Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The large family of papillomaviruses are important human pathogens that infect epithelial tissues at preferred body sites, typically causing papillomas or condylomata. Over three dozen human papillomavirus types are tropic for genital mucosal epithelia, and some of the high-risk genotypes can cause dysplasias and carcinomas. Among the low-risk, mucosotropic human papillomaviruses (HPVs), type 6 and 11 are additionally tropic for airway epithelium causing recurrent respiratory papillomatosis (RRP) in children and adults (for reviews, see References 1 and 2). An insidious aspect of HPV infections is their persistence for years and decades. Papillomas may regress, but the viral DNA characteristically is maintained subclinically as episomes in stem cells and often reactivates during transient or long-term immune suppression (3). Children with untreated laryngeal papillomatosis may succumb to suffocation. With advanced medical care systems, RRP becomes a chronic disease that can require four to eight (or more) laser operations per year to maintain an airway and the voice. The repeated surgeries to clear the airway in these patients invariably reactivate HPV in latently infected surrounding tissues, making the infection recurrent and persistent (4). A small percentage of the children eventually develop severe pulmonary involvement by the age of 20, causing obstruction of air exchange, cystic destruction of the lungs, pneumonias, and death. In situ hybridization of tissue specimens from trachea and lung parenchyma indicate that HPV gene expression is highly active (Broker and colleagues, unpublished results). In addition, there are reports that some adenocarcinomas, squamous carcinomas, and small-cell carcinomas of the lung also contain HPV-6, HPV-11, or the genital tropic HPV-16 and HPV-18 DNA (5). Although the number of patients who suffer from RRP or have pulmonary involvement is relatively small compared with those with HPV genital tract infections, the lack of vaccines or effective treatment, short of surgery, makes the airway infection a devastating, lifelong disease. Thus, an improved understanding of HPV replication and transactivation in airway epithelial cells has direct medical significance.

HPV DNA replication depends on origin sequences and virally derived proteins complemented by host enzymes that synthesize deoxyribonucleoside triphosphates, and host DNA replication proteins, including DNA polymerases, proliferating cell nuclear antigen, topoisomerases, and others (15). In transiently transfected cells and in cell-free replication assays, three HPV components have been implicated in this process: the origin of replication (ori) sequences, and the origin recognition and initiator proteins E1 and E2 (for reviews, see References 16 and 17). Whether these same elements are necessary or sufficient for plasmid replication in human airway cells is not known. The ori sequence is located within a noncoding region, called the upstream regulatory region (URR), of approximately 0.7 kb. In this region of all mucosotropic HPV types, there are four copies of the E2 protein binding site (E2BS) and one copy of the E1 protein binding site (E1BS). The E1 protein binds to E1BS with low affinity. It functions as helicase and recruits the host DNA polymerase . The E2 protein is also a transcription factor, associates with the E2BS with high affinity, and can prevent nucleosome formation around the origin. It helps recruit and stabilize E1 binding to the E1BS, thereby facilitating the assembly of the replication preinitiation complex on the origin (18). Viral E6 and E7 oncoproteins that have the ability to dysregulate host cell growth and differentiation and are responsible for initiating carcinogenesis are not directly required for the process of viral DNA replication (reviewed in Reference 19). The URR is also the main transcription regulatory region and contains many host transcription-factor binding sites as well as one or more promoters. HPV E2 proteins have been shown to modestly activate or repress HPV promoters (in some cases by 4- to 5-fold), depending on the sequence context or the strength of the E2 expression vector (reviewed in References 20 and 21). No transcriptional regulation of the URR in airway cells has been reported previously.

From another perspective, the tissue tropism and recalcitrant nature of HPV infections could potentially be useful in augmenting gene expression and persistence in transfected airway cells, but these possibilities have not been examined previously. In the present study, we performed experiments to address the following questions: (1) Does the URR of HPV-11 alone confer transcriptional activation in human airway cells? (2) Can either E1 or E2 protein regulate the transcriptional activity of the URR in these cells? (3) Can this first-generation three-plasmid system promote transcriptional upregulation and plasmid replication in airway cells in vitro and in mouse lungs in vivo? We interpret these results from the standpoint of HPV pathogenesis in lung cells, and also consider ways in which this new information might be applied to improve the magnitude or duration of gene transfer to airway epithelial tissues.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Plasmids

pMT2-11URRE1 and pMT2-11URRE2 were constructed by inserting a fragment spanning the HPV-11 URR (nucleotides 7072-7933/1-99) removed from clone 23-3 (22) with HindIII digestion into the StuI site in the SV40 enhancer in pMT2-11E1 and pMT2-11 E2 (23). To prepare p11LUC, the HPV-11 URR previously cloned into the shuttle vector pGEM-1 (Promega Corp., Madison, WI) was recovered from a BamHI-NheI double digestion. The URR-containing fragment was then inserted into the BglII/NheI sites in pGL3c (Promega), which expresses luciferase under the control of the SV40 early promoter and enhancer. The structures of these new plasmids are shown in Figure 1. In studies conducted in cell lines, a total of at least 20 plasmid preparations (at least four plasmid preparations for each of the four constructs) were used. The results were not influenced by a particular plasmid preparation, and were consistent when repeated for each new plasmid isolation.

