Recombinant adenoviral vectors can induce expression of p73 via the E4-orf6/7 protein

Abstract

Despite the utility of recombinant adenoviral vectors in basic research, their therapeutic promise remains unfulfilled. Most engineered adenoviral vectors use a heterologous promoter to transcribe a foreign gene. We show that adenoviruses containing the cytomegalovirus immediate-early promoter induce the expression of the proapoptotic cellular protein TAp73 via the cyclin-dependent kinase-retinoblastoma protein-E2F pathway in murine embryonic fibroblasts. Cells transduced with these vectors also expressed high levels of the adenoviral E4-orf6/7 and E2A proteins. By contrast, adenoviruses containing the ubiquitin C promoter failed to elicit these effects. E4-orf6/7 is necessary and sufficient for increased TAp73 expression, as shown by using retrovirus-mediated E4-orf6/7 expression and adenovirus with the E4-orf6/7 gene deleted. Activation of TAp73 likely occurs via E4-orf6/7-induced dimerization of E2F and subsequent binding to the inverted E2F-responsive elements within the TAp73 promoter. In addition, adenoviral vectors containing the cytomegalovirus immediate-early promoter, but not the ubiquitin C promoter, cooperated with chemotherapeutic agents to decrease cellularity in vitro. In contrast to murine embryonic fibroblasts, adenoviruses containing the ubiquitin C promoter, but not the cytomegalovirus immediate-early promoter, induced both E4-orf6/7 and TAp73 in human foreskin fibroblasts, emphasizing the importance of cellular context for promoter-dependent effects. Because TAp73 is important for the efficacy of chemotherapy, adenoviruses that increase TAp73 expression may enhance cancer therapies by promoting apoptosis. However, such adenoviruses may impair the long-term survival of transduced cells during gene replacement therapies. Our findings reveal previously unknown effects of foreign promoters in recombinant adenoviral vectors and suggest means to improve the utility of engineered adenoviruses by better controlling their impact on viral and cellular gene expression.

FIG. 1.

Adenoviral vectors used in these studies are presented schematically, with modifications of the E1 and E4 regions indicated, along with the location of the pIX gene (Ptn IX). Note that all vectors have the E1A and E1B genes deleted. Not drawn to scale.

FIG. 2.

Adenoviral vectors expressing GFP from the CMV promoter, but not the UbC promoter, induce TAp73 expression. MEFs suspended in serum-free medium were transduced with adenovirus at the indicated MOI, plated at subconfluency for 16 h in complete medium, changed to starvation medium (0.2% serum) for 2 days to synchronize their cell cycles, and then fed for 22 h in complete medium prior to harvest. (A) The transcript levels of TAp73 relative to 18S rRNA were determined by real-time RT-PCR. Indistinguishable results were obtained by normalization to β-actin. Parallel samples were used for GFP analysis (gray diamonds). The standard error of the mean is included (obscured when the error is very small). (B) TAp73 protein was detected by immunoblotting. A nonspecific band served as the loading control (lower blots).

FIG. 3.

Adenoviral vectors expressing GFP from the CMV promoter, but not the UbC promoter, induce ΔNp73 expression. (A) MEFs were transduced with adenovirus at the indicated MOI, serum starved, and then restimulated with serum-containing medium as described in the legend to Fig. 2. The transcript levels of ΔNp73 relative to 18S rRNA were determined by real-time RT-PCR. Parallel samples were used for GFP analysis (gray diamonds). The standard error of the mean is included (obscured when the error is very small). (B) EL4 thymoma cells expressing the coxsackievirus/adenovirus receptor required for adenoviral entry (26) were transduced over a wide range of MOIs, and the percentage that expressed GFP was quantified via flow cytometry 16 h later.

FIG. 4.

Neither GFP expression nor adenoviral DNA replication is responsible for induction of TAp73 and adenoviral pIX RNA. MEFs were transduced as before with adenoviral vectors that (A) possess the CMV promoter but do not encode GFP (or any foreign gene) or (B) express GFP from the CMV promoter but cannot replicate due to deletion of the pTP gene. TAp73 mRNA (black bars, left y axis) and adenoviral pIX RNA (gray bars, right y axis) relative to 18S rRNA were quantified by real-time RT-PCR (ND, not determined; “DEAD” indicates that cell killing at these MOIs prevented mRNA analyses). (C) Adenoviral pIX is not responsible for increased transcript levels of TAp73. Prior to adenoviral transductions (MOI, 10), MEFs were transduced with an MSCV-based retrovirus expressing GFP alone or in conjunction with adenoviral pIX. TAp73 mRNA relative to 18S rRNA was quantified by real-time RT-PCR. The expression of pIX mRNA was confirmed by RT-PCR and found to be substantially higher than levels observed in cells transduced with Ad CMV-GFP (data not shown). The standard error of the mean is included (obscured when the error is very small).

