Activation of Papillomavirus Late Gene Transcription and Genome Amplification upon Differentiation in Semisolid Medium Is Coincident with Expression of Involucrin and Transglutaminase but Not Keratin-10

Abstract

The life cycle of the papillomaviruses is closely linked to host cell differentiation, as demonstrated by the fact that amplification of viral DNA and transcription of late genes occur only in the suprabasal cells of a differentiated epithelium. Previous studies examining the pathogenesis of papillomavirus infections have relied on the use of organotypic raft cultures or lesions from patients to examine these differentiation-dependent viral activities. In this study, we used a simple system for epithelial differentiation to study human papillomavirus (HPV) late functions. We demonstrate that the suspension of HPV-infected keratinocytes in semisolid medium containing 1.6% methylcellulose for 24 h was sufficient for the activation of the late promoter, transcription of late genes, and amplification of viral DNA. These activities were shown to be linked to and coincide with cellular differentiation. Expression of the late protein E1^∧E4 and amplification of viral DNA were detected in the identical set of cells after suspension in methylcellulose. This technique was also used to analyze the differentiation properties of the cells which expressed the late protein E1^∧E4. While induction of the spinous layer markers involucrin and transglutaminase was compatible with late promoter induction, expression of the differentiation-specific keratin-10 was shown not to be required for HPV late functions. Interestingly, while the majority of normal human keratinocytes induced filaggrin expression by 24 h, this marker of the granular layer was induced in a smaller subset of HPV type 31 (HPV-31)-positive cells at this time point. The HPV-31-positive cells which expressed filaggrin did not induce the late protein E1^∧E4. Use of the methylcellulose system to induce epithelial differentiation coupled with the ability to perform a genetic analysis of HPV functions by using transfection of cloned viral DNA will facilitate the study of the regulation of the papillomavirus life cycle.

Human papillomaviruses (HPVs) are a group of small, double-stranded DNA viruses which infect the keratinocytes of differentiating epithelia and induce hyperproliferative lesions (17, 21, 24). Roughly one-third of the 70 identified HPV types infect the epithelium of the anogenital tract, and a subset of these are termed high risk because they are the etiologic agents of cervical cancer (31). Productive HPV infections are dependent on epithelial differentiation, and the duplication of this process in vitro has made studies of the complete viral life cycle difficult. The production of HPV virions from infected cells in the laboratory was first achieved through the use of organotypic raft cultures which are capable of inducing faithful keratinocyte differentiation (6, 25).

HPV infection is believed to occur through small wounds in the epithelium. Following entry, HPV DNA is established extrachromosomally as episomes at roughly 20 to 50 copies per cell (23). All HPV genomes contain approximately eight open reading frames which are transcribed as polycistronic messages from a single DNA strand. Regulatory sequences for early viral transcription and replication are concentrated in a small noncoding region termed the upstream regulatory region. In basal cells, transcripts from the high-risk types such as HPV type 16 (HPV-16) and HVP-31 are initiated from a promoter in the upstream regulatory region at nucleotide 97 (p97) and encode the transforming proteins of high-risk genital HPV types, E6 and E7 (2, 18, 27).

Epithelial differentiation induces a dramatic increase in viral DNA replication and late gene transcription, both of which are required for the production of progeny virions. In HPV-16 and -31, a differentiation-dependent promoter (p742 in HPV-31 and p670 in HPV-16) has been identified at the 3-prime end of the E7 open reading frame (15, 18). In organotypic raft cultures of cells productively infected with HPV-31b, transcripts initiating at nucleotide 742 and expressing the E4 and E5 open reading frames increased dramatically upon differentiation (18, 27). Consistent with this observation, levels of E4 protein have also been shown to increase upon differentiation (8, 28). The E4 protein is synthesized as a fusion with the five amino-terminal amino acids of E1 (E1^∧E4), which provides the methionine for initiation of translation. It has been suggested that E4 facilitates viral egress by causing the collapse of cytokeratin filaments (7), but these studies are controversial and its actual function remains largely undefined. Expression of the capsid protein genes L1 and L2 is dependent on initiation from the differentiation-dependent promoter (p742) but also requires differentiation-induced changes in splicing and polyadenylation site usage (19).

In addition to changes in transcription, differentiation of HPV-infected cells results in a dramatic increase in viral replication (3). Amplification of HPV DNA to thousands of copies per cell occurs in the suprabasal cells of a differentiating epithelium and is essential for the production of new virions. It has been hypothesized that viral DNA amplification and differentiation-dependent transcription may be linked since cell lines which contain HPV DNA integrated into the host cell chromosomes are unable to induce significant transcription of late genes despite full epithelial differentiation (11). This finding is consistent with the observation that replication of viral DNA and the expression of the late protein E1^∧E4 coincide in specimens from low-grade HPV-16 lesions and cutaneous warts (8).