View larger version (14K): [in this window] [in a new window]   Figure 1.   Structures of p11LUC, pMT2-11URRE1, and pMT2-11URRE2. Schematic representations of plasmids used in this study. p11LUC is derived from pGL3c (Promega) by inserting the 1-kb HPV-11 URR into the parental clone. pMT2-11URRE1 and pMT2-11URRE2 are expression vectors derived from pMT2- 11E1 and pMT2-11E2 (23), respectively, by inserting the approximately 1-kb HPV-11 URR into the parental vectors as described in MATERIALS AND METHODS.

Gene Transfer in Cell Lines and In Murine Airway Cells

IB3-1 cells, airway cells derived from a patient with cystic fibrosis (CF) (a kind gift from Dr. Pam Zeitlin of the Johns Hopkins University, Baltimore, MD), were grown in 48-well trays in LHC-8 media (Biofluids, Inc., Rockville, MD) for luciferase expression studies. The DNAs (pGL3c, p11LUC, pMT2-11URRE1, pMT2-11URRE2, and carrier DNA) were complexed with DOTAP:DOPE cationic liposomes (Avanti, Birmingham, AL) in a 4:1 lipid-to-DNA molar ratio. Each well was transfected with 150 ng of pGL3c, or 150 ng of p11LUC with either 600 ng each of E1 and/or E2 or with an equal amount of carrier plasmid DNA. The use of carrier DNA ensured the consistency of the total amount of transfected DNA as well as the ratio of lipid to DNA in all transfections. All transfections were performed in triplicate. Within any one experiment, aliquots of the same batch of IB3-1 cells were transfected under identical conditions. BEP1E and BEP2D airway cells were a gift from Dr. Curtis Harris of the National Institutes of Health (Bethesda, MD).

For in vivo gene transfer into FVB/N-C57BL/6 mice, lipid GL-67 (Genzyme Corp., Farmington, CT) was complexed with the DNA of interest. The ratio of lipid to plasmid DNA has previously been optimized to be 1:6 (24). Briefly, 108 µg of GL-67, 250 µg of pGL3, and 200 µg each of pMT2-11E1 and pMT2-11E2 or, alternatively, equivalent amounts of carrier DNA, were mixed in 550 µl. After a 15-min incubation at 30°C, 100 µl of the lipid-DNA mixture were pipetted on the tip of the nares of lightly anesthetized FVB/N-C57BL/6 mice over the duration of 1 min. Three days later, mice were killed by CO₂ euthanasia and their lungs and tracheas were removed. The tissues were processed as previously described (25). Low molecular- weight DNA was extracted from 10 to 30 mg of the tissues (RNeasy Kit; Qiagen Corp., Valencia, CA) for replication assays as described below. The kit is suitable for extracting both nuclear RNA and plasmid-based DNA.

Plasmid Replication Assays

Polymerase chain reaction (PCR)-based assays were used to detect plasmid replication in IB3-1 cells and in mouse airway tissues. Briefly, low molecular-weight DNA was harvested and then digested with restriction enzyme StuI or enzymes StuI plus DpnI as described previously (23). Thirty cycles of PCR amplification were then conducted using primers 5'-TAAAAAGCTTATGGAAGACGCCAAAAACATAAAGAAA-3' and 5'-GCCCAAGCTTATCGATTACACGGCGATCTTTCCGCCCTTCTT that target a 1.6-kb region of the luciferase gene. Each cycle consisted of 30 s at 95°C for denaturation, 30 s at 60°C for annealing, and 90 s at 72°C for polymerization. The products were separated in a 0.8% agarose gel and revealed by ethidium bromide staining. A 1.6-kb PCR product amplified from the pGL3c plasmid was used as a positive control.