FIG. 5.

CDK-pRb-E2F pathway is required for adenoviral induction of TAp73. (A) MEFs were transduced as before (MOI, 10), and TAp73, CycA2, DHFR, and Cdc6 RNAs relative to 18S rRNA were quantified by real-time RT-PCR. (B) MEFs were incubated in 30 μM roscovitine or dimethyl sulfoxide (DMSO) control, and CycA2, DHFR, and Cdc6 RNAs relative to 18S rRNA were quantified by real-time RT-PCR. (C) MEFs were transduced as before (MOI, 10), except the 0.2% starvation medium was not replaced or (D) roscovitine or DMSO control was added when the starvation medium was replaced with 10% FBS-DMEM. (E) MEFs were transduced with the indicated adenoviral vectors at the indicated MOIs. Additional Ad CMV-GFP was added to equalize the MOIs when Ad CMV-p16 was used. TAp73 mRNA (black bars, left y axis) and adenoviral pIX RNA (gray bars, right y axis) relative to 18S rRNA were quantified by real-time RT-PCR. The standard error of the mean is included (obscured when the error is very small).

FIG. 6.

Adenoviral protein E4-orf6/7 is necessary and sufficient to induce p73. (A and B) MEFs were transduced with the indicated viruses or (C) prior to adenoviral transductions, MEFs were transduced with an MSCV-based retrovirus expressing GFP alone or in conjunction with E4-orf6/7. TAp73 RNA (black bars, left y axis) and adenoviral pIX RNA (gray bars, right y axis) relative to 18S rRNA were quantified by real-time RT-PCR. The standard error of the mean is included (obscured when the error is very small). “DEAD” indicates that cell killing at these MOIs prevented mRNA analyses. (D) Adenoviral DNA DBP and E4 proteins were detected by immunoblotting. β-Actin detection served as the loading control. (E) Adenoviral DBP and E4-orf6/7 protein were detected by immunoblotting. β-Actin served as the loading control.

FIG. 7.

Adenoviral vectors possessing only endogenous elements induce adenoviral pIX and TAp73 expression. MEFs were transduced as before with adenoviral vectors that maintain the natural E1A promoter but lack foreign promoter sequences. As indicated in Fig. 1, the two adenoviral vectors used differed in that only the first possessed an SV40 polyadenylation sequence between the E1A promoter and the pIX coding sequence. TAp73 mRNA (black bars, left y axis) and adenoviral pIX RNA (gray bars, right y axis) relative to 18S rRNA were quantified by real-time RT-PCR. “DEAD” indicates that cell killing at these MOIs prevented mRNA analyses. For both viruses, a linear correlation (correlation coefficient, 0.99) exists between pIX RNA and TAp73 induction as a function of the MOI used for transduction being decreased. The standard error of the mean is included (obscured when the error is very small).

FIG. 8.

Chemotherapeutics and Ad CMV-GFP cooperatively decrease cellularity. MEFs were transduced with adenovirus (MOI, 30) as before, except 0.2% FBS starvation medium was replaced with 10% FBS-DMEM containing DMSO, etoposide (10 μM), or 5-FU (100 μM). Four days following the addition of chemotherapeutics, triplicate wells were harvested and the quantity of live cells was assessed by forward and side scatter as described in Materials and Methods. The standard error of the mean is included. *, P < 0.005; **, P < 0.0005 using Student's two-sample homoscedastic t test.

FIG. 9.

Adenoviral vectors expressing GFP from the UbC promoter, but not the CMV promoter, induce E4-orf6/7 and TAp73 expression in human foreskin fibroblasts. HFFs were transduced with adenovirus (MOI, 10) as previously described. (A) The transcript levels of TAp73 relative to 18S rRNA were determined by real-time RT-PCR. The standard error of the mean is included (obscured when the error is very small). To rule out a “virus-switching” mistake, these experiments were repeated several times, the virus stocks used were the same as those used for the experiments above with MEFs, and PCR amplification of the promoters followed by sequencing was used to confirm the identities of the viral stocks. (B) E4-orf6/7 protein was detected by immunoblotting. β-Actin served as the loading control (lower blots).

FIG. 10.

Alignment of E2F-responsive sites in the TAp73, Ad5 E2A, and E2F1 promoters. (A) Model for adenoviral-vector-mediated deregulation of cellular gene expression dependent on both the choice of promoter driving ectopic gene expression and E4-orf6/7. The CMV promoter is used in this example, but the extent of E4-orf6/7-dependent deregulation of cellular gene expression mediated by a particular ectopic promoter depends on the cellular context. (B) Alignment of human and mouse TAp73, Ad5 E2A, and human E2F1 promoter sequences with predicted (for p73) or demonstrated (E2A and E2F1) E2F binding sequences (arrows).