The majority of studies which have examined the differentiation-dependent functions of HPV have relied on the use of organotypic raft cultures or lesions obtained from patients (4, 5, 19, 27). While raft cultures can be used to study the spatial distribution of HPV late functions within a differentiated epithelium, it is a time-consuming technique prone to variability. Moreover, since a mature raft culture is heterogeneous, it is not possible to isolate the specific cell layers which represent discrete stages of differentiation. We have examined the utility of an alternative method for the rapid analysis of HPV late functions. Green first demonstrated that suspension of keratinocytes in semisolid medium results in their differentiation (16). Subsequent studies have used suspension in semisolid medium to show that a major signal for keratinocyte differentiation is the disengagement of integrins from their receptors following detachment from the basement membrane (1, 30). Studies by Flores and Lambert demonstrated the utility of suspension in methylcellulose to induce a change in the mode of HPV replication following differentiation (10). However, suspension in methylcellulose is only a simplified system for the induction of keratinocyte differentiation since it does not include the effects of diffusible factors or cell-cell adhesion, which are undoubtedly important for complete differentiation (9, 12, 14). Despite this limitation, we investigated whether this system could provide sufficient differentiation for the study of the mechanisms responsible for inducing HPV late gene expression and amplification.

MATERIALS AND METHODS

Cell culture and suspension in semisolid medium.

Normal human keratinocytes (NHKs) were maintained in KGM (Clonetics, San Diego, Calif.) or in E medium with mitomycin C (Boehringer Mannheim, Indianapolis, Ind.)-treated fibroblast feeders (25, 26). NHKs were isolated from epidermis which was removed from neonatal foreskins after overnight incubation in 5 ml of dispase (2.4 U/ml) at 4°C (Boehringer Mannheim). After 15 min at 37°C in 0.25% trypsin–1 mM EDTA (Gibco BRL, Grand Island, N.Y.), epidermis was mechanically disrupted, trypsin was inactivated, and cells were plated. Transfection of NHKs to create LKP-31-1 (MR6/18wt) and LKP-31-2 (FSw-wt2) mass cell lines has been described elsewhere (11). Cervical biopsy-derived CIN-612 cells are described in reference 25. LKP-31 and CIN-612 cells were maintained in E medium with mitomycin C (Boehringer Mannheim)-treated fibroblast feeders (25, 26). Keratinocytes were suspended in 1.6% methylcellulose to induce differentiation (1, 16). To prepare methylcellulose, half of the final volume of E medium containing 5% fetal bovine serum was added to dry, autoclaved methylcellulose (4,000 cps) (Sigma, St. Louis, Mo.), and heated to 60°C for 20 min. The remaining volume consisting of E medium and an additional 10% fetal bovine serum was added, and the mixture was stirred at 4°C overnight. Three to five million cells were suspended in 1 to 2 ml of E medium and added dropwise to a 10-cm-diameter petri dish containing 25 ml of 1.6% methylcellulose. Cells were stirred with a pipette and incubated for the indicated times in a 37°C 5% CO₂ humidified incubator. Cells were harvested by being scraped into three 50-ml conical tubes containing phosphate-buffered saline (PBS; Gibco BRL) and collected by centrifugation.

Plasmids.

pBR322-HPV31 contains the HPV-31 genome inserted into the EcoRI site of pBR322 (11). pRP-p742 was constructed by amplifying the region of HPV diagrammed in Fig. 3A with primers containing BamHI sites and cloning into the BamHI site of pcDNAII (Stratagene, La Jolla, Calif.) (20). pRPA31L1 was constructed by amplifying the region of HPV diagrammed in Fig. 3B, which contains a 5-prime SalI site. A 3-prime BamHI site was created with a PCR primer, and this SalI-BamHI fragment was cloned into pSP-72 (Promega, Madison, Wis.).

FIG. 3.

Coordinate induction of messages initiated from p742 and messages containing L1. RPA was performed on total RNA isolated from LKP-31-1 cells cultured in methylcellulose (mcell.) for the indicated times. (A) RPA for messages initiated at p742. The bands of expected sizes corresponding to spliced and unspliced messages initiated at the p97 and p742 promoters are indicated. Undigested probe is shown in lane P, size markers are shown in lane M, and a control hybridization with tRNA is shown in lane t. The probe which spans the p742 promoter and the splice donor at nucleotide 877 is diagrammed. Levels of spliced p742 were quantitated by PhosphorImager analysis and are illustrated graphically. (B) RPA for messages containing the L1 open reading frame. The bands of expected sizes corresponding to spliced or unspliced L1 messages are indicated. Undigested probe is shown in lane P, size markers are shown in lane M, and a control hybridization with tRNA is shown in lane t. The probe which spans the splice acceptor site at nucleotide 5552 and the amino terminus of the L1 message are diagrammed. Levels of spliced L1 message were quantitated by PhosphorImager analysis and are illustrated graphically.