Luciferase Assay

At 4 d after transfection of IB3-1 cells, cell lysates were harvested with 50 µl of luciferase assay buffer (Promega) and centrifuged to remove debris. The lysates were combined with luciferin reagent and read for light activity in a luminometer with signal integration for 10 s.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Upregulation of Reporter Expression from a Vector Containing the HPV-11 URR in a Three-Plasmid Replication System

We introduced a 1-kb segment of HPV-11 DNA containing the entire HPV-11 URR into a plasmid in which luciferase reporter gene expression was driven by the SV40 enhancer and early promoter (p11LUC, Figure 1). The 1-kb HPV-11 URR was also introduced into the E1 and E2 protein expression vectors, pMT2-11URRE1 and pMT2-11URRE2 (Figure 1). Unlike the parental E1 and E2 expression vectors that do not have URRs and do not replicate (23), these two modified expression vectors can support the replication of not only a third URR-containing plasmid (see below) but also themselves (data not shown).

The effects of each HPV component on reporter gene expression were tested in IB3-1 airway cells, a cell line established from a patient with CF. A cationic lipid (DOTAP:DOPE) was used to mediate gene transfer into these cells (Figure 2). This cell line was chosen because it is one of a few human airway cell lines that does not contain the SV40 T antigen. Because the plasmids in our study contain the origin of SV40 replication, the T antigen would support their replication, confounding the interpretation of experiments designed to test HPV components. The method of DNA transfection was chosen because it is reproducible and efficient for airway cells. Using the bacterial LacZ gene product -galactosidase as a reporter, we have consistently observed that 5 to 10% of IB3-1 cells are positive for reporter activity by this technique.

View larger version (33K): [in this window] [in a new window]   Figure 2.   Augmentation of luciferase expression by HPV-11 cis elements and trans factors in the human IB3-1 CF cell line. IB3-1 cells were transfected with DOTAP:DOPE cationic liposomes and various plasmids (+). The activity from the parental clone pGL3c was used as a reference. Transfection with only E1 and E2 expression vectors served as a negative control. At 4 d after transfection, luciferase activities from triplicate dishes were determined and plotted on a log scale. Bars indicate standard error.

Interestingly, the presence of HPV-11 URR by itself increased luciferase expression by 10- to 15-fold relative to pGL3c, the parental plasmid (Figure 2). Thus, the URR functions as a powerful enhancer in human airway epithelial cells in response to host transcription factors. Previous studies have shown that HPV URR is active in human dermal keratinocytes or epithelial cell lines, and inactive or less active in monkey epithelial cells, human mammary tumor cell lines, or human or mouse fibroblasts (22, 26). Collectively, these results support the hypothesis that the HPV-11 URR might contain sequences that confer transcriptional tissue tropism to human epithelial cells, including airway epithelial cells. When p11LUC was cotransfected with the E1 expression plasmid alone, there was a moderate increase in the luciferase activity, about 7-fold over p11LUC or 100-fold over pGL3c. A transcription activation by an HPV E1 protein has not been reported previously. When p11LUC was cotransfected with E2 expression plasmid, there was an almost 500-fold increase in luciferase activity compared with pGL3c, or a 33-fold increase relative to that of p11LUC. This observation is consistent with the interpretation that, with its four copies of E2BS, the URR acts as a strong E2-responsive enhancer in the presence of the E2 protein. When all three plasmids were cotransfected into the IB3-1 cells, the luciferase activity increased approximately 330-fold relative to p11LUC or 5,000-fold over pGL3c (Figure 2). Similar transcriptional transactivation was also obtained with two other non-CF human airway epithelial cell lines (BEP1E and BEP2D; transactivation by two to three orders of magnitude). In contrast, cotransfection of E1 and E2 expression plasmid with the parental reporter plasmid (without HPV-11 URR) elicited a much smaller (10-fold) stimulation of reporter activity (data not shown). Collectively, these results indicate pronounced enhancement of reporter gene expression in human airway cells by incorporating HPV-11 cis sequence elements, trans-acting factors, or both. Such enhancement might be attributable to transcriptional transactivation, plasmid replication, or both (see below).

Evidence of Plasmid Replication Using a Three-Plasmid System in CF Airway Epithelial Cells

Partly due to the limitation of the sensitivity of Southern blot hybridization and partly due to a reduced cotransfection efficiency of three plasmids relative to one plasmid, transient replication by Southern blotting can be detected only when at least 10 to 20% of cells are transfected (as estimated by a LacZ reporter gene [31]). Not unexpectedly, our attempts to detect replicating p11LUC by Southern blot hybridization in transfected IB3-1 cells were unsuccessful. To circumvent this limitation of low transfection efficiency, we adapted a PCR-based method to investigate plasmid replication in these cells. PCR amplification is an established method for demonstrating plasmid persistence in mouse tissues after injection with DNA liposomes (25).

