirmed by immunoblotting for β-actin. B, Primary keratinocytes were infected with lentivirus encoding ΔNp63α or GFP at day 3 after plating. Total RNA was harvested at day 3, 5, 7 and 10 post-infection and reverse transcribed. Expression of p16^ink4a and p19^arf was determined by PCR amplification. PCR amplification of GAPDH was used as a loading control. C, Primary keratinocytes were cultured for 7 days after plating and then infected with adenovirus encoding ΔNp63α or LacZ. Whole cell lysates were collected at day 2, 3, 6 and 8 post-adenovirus infection (equivalent to day 9, 10, 13 and 15 post-plating). The expression levels of p16^ink4a, p19^arf and p63 were detected by western blot. Equivalent protein loading was confirmed by immunoblotting for actin.

Figure 4. Elevated expression of ΔNp63α enhances malignant conversion in v-ras^Ha-expressing keratinocytes.

A, In vivo phenotype of ΔNp63α-overexpressing keratinocytes in the presence and absence of oncogenic ras. Primary mouse keratinocytes were infected with lentivirus encoding lenti-GFP or lenti-ΔNp63α or sequentially with retrovirus encoding v-ras^Ha followed by lenti-GFP or lenti-ΔNp63α. The final tumor phenotype was assessed at 5 weeks following grafting. Final tumor volumes at 4 weeks are presented as the mean tumor volume ± S.E. B, Upper panel, representative H&E sections of tumor tissues obtained from grafting sites. Lower panel, BrdU incorporation in grafted tissue samples. C, Quantification of BrdU incorporation levels. Data are presented as the mean % of BrdU positive cells ± S.E. * indicates a statistically significant difference between lenti-GFP/v-ras^Ha and lenti-ΔNp63α/v-ras^Ha groups at p<0.05.

Figure 5. Long term ΔNp63α overexpression blocks oncogene-induced senescence.

A, Primary keratinocytes were infected sequentially with retrovirus encoding oncogenic v-ras^Ha, followed by lentivirus encoding GFP or ΔNp63α. Oncogene-induced senescence was assessed 14 days post-lentivirus infection by immunofluorescent analysis of nuclear foci of HP-1γ (upper panel), or enzymatic activity of SA-β-gal (lower panel). The images shown are representative of three independent experiments. B, Whole cell protein was collected from keratinocytes infected sequentially with retrovirus encoding oncogenic v-ras^Ha followed by lentivirus encoding GFP or ΔNp63α at timepoints indicated following lentiviral gene transduction. The levels of phosphorylated AKT, total AKT, p19^arf and p63 were detected by western blot (upper panels). p16^ink4a and E2F1 levels with corresponding p63 expression are presented in the lower panels. Equal protein loading was confirmed by immunoblotting for β-actin.

Figure 6. Overexpression of ΔNp63α supports long term survival of v-ras^Ha-expressing primary keratinocytes.

Primary keratinocytes were infected sequentially with retrovirus encoding oncogenic v-ras^Ha and lentivirus encoding GFP or ΔNp63α, as noted. On post-infection day 14, the cells were trypsinized and reseeded. A, cell morphology was recorded at day 3 after the initial reseeding. B, growth curves of keratinocytes expressing v-ras^Ha in conjunction with lenti-GFP or lenti-ΔNp63α. Cell numbers were counted at days 1, 3, 7, 10 and 14 post-lentivirus infection (passage 0), or at the same timepoints following each reseeding. Data are expressed as cell number per well and represent the means ± S.D. of two independent experiments.

Figure 7. v-ras^Ha-expressing p19^arf null keratinocytes display a tumor phenotype similar to v-ras^Ha/ΔNp63α overexpressing keratinocytes.

A, p19^arf null primary keratinocytes were infected with lenti-GFP or lenti-ΔNp63α in combination with v-ras^Ha, and grafted onto the dorsal side of nude mice as previously described . Wild type primary keratinocytes expressing lenti-GFP or lenti-ΔNp63α in combination with v-ras^Ha were used as negative and positive controls. The final tumor phenotype was assessed at 5 weeks after grafting. B, Final tumor volumes at 5 weeks are presented as the mean tumor volume ± S.E. * indicates a statistically significant difference between p19^arf null/v-ras^Ha and p19^arf null/lenti-ΔNp63α/v-ras^Ha groups at p<0.05.