Northern blot analysis.

Total RNA was isolated from monolayer cells or cells harvested from methylcellulose with TriZOL reagent (Gibco BRL) as described by the manufacturer. For Northern analysis, 10 μg of total RNA was electrophoretically separated on a 0.8% agarose–2.2 M formaldehyde gel and transferred to MSI paper (Micron Separations Inc., Westborough, Mass.). Equal loading (less than 10% variability) was verified by densitometric analysis of the rRNA bands on the ethidium bromide-stained gel prior to transfer. Probes were made from a plasmid containing the E4 and E5 open reading frames (18) or the cellular involucrin gene, using a High Prime random labeling kit (Boehringer Mannheim). Hybridization and high-stringency washing were performed as described previously (18). Results were quantitated by using a Molecular Dynamics PhosphorImager:SI with ImageQuant software.

RPA.

Total RNA was prepared as described above. Antisense riboprobes were prepared by using the Riboprobe Combination System-SP6/T7 (Promega) according to the manufacturer’s instructions. SP6 was used to synthesize the p742 probe from gel-purified pRP-p742 linearized with SpeI, and T7 was used to synthesize the L1 probe from gel-purified pRPA31L1 linearized with SalI. The RNase protection assay (RPA) was performed with 15 μg of total RNA or 15 μg of yeast tRNA as a negative control as previously described (20). Results were quantitated by using a Molecular Dynamics PhosphorImager:SI with ImageQuant software.

Southern blot analysis.

Total genomic DNA was prepared by suspending the cell pellet in lysis buffer (400 mM NaCl, 10 mM Tris-HCl [pH 7.4], 10 mM EDTA) and digestion with RNase A (50 μg/ml) for 1 h at 37°C followed by incubation with proteinase K (50 μg/ml) and 0.2% sodium dodecyl sulfate (SDS) at 37°C overnight. Then 10 μg of DNA was digested with either a noncutting enzyme for the HPV genome (BamHI) or a single-cutting enzyme for the HPV genome (EcoRV). Digested DNA was separated on a 0.8% agarose gel, agitated in 0.25 N HCl for 15 min, and transferred to GeneScreen Plus nylon membrane, using the alkaline transfer method as instructed by the manufacturer (NEN Life Sciences, Boston, Mass.). The HPV-31 probe was synthesized with 25 ng of HPV genome released from pBR322-HPV31 by EcoRI digestion followed by gel purification. The HPV-31 fragment was random prime labeled by using a High Prime kit (Boehringer Mannheim). The membrane was prehybridized in 4× SSPE (20× SSPE [pH 7.4] is 3 M NaCl, 0.2 M NaH₂PO₄-H₂O, and 0.02 M EDTA-Na₂)–4× Denhardt’s solution–10% dextran sulfate–50% formamide–1% SDS–100 μg of denatured salmon sperm DNA per ml; 10 × 10⁶ cpm of probe was added to fresh prehybridization solution after heat denaturation and hybridized overnight. The membrane was washed in 2× SSC (20× SSC is 3 M NaCl plus 0.3 M sodium citrate dihydrate)–0.1% SDS three times at room temperature for 15 min followed by two 15-min washes in 0.1× SSC–0.1% SDS at room temperature and one 30-min wash in 0.1 SSC–1% SDS at 50°C. Results were quantitated by using a Molecular Dynamics PhosphorImager:SI with ImageQuant software.

Western blot analysis and immunofluorescence.

For E1^∧E4 Western analysis, urea-solubilized whole-cell extracts were prepared, separated by SDS-polyacrylamide gel electrophoresis, transferred to an Immobilon-P membrane (Millipore, Bedford, Mass.), and probed with anti-HPV-31bE1^∧E4 antibody as previously described (28). Immunofluorescence was performed on cells dried on specimen slides (Enzo Diagnostics, Farmingdale, N.Y.) and fixed for 5 min in 4% paraformaldehyde–PBS at room temperature followed by permeabilization with 100% methanol for 2 min at −20°C. Air-dried slides were blocked in PBS–0.5% Nonidet P-40–1% bovine serum albumin for 30 min at room temperature. The following primary antibodies were diluted in blocking solution and incubated with slides for 1 h at room temperature in a moist chamber: anti-cytokeratin 8.6 (C-7284; Sigma) at 1:100, anti-involucrin (I-9018; Sigma) at 1:100, antifilaggrin (BT-576; Biomedical Technologies, Stoughton, Mass.) at 1:50, and antitransglutaminase (BT-621; Biomedical Technologies) at 1:20. Following three PBS washes, fluorescein-conjugated anti-mouse immunoglobulin secondary antibody (Amersham Life Sciences, Arlington Heights, Ill.) was diluted 1:50 in blocking solution and incubated with slides for 1 h at room temperature. Slides were washed a final three times with PBS. For dual immunofluorescence with E1^∧E4, anti-HPV-31b E1^∧E4 antibody diluted 1:100 in blocking solution was incubated for 1 h following the final PBS washes. Following three PBS washes, a Texas red-conjugated anti-rabbit immunoglobulin secondary antibody (Amersham Life Sciences) was diluted 1:200 in blocking solution and incubated with slides for 1 h. After the final secondary antibody incubation, slides were washed gently in PBS three times and coverslips were mounted in 90% glycerol–10% PBS–50 μg of n-propyl gallate per ml. To quantitate induction of differentiation markers, 250 to 300 cells were counted in three randomly chosen fields. To visualize dual immunofluorescence, the fluorescein and Texas red images of the same fields were overlaid by using Adobe Photoshop software.