IB3-1 cells were transfected with different combinations of the reporter plasmids, pMT2-11URRE1 and pMT2- 11URRE2. Low molecular-weight DNA was harvested 4 d after transfection, digested with either StuI to linearize the reporter plasmid or with StuI and DpnI. DpnI recognizes sites that are methylated in bacteria and cuts input plasmids numerous times. The DpnI sites in DNA replicated in eukaryotic cells are not methylated and are thus resistant to the digestion. PCR amplification of digested low molecular-weight DNA with primers targeted to a 1.6-kb region of the luciferase complementary DNA (cDNA) was performed as described in MATERIALS AND METHODS. Because of the presence of seven DpnI sites in the luciferase cDNA, the probability of amplifying an uncut luciferase cDNA from the input plasmid was extremely low. Consistent with this predication, a 1.6-kb PCR product was detected after DpnI digestion of low molecular- weight DNA only after the cotransfection of p11LUC and E1 and E2 expression plasmids (Figure 3, lane 3 versus lane 9), but not after transfection of pGL3c alone (lanes 1 versus 7), p11LUC alone (lanes 2 versus 8) or together with either E1 (lanes 5 versus 11) or E2 (lanes 6 versus 12) expression vector, or with the E1 and E2 expression vectors in the absence of p11LUC (lanes 4 versus 10). These results indicate that plasmid replication in human airway cells requires three components: the URR in cis, and E1 and E2 proteins that function in trans. To substantiate this conclusion, transient replication of p11LUC was confirmed by Southern blot hybridization of low molecular- weight DNA recovered from transfected human 293 cells without prior PCR amplification. This cell line was readily transfected to over 20% efficiency (data not shown).

View larger version (33K): [in this window] [in a new window]   Figure 3.   Replication of p11LUC in IB3-1 cells in the presence of pMT2-11URRE1 and pMT2-11URRE2. At 4 d after transfection of plasmid or plasmids (X), the low molecular-weight DNA was harvested and digested with StuI alone (lanes 1-6) or with StuI plus DpnI (X; lanes 7-12). Thirty cycles of PCR amplification were then conducted as described in MATERIALS AND METHODS to detect DpnI-resistant replicated input DNA. DNA size markers are shown in lanes marked M. The arrow points to the 1.6-kb products from PCR amplification.

Three-Plasmid System in Mouse Airways

We next tested transcriptional activation in mouse lungs with plasmids in complex with lipid GL-67. Only when all three URR-containing plasmids were cotransfected did we detect luciferase activity. The inactivity of p11LUC or lack of transactivation of p11LUC by E1 alone or E2 alone may be attributed to several factors. For example, the parental plasmid's yielding no detectable activity could indicate that the SV40 promoter is rather inactive in this setting, the transfection efficiency was low, or both. The URR has a low activity in mouse cells (Zhao, Chow, and Broker, unpublished results) and by itself may not be able to augment the SV40 promoter to detectable levels. Thus, transactivation of p11URR by E1 or E2 might be anticipated to be much less in murine compared with human cells. This has been our experience in transactivation of URR by E2 in monkey CV1 cells (22, 30). Nevertheless, the result that the reporter activity was highest when all three plasmids were cotransfected is consistent with results from IB3-1 cells (Figure 2). We used the DpnI/PCR amplification method to detect plasmid replication in proliferating cells within the mouse airway. The low molecular-weight DNA from mouse airway tissues was extracted 3 d after transfection with plasmids in complex with lipid GL-67. After StuI/DpnI digestions, PCR amplification generated a 1.6-kb fragment when all three plasmids were cotransfected (Figure 4, lanes 3 and 4) but not when the pGL3c was transfected alone (Figure 4, lanes 1 and 2). These results indicated that at least some proliferating cells in the mouse lung were cotransfected by all three plasmids, resulting in transient reporter plasmid replication and augmented reporter gene signal. However, whether the enhanced reporter activity in our studies was due to plasmid replication alone or was additionally caused by synergistic transcriptional activation from the E1 and E2 proteins cannot be determined.

