Expression of HPV16 E5 down-modulates the TGFbeta signaling pathway

In order to investigate the possible existence of a relationship between the E5 viral protein of HPV16 and the TGFβRII expression, we first quantified the 16E5 and TGFβRII mRNA levels in various samples of low grade (LSIL) and high grade (HSIL) squamous intra-epithelial lesions by real-time RT-PCR and normalized them respect to their levels in the HPV16-positive cervical epithelial cell line W12 [19] at the passage 6 (W12p6), in which ~100 to 200 copies per cell of the E5-expressing HPV episomes were retained [20]; our unpublished data. Results showed a great variability of 16E5 transcript levels in LSILs, which could be related to a high variability of episomal/integrated HPV16 distribution (Figure 1A, upper panel) as expected [5]. The primary culture of normal human ectocervical keratinocytes (HCK) as well as the negative controls (C2-C5) were all negative for 16E5 mRNA (Figure 1A, upper panel). In contrast, as previously reported [5], in all HSILs the expression of 16E5 mRNA resulted lower if compared to W12p6, presumably as a result of viral integration. The parallel analysis of TGFβRII mRNA showed that the receptor transcript levels were also highly variable in LSILs and appeared inversely linked to those of the viral protein (Figure 1A, lower panel): in fact, particularly in P90, P227 and P608 samples, high levels of 16E5 transcript corresponded to low levels of the receptor mRNA, while in p470 and P385 samples very low levels of the viral protein transcript corresponded to high levels of TGFβRII mRNA (Figure 1A, lower panel, arrows). Only in a single sample (p267) with high levels of 16E5 mRNA the amount of TGFβRII appeared not modulated and comparable to that of control samples (Figure 1A, lower panel, arrowhead). In contrast, a significant homogeneous down-modulation of TGFβRII mRNA expression compared to either the controls and the W12p6 cells (p < 0.01 vs controls) was found in HSILs (Figure 1B, lower panel), suggesting here the possible independence of the receptor modulation from 16E5 expression when viral integration or mixed episomal/integrated states are expected with more frequency (Figure 1B upper panel) and consistent with the proposed down-modulating effect of the HPV E7 protein on TGFβRII in the mouse model [15]. Results are expressed as mean ± 95% confidence interval (CI). Statistical analysis was performed as reported in the Additional file 1 Materials and Methods. Therefore, while in early steps of HPV infection the expression of TGFβRII can be regulated by the E5 protein, in more advanced lesions, characterized by frequent viral integration and loss of E5 expression, the receptor transcription would be stably controlled by other HPV transforming proteins, such as the E7 oncogene product.

To evaluate if the expression of 16E5 alone, in the absence of other HPV proteins, could be responsible for the down-modulation of TGFβRII observed in vivo in LSILs, we performed a real-time RT-PCR analysis as above using the human keratinocyte cell line HaCaT [21] stably transfected with the construct pMSG 16E5 (HaCaT pMSG E5) [22], in which the expression of the viral protein was progressively induced, by treatment with dexamethasone 1 μM, in a time-dependent manner (6 h, 12 h, 24 h of treatment, Figure 2A, left panel). The HaCaT pMSG cells were used as negative control. 16E5 mRNA levels were normalized with respect to the amount of the viral protein transcript in W12p6. Results are expressed as mean values ± standard deviation (SD). Student’s t test was performed as reported in the Additional file 1 Materials and Methods. Results demonstrated that, upon treatment with dexamethasone TGFβRII transcript levels progressively decreased, at least up to 12 h, (Figure 2A, right panel), in cells expressing increasing amounts of 16E5 mRNA (Figure 2A, left panel), suggesting that 16E5 is directly responsible for the receptor transcript down-modulation.

To further demonstrate the role of 16E5 expression in TGFβRII down-regulation, we transiently transfected HaCaT cells using increasing amounts (0.5 μg, 1 μg, 2 μg) of pCI-neo E5-HA expression vector [23] (HaCaT E5), in order to induce an increased expression of 16E5 in a dose-dependent manner (Figure 2B, left panel). HaCaT cells transfected with the empty vector pCI-neo (HaCaT pCI-neo) were used as negative control. Results demonstrated that the TGFβRII transcript levels progressively decreased with the increase of 16E5 mRNA amount (Figure 2A and B, right panels), demonstrating that 16E alone is able to down-modulate the receptor in a dose-dependent and time-dependent manner. Interestingly, this down-modulating effect appears more evident when the viral protein expression is obtained with 1 μg cDNA and is similar to the 16E5 expression in the W12 cell model, representing the most physiological condition.

To assess whether the down-regulation of TGFβRII expression induced by 16E5 leads to attenuation of TGFβ1 signaling, HaCaT E5 and HaCaT pCI-neo cells were stimulated with 20 ng/ml TGFβ1 for 1 h at 37°C. Western blot analysis confirmed that also the protein expression of TGFβRII is strongly reduced in the presence of E5 (Figure 3A). To analyze the TGFβ/Smad signaling, we first estimated the level of Smad2 phosphorylation in cells stimulated or not with TGFβ1 as above and we found that the ligand-dependent phosphorylation of Smad2 was decreased in cells expressing E5, although the Smad2 total protein amount was not modified (Figure 3A). Densitometric analysis and Student’s test were performed as reported in the Additional file 1 Material and Methods. Since it is well known that, following phosphorylation, Smad2 binds to Smad4 and the Smad2/Smad4 complexes translocate into the nucleus for gene expression regulation [9], we wondered if 16E5 expression could also affect this late event of TGFβ1-mediated signaling. Quantitative double immunofluorescence analysis performed as reported in the Additional file 1 Materials and Methods, showed that, among the cells clearly positive for E5 (Figure 3B), the percentage of cells showing Smad4 nuclear translocation upon TGFβ1 treatment (Figure 3B, arrowhead) was significantly decreased compared to the surrounding cells that did not show E5 positivity (Figure 3B). Only a minor fraction of the E5 positive cells showed a Smad4 nuclear translocation (Figure 3B, arrowhead). Thus, taken together our results indicate that 16E5 expression is able to attenuate the TGFβ1/Smad signaling and this loss of signal transduction, leading to destabilization of the epithelial homeostasis [14], may represent an additional mechanism of promotion at very early stages of the HPV-mediated cervical carcinogenesis. Since specific inhibition of the viroporin channel activity of E5 reduces the EGF stimulated mitogenic ERK signaling [24], future work might be addressed to investigate the possible interplay among the TGFβ1/Smad pathway and the EGF/ERK signaling in the presence of 16E5.