FISH and immunofluorescence.

Fluorescence in situ hybridization (FISH) for HPV DNA was performed as instructed for the BioPap HPV In Situ Typing Assay (Enzo Diagnostics) kit. Cells were dried on specimen slides (Enzo Diagnostics) and fixed for 5 min in 100% acetone at −20°C. HPV 31/33/51 Probe Mix was denatured on the slide for 5 min on a 95°C heat block, hybridized for 2 h on a 37°C heat block, and washed according to the manufacturer’s instructions. Hybridized probe was visualized with the Simply Sensitive In Situ Detection Fluorescent Streptavidin system (Enzo Diagnostics) according to the manufacturer’s instructions. For FISH followed by immunofluorescence, the denaturation temperature was reduced to 70°C and immunofluorescence for E1^∧E4 was performed after completion of FISH.

RESULTS

Keratinocytes containing HPV-31 express transglutaminase and involucrin but fail to express keratin-10 (K10) or filaggrin upon suspension in methylcellulose.

We first examined the ability of cell lines containing episomal copies of HPV-31 to induce markers of epithelial differentiation following growth in semisolid media containing methylcellulose. Keratinocytes containing episomal copies of HPV-31 DNA (LKP-31-1) (11) were generated by transfection of cloned viral DNA and compared with NHKs following suspension in medium containing 1.6% methylcellulose. The first markers examined were involucrin and transglutaminase, which are first expressed in cells of the spinous layer, the layer just above the basal cells in an epithelium (9). While roughly 25% of monolayer LKP-31-1 or NHK cells stained positively for involucrin and transglutaminase (Fig. 1A and D and data not shown), nearly 100% of LKP-31-1 cells and NHKs stained positively for these two markers of differentiation after suspension in methylcellulose for 24 h (Fig. 1B, C, E, and F).

FIG. 1.

Suspension of LKP-31-1 cells in methylcellulose induces expression of the differentiation markers involucrin and transglutaminase but not K10 or filaggrin. LKP-31 cells and NHKs were isolated from monolayer cultures or after suspension in 1.6% methylcellulose (mcell.) for 24 h and examined by immunofluorescence. (A, D, G, and J) Representative fields of LKP-31-1 monolayer cells; NHK monolayer cells were stained similarly. (B, E, H, and K) NHKs in methylcellulose. (C, F, I, and L) LKP-31-1 cells in methylcellulose. (A to C) Stained with an antibody to involucrin; (D to F) stained with an antibody to transglutaminase; (G to I) stained with an antibody to K10; (J to L) stained with an antibody to filaggrin. Magnifications: (A to I) ×200; (J to L) ×400. All antibodies were followed by a fluorescein-conjugated anti-mouse secondary antibody.

We next analyzed the expression of K10 and filaggrin following suspension in methylcellulose. K10 is a differentiation-specific keratin which is expressed beginning in the spinous layer (13), while filaggrin starts to be expressed in the granular layer just above the spinous layer (9). Approximately 10% of cells in monolayer LKP-31-1 cultures were K10 positive (Fig. 1G). Following culture in methylcellulose, the number increased slightly to 12 to 20% in either NHKs or LKP-31-1 cells (Fig. 1H and I), indicating that suspension in semisolid medium does not induce a high level of expression of the differentiation-specific keratin K10 in either cell type. In contrast, filaggrin was not expressed in monolayer cultures but was induced by roughly 70% of NHK cells after 24 h in methylcellulose (Fig. 1K). Interestingly, only 9% of LKP-31-1 cells harvested from methylcellulose at this same time point induced filaggrin expression (Fig. 1L). We conclude that while the induction of involucrin and transglutaminase in LKP-31-1 cells is similar to that in NHKs, the expression of the granular layer marker filaggrin is reduced in the LKP-31-1 cells at this time point. These results were confirmed with a second cell line containing episomal HPV-31 DNA (LKP-31-2). Consistent with studies in organotypic rafts in the absence of activation of protein kinase C (PKC) (25), the cervical biopsy-derived line CIN-612 which contains episomal copies of HPV-31b did not induce filaggrin expression following culture in methylcellulose (data not shown). When cells were examined after 42 h in methylcellulose, the percentage which induced filaggrin increased in the LKP-31-1 and LKP-31-2 lines (data not shown).