View larger version (52K): [in this window] [in a new window]   Figure 4.   Representative replication of p11LUC in mouse lung tissues in the presence of pMT2-11URRE1 and pMT2-11URRE2. At 3 d after transfection with plasmid or plasmids complexed with lipid 67 as indicated, low molecular-weight DNA was harvested and digested with StuI alone () or with StuI and DpnI (+). DpnI-resistant, newly replicated luciferase plasmid DNA was revealed after PCR amplification as described in Figure 3. Lanes 1 and 2, pGL3c plus carrier plasmid; lanes 3 and 4, p11LUC, pMT2-11URRE1, pMT2-11URRE2; lane C, PCR product from p11LUC as a positive control; lane M, DNA size markers. The arrow points to the 1.6-kb products from PCR amplification.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The results presented here establish that HPV sequence elements and protein products can augment reporter gene expression by several orders of magnitude in airway epithelial cells in vitro (Figure 2). The incorporation of the HPV-11 URR alone increased luciferase activity over the parental vector, in which the reporter is under the control of a fully active SV40 early promoter. This result indicates that the URR acts as a strong enhancer in airway cells. Cotransfection with pMT2-11URRE2 strongly stimulated p11LUC. The fold of stimulation was much higher than our previous observations when the HPV-11 URR was used to activate a minimal SV40 promoter in nonrespiratory cell lines (22, 30). We suggest that presence of the URR in the pMT2 protein expression vector (Figure 1), together with the strong enhancer activity of the URR in airway cells (Figure 2), may have produced high levels of E2 protein and augmented transcription activation. The HPV E1 protein has not been previously reported to modulate the URR transcriptional activity. Therefore it was quite unexpected that cotransfection with pMT2-11URRE1 also resulted in a significant stimulation of the reporter activity, though not as high as that achieved by the E2 expression vector. Because E1 alone did not support reporter plasmid replication (Figure 3; compare lanes 5 and 11), one possible explanation is that at high concentrations of E1 protein expressed from the URR-containing vector, E1 was able to associate with the origin efficiently, alter the nucleosome structure, and function as a transcription factor. Cotransfection with E1 and E2 expression vectors had a synergistic rather than an additive effect on augmenting luciferase activity. Thus, there may also be a collaboration in the transcriptional upregulation by the E1 and E2 proteins.

We consider it unlikely that the large increases in luciferase signal in these studies (by several orders of magnitude; Figure 2) can be attributed to differences in transfection efficiency. In IB3-1 cells, the transfection efficiency, as judged by measurable LacZ reporter signals, has consistently been between 5 and 10% within any one experiment. Moreover, all transfections were internally controlled. Within any one experiment, aliquots of the same batch of IB3-1 cells were tranfected using identical conditions. A carrier plasmid DNA was substituted for the omitted plasmids to keep the total amount of transfected DNA as well as the lipid-DNA ratio constant.

An alternative interpretation for the synergistic transcriptional activation of the reporter in p11LUC by E1 and E2 proteins is reporter plasmid replication in the presence of the HPV proteins. To confirm plasmid replication in the setting of low transfection efficiency in CF cells, we developed a DpnI/PCR amplification method and demonstrated that p11LUC plasmid indeed replicated in these cells (Figure 3). By testing different combinations of plasmids, we show that replication in these cells requires URR, which is known to contain an origin of replication, and E1 and E2 proteins. Further, by using this modified replication assay, evidence of replication in cells from mouse lung tissues was also obtained. When the E1 and E2 expression vectors were omitted, no replication was detected (Figure 4).

The experiments described here suggest that the strong transcriptional transactivation in human airway cells conferred by the HPV cis elements and trans factors might contribute to the aggressive nature of this HPV type in airway epithelium and in lungs. We also note that an improved understanding of the mechanisms by which HPV augments gene expression and persistence in airway cells could be applied to experimental trials of pulmonary gene transfer. For example, using either virus- or nonvirus- mediated approaches, gene transfer in CF airways has been limited by low levels of expression and by lack of transgene persistence. Considerable efforts have been devoted to improving methods for DNA delivery into airway cells (24, 32, 33). Our study addresses features that historically have limited gene transfer in both the preclinical and clinical settings: we suggest that elevated gene expression and extrachromosomal plasmid replication might be accomplished in patient airways by incorporation of HPV components into plasmid DNA molecules, as shown in the human airways cell lines and in mouse lung. We note that plasmids containing the oriP of the Epstein-Barr virus (EBV) can also replicate in proliferating cells in the presence of the EBV-encoded nuclear antigen-1 protein (34). Similar to the HPV E2 protein, EBNA-1 can also transactivate its promoter via elements in the oriP in a cell type- specific manner (35). The extent to which this EBV system might activate a reporter gene in airway cells has not been investigated.

Difficulties transfecting proliferating airway cells in vivo have been noted in the past, and the cell types transfected are often not known with certainty (24, 33). In addition, the development of single-plasmid replicons and characterization of the in vivo immune response to transfected DNA and transgene expression will be required before it is clear whether an HPV-based strategy could contribute to applications in CF or other human lung diseases. The experiments shown here describe a means by which HPV infection of human airway cells, and the possible use of HPV elements in studies of gene transfer, can be more fully explored.