The induction of the HPV-31 late promoter p742, transcription of L1 message, and onset of keratinocyte differentiation coincide.

We next investigated whether the degree of differentiation achieved by LKP-31-1 cells after suspension in methylcellulose was sufficient for the induction of viral late functions. LKP-31 cells have previously been shown to induce late viral functions as well as virion biosynthesis following growth in organotypic rafts (11). We first determined if the viral late promoter p742 was induced in LKP-31-1 cells following culture in methylcellulose. As diagrammed in Fig. 2A, the major transcripts expressed in undifferentiated keratinocytes initiate from the p97 promoter, encode the open reading frames E6 (or the splice variant E6*), E7, E1∧E4, and E5, and terminate at the early polyadenylation site (18). Following differentiation, the most abundant late transcript initiates at p742, contains the open reading frames E1∧E4 and E5, and terminates at the early polyadenylation site (Fig. 2A) (18). We performed Northern analysis on RNA from LKP-31-1 cells cultured for various times in methylcellulose, using a probe from the E4/E5 region which allows for the simultaneous identification of the predominant early and late messages (Fig. 2B). While the expression of the p97 transcripts containing the E4/E5 region does not change dramatically upon differentiation, expression of messages from p742 encoding the E4/E5 region increased significantly by 16 h, with peak induction at 24 h (Fig. 2B).

FIG. 2.

Coincident induction of involucrin and HPV late messages after 12 to 16 h in methylcellulose. Northern analysis was performed on total RNA isolated from LKP-31-1 cells cultured in methylcellulose (mcell.) for the indicated times. (A) The most abundant HPV messages initiated at the p97 (early) and p742 (late) promoters as well as the message encoding L1 are diagrammed. The upper two messages are the most abundant early transcripts which are initiated at p97. The third message shown is the most abundant p742 message, and the fourth is the p742-initiated message which contains the L1 open reading frame. (B) Northern blot for the major message initiated at p97 (E6 or E6∗,E7,E1^∧E4,E5) and p742 (E1^∧E4,E5), using a probe of the HPV E4 and E5 open reading frames. Quantitation of the p742 message (E1^∧E4,E5) by PhosphorImager is shown as a graph below. (C) Northern blot for cellular involucrin message. Quantitation of the involucrin message by PhosphorImager is shown as a graph below.

Since involucrin transcription increases during epithelial differentiation, we next examined how the induction of this cellular differentiation marker corresponded to that of the p742 promoter. As shown in Fig. 2C, involucrin message levels increased by 12 h in methylcellulose, slightly preceding induction of p742 message, but exhibited a time course similar to that seen with p742 induction. Additional experiments demonstrated that the levels of both p742 message and involucrin message peaked at 24 h and then decreased between the 24- and 48-h time points (data not shown). The coordinate induction of involucrin and p742 messages was seen in the second independently derived LKP-31-2 line and confirmed by the observation that the appearance of cells positive for involucrin and E1^∧E4 proteins is coincident (data not shown).

Previous studies by our laboratory demonstrated that the induction of the late promoter p742 was not sufficient for the appearance of L1 messages and that additional posttranscriptional changes in splicing and polyadenylation site usage were required (19, 25). These posttranscriptional changes correlated with a more complete program of epithelial differentiation evidenced by filaggrin expression. We next analyzed the induction of p742 and the appearance of L1 message by RPA performed with total RNA isolated from LKP-31-1 cells cultured for different times in methylcellulose. Using a probe spanning the late promoter p742, we confirmed the results from the Northern analysis, showing induction of p742 beginning by 16 h (Fig. 3A). Using a probe spanning the L1 splice junction, we observed that the appearance of messages encoding L1 paralleled the induction of p742 (Fig. 3B). This finding indicates that culture of HPV-containing cells in methylcellulose for a short period of time induces a sufficient degree of differentiation to activate expression of the late promoter as well as provides for the posttranscriptional modifications required for L1 transcription. In our studies, these activities occurred at similar times. The induction of the late promoter p742 and transcription of L1 messages following suspension in methylcellulose have been confirmed in assays using five independently isolated LKP-31 lines established from different transfections.

It was next important to determine if activation of late transcription resulted in the appearance of late proteins, as this process may also be regulated by differentiation. For these studies, we performed Western analysis on protein extracts from cells cultured in methylcellulose. A low level of E1^∧E4 protein was detected in monolayer cultures, as has previously been reported (28), and this increased significantly following suspension in methylcellulose (Fig. 4). Using both Western blot analysis and immunofluorescence, we could not convincingly demonstrate the synthesis of capsid protein after suspension in methylcellulose despite the presence of L1 message (data not shown). It is not clear if this is due to low levels of synthesis, or if the level of differentiation provided by the methylcellulose system is not sufficient for the differentiation-dependent translation of L1 protein.

FIG. 4.

Induction of the late protein E1^∧E4 upon suspension in methylcellulose. LKP-31-1 cells were recovered from suspension in methylcellulose (mcell.) at the indicated times. Equal amounts of urea-solubilized whole-cell extracts were separated by SDS-PAGE and examined by Western blot analysis using an antibody to the HPV-31 E1^∧E4 protein. Whole-cell extract from NHKs containing an HPV-18 E7-expressing retrovirus is included as a negative (neg.) control. The induction of E1^∧E4 protein was found to be roughly threefold when quantitated by densitometer.

Amplification of viral DNA upon differentiation in methylcellulose.

In previous studies, we demonstrated that high level expression from the late HPV-31 promoter required that the viral DNA be present as episomes (11). This finding suggested a link between late promoter induction and HPV DNA amplification. Therefore, we examined whether viral copy number also increased upon suspension in methylcellulose. Total genomic DNA from LKP-31-1 cells cultured in methylcellulose was digested with an enzyme which does not cut the HPV genome, and Southern analysis was performed. Figure 5A demonstrates that viral DNA is maintained episomally following suspension in methylcellulose, as indicated by the presence of closed circular DNA. The amount of HPV DNA increased approximately 3.5-fold upon culture in methylcellulose for 40 h. The ability to detect HPV DNA amplification by Southern analysis was surprising, as previous attempts to detect amplification in organotypic raft cultures by this method were unsuccessful (22). We next examined viral DNA in cells harvested from methylcellulose as a function of time to determine if DNA amplification coincides with late promoter induction. Total genomic DNA from LKP-31-1 cells was digested with an enzyme which cuts the HPV genome once and examined by Southern analysis (Fig. 5B). HPV copy number was found to increase by 16 h in the LKP-31-1 cells and reached a maximum after 24 h. This pattern of induction parallels that seen for p742 message and capsid protein message in these LKP-31-1 cells (Fig. 3 and 5). To ensure that this effect was not specific to LKP-31 cells, we also analyzed the amount of HPV DNA in CIN-612 cells isolated from an HPV-31b cervical lesion at the same time points following methylcellulose culture (Fig. 5B). Viral DNA was again found to increase by 16 h but continued to accumulate after 24 h (Fig. 5B).

FIG. 5.

Southern analysis demonstrating amplification of HPV DNA upon culture in methylcellulose (mcell.). (A) LKP-31-1 cells were harvested from monolayer culture (0 h in methylcellulose) or after 40 h in methylcellulose. Equal amounts of total genomic DNA were digested with a restriction enzyme which does not cut the HPV-31 genome, and Southern analysis was performed with a probe of the entire HPV genome. An increase in viral DNA maintained episomally is marked with the arrow labeled ccc. Controls roughly equivalent to 5 or 50 copies of viral DNA per cell are also shown. (B) LKP-31-1 or CIN-612 cells were harvested after suspension in methylcellulose for the indicated times. Equal amounts of total genomic DNA were digested with a restriction enzyme which cuts the HPV-31 genome once, and Southern analysis was performed with a probe of the entire HPV genome. Linearized viral DNA is marked with an arrow. Fold induction in HPV DNA was quantitated by PhosphorImaging and is illustrated graphically below each lane. Controls roughly equivalent to 5 or 50 copies of viral DNA per cell are also shown.

Induction of E1^∧E4 protein and DNA amplification occur in the same cells.

Since late promoter induction and viral DNA amplification occurred at similar times of culture in methylcellulose, we next examined whether these two processes were occurring in the same cells. For these studies, we used FISH to examine viral DNA copy number followed by immunofluorescence for the late protein E1^∧E4. By using fluorescein detection reagents for the FISH analysis and Texas red detection reagents for the E1^∧E4 immunofluorescence, we were able to determine if amplification and late protein expression occurred in the same cells. A small number of cells positive by FISH for HPV DNA were evident in monolayer LKP-31-1 cells (Fig. 6A), but the number of cells containing FISH reactivity as well as the amount of hybridization in individual cells increased after 24 h of growth in methylcellulose (Fig. 6C), in agreement with the Southern analysis described above. Only a very rare monolayer cell contained FISH reactivity comparable to that seen in the methylcellulose cells (Fig. 6G). The FISH assay used here is likely not sensitive enough to detect the low copy number of HPV DNA in most monolayer cells and identifies only those cells which have substantially increased viral copy number.

FIG. 6.

Amplification of viral DNA and induction of the late protein E1^∧E4 occur in the same subset of cells. Monolayer LKP-31-1 cells or LKP-31-1 cells after 24 h in methylcellulose were fixed on slides and analyzed by FISH for HPV DNA followed by immunofluorescence for E1^∧E4 protein. FISH analysis using a fluorescein detection system is shown in panels A, C, E, and G. Immunofluorescence of the same fields of cells for E1^∧E4 protein using a Texas red-conjugated secondary antibody is shown in panels B, D, F, and H. A representative monolayer field is shown in panels A and B. Representative fields of cells after 24 h in methylcellulose are shown at magnifications of ×156 in panels C and D and ×312 in panels E and F. A field containing a rare monolayer cell positive for FISH and E1^∧E4 is shown in panels G and H at a magnification of ×312.

An examination of LKP-31-1 cells by immunofluorescence for E1^∧E4 protein demonstrated that only a rare brightly positive cell was present in monolayer cultures (Fig. 6B and H). Since the Western analysis demonstrated that E4 protein was detectable in monolayer cultures (Fig. 4), we suspect that some E4 was expressed in all monolayer cells which was not evident by this immunofluorescence technique. After suspension in methylcellulose, roughly 25% of LKP-31-1 cells induced a large amount E1^∧E4 protein (Fig. 6D and F). When FISH analysis was coupled with E1^∧E4 immunofluorescence, we found that the LKP-31-1 cells which induce E4 protein expression were the same cells which contained amplified HPV DNA (Fig. 6C to F). Most cells which were positive for DNA amplification in methylcellulose also expressed the E1^∧E4 protein, and all E1^∧E4-positive cells had FISH reactivity. Moreover, induction of E4 protein and amplification of viral DNA occurred coincidentally, with the appearance of both after suspension for 16 h in methylcellulose (data not shown). Interestingly, the rare brightly positive E1^∧E4 cell in the LKP-31-1 monolayer was also found to contain a large amount of viral DNA (Fig. 6G and H). This finding suggests that significant late protein induction occurs only in cells undergoing viral DNA amplification. Taken together, these observations support the hypothesis that viral replication and induction of the late promoter are processes which are coupled and dependent on the differentiation of the host cell. These studies were repeated with CIN-612 cells, a clonal line derived from a cervical biopsy, and similar results were obtained. It is interesting that despite our use of cells which all contain HPV-31 episomes, only a subset amplified viral DNA and induced E1^∧E4 expression following differentiation.

Induction of the late protein E1^∧E4 is compatible with involucrin and transglutaminase expression but does not require K10 or filaggrin induction.

Since induction of E1^∧E4 synthesis and DNA amplification occurred only in a subset of cells in methylcellulose, we next examined whether these cells were the same ones which expressed the differentiation markers K10 and filaggrin. Our initial studies indicated that, like E1^∧E4, these markers were also expressed in only a subset of cells. We first investigated if cells which induced E1^∧E4 protein were also positive for involucrin, transglutaminase, and K10. While the expression of all three of these markers begins in the spinous layer of a differentiated epithelium, they are regulated by different mechanisms. As seen in Fig. 1, nearly all LKP-31-1 cells and NHKs expressed the suprabasal markers involucrin and transglutaminase after suspension for 24 h in methylcellulose. At this time point there is significant E1^∧E4 synthesis, and the cells which induced E1^∧E4 expression also expressed both involucrin and transglutaminase (Fig. 7A to D). Since many cells were positive for involucrin and transglutaminase but failed to induce E1^∧E4 protein, expression of these markers of differentiation is clearly not sufficient for the induction of viral late functions. We next examined the LKP-31-1 cells by immunofluorescence for the expression of the suprabasal keratin K10, which we observed to be expressed in only a small number of cells after differentiation in methylcellulose (Fig. 1). The vast majority of E1^∧E4-positive cells were not K10 positive, although a rare double-positive cell could be found (Fig. 7E and F). This finding indicates that the differentiation signals required for K10 synthesis in LKP-31-1 cells are not well induced by suspension in methylcellulose. More importantly, these signals are not required for the induction of the late promoter and synthesis of E1^∧E4 protein. We next examined whether cells which induced E1^∧E4 protein were filaggrin positive. Filaggrin is a marker of differentiation which appears in the granular layer, and in previous studies in organotypic raft cultures, its synthesis correlated well with expression of L1/L2 message and protein (25). As shown in Fig. 1, filaggrin was induced in only 9% of LKP-31-1 cells after 24 h in methylcellulose. We found that the small number of cells which induced filaggrin expression after growth in methylcellulose were not E1^∧E4 positive (Fig. 7G and H). This result indicates that the differentiation-dependent signals necessary for filaggrin synthesis are not required for induction of the late protein E1^∧E4.

FIG. 7.

Induction of the late protein E1^∧E4 occurs in cells positive for involucrin and transglutaminase but does not require K10 or filaggrin induction. Dual immunofluorescence for E1^∧E4 protein and cellular differentiation markers was performed on LKP-31-1 cells after 24 h in methylcellulose. (A, C, E, and G) E1^∧E4 immunofluorescence, using a Texas red-conjugated secondary antibody; (B, D, F, and H) immunofluorescence for cellular differentiation markers, using a fluorescein-conjugated secondary antibody which has been overlaid onto the E1^∧E4 staining of the same field. Cells which are positive for the differentiation marker, but not for E1^∧E4, are green. Cells which are positive for both the differentiation marker and E1^∧E4 are yellow-orange. Cells which are positive for E1^∧E4, but not for the differentiation marker, remain red. Cells were stained for involucrin (B), transglutaminase (D), K10 (F), and filaggrin (H).

DISCUSSION

The close association of the HPV life cycle and the differentiation state of its host cell is demonstrated by the restriction of late gene transcription and amplification of viral DNA to suprabasal epithelial cells. In this study, we have used a simple method of culturing keratinocytes to induce differentiation-dependent late viral functions. Previous studies of differentiation-dependent HPV late functions have used culture in organotypic rafts, which is technically challenging and requires extended periods of time for growth (4, 5, 19, 27). Suspension of cell lines which maintain HPV-31 episomes in methylcellulose results in the rapid induction of the late promoter p742, appearance of L1 transcripts, synthesis of the late protein E1^∧E4, and differentiation-dependent amplification of viral DNA. There are several advantages of the methylcellulose system for the analysis of HPV late functions. While raft cultures provide a spatial separation of cells at various stages of differentiation, separate layers are not easily isolated. The methylcellulose system allows for the analysis of differentiation-dependent activities as a function of time and permits isolation of cells which have progressed to similar degrees of differentiation. Furthermore, these late functions were induced in only 1 day instead of the 2 weeks required for raft culture. The methylcellulose system also allowed us to quantitate the degree of differentiation-dependent viral amplification by Southern analysis. In previous studies, we were able to use only the more qualitative method of in situ hybridization of raft cultures to detect DNA amplification. Use of these methods should allow for a quick and easy assay to study the mechanisms which regulate these differentiation-dependent viral functions.

We have used this system to provide support for the hypothesis that replication of viral DNA and induction of transcription from the late promoter are interdependent processes which are triggered by cellular differentiation (11). In agreement with studies by Doorbar et al., we found that viral DNA amplification and late promoter induction, assayed by expression of the E1^∧E4 late protein, occur in the same cells (8). Furthermore, our analysis of these functions over time demonstrates that p742 induction, L1 transcription, and viral amplification are coincident with one another as well as with the onset of cellular differentiation. It is possible that there are small differences in the time of induction of these processes, but we have not been able to identify those by our methods. These data are consistent with the idea that an increase in HPV DNA copy number is required for the induction of high levels of late protein. This could either reflect a requirement for a change in chromatin configuration around the late promoter brought on by replication or indicate that an increase in the number of templates expressing low levels of the late promoter transcripts results in high levels of late transcription. In further support of this hypothesis, we found that although only a rare cell in monolayers expressed a large amount of E1^∧E4 protein, this cell always contained dramatically amplified amounts of viral DNA. It remains possible that the coincidence of these late processes is a result of their responding to the same signals, but our failure to detect late gene expression from integrated templates makes this unlikely (11).

The differentiation requirements for induction of the late promoter, L1 transcription, and viral DNA amplification were surprisingly few. While suspension in methylcellulose did not provide the signals necessary for high levels of K10 induction, these signals are clearly not required for induction of E1^∧E4 expression or DNA amplification. The early suprabasal markers involucrin and transglutaminase were found to be well induced in this system and compatible with late promoter induction, since cells which induced E1^∧E4 were positive for both of these markers. Since we observed cells which were positive for involucrin and transglutaminase but did not express E1^∧E4, expression of these markers is not sufficient for activation of the late promoter. These studies demonstrate that detachment from the basement membrane for only 16 h provides signals sufficient for all of the HPV late functions discussed above.

Interestingly, induction of filaggrin, a marker of the granular layer, appeared to be reduced in LKP-31 cells compared to NHKs after 24 h in methylcellulose. This result is in agreement with studies which demonstrated a reduced commitment to terminal differentiation by squamous cell carcinoma lines compared to normal keratinocytes upon suspension in semisolid medium (29). This observation also highlights the fact that suspension in methylcellulose does not induce faithful terminal differentiation of the HPV-31-containing keratinocytes. In previous studies using the biopsy-derived cell line CIN-612, we observed the synthesis of HPV-31b virions following growth in raft cultures. This synthesis was dependent upon the addition of activators of PKC and correlated with the synthesis of filaggrin in suprabasal layers (25). Further studies demonstrated that addition of PKC activators alleviated a block to late gene transcription through changes in splicing and polyadenylation (19). In our current studies, filaggrin expression was induced following growth in methylcellulose but not in the HPV-31-positive cells which expressed E1^∧E4 and amplified viral DNA. One explanation for this could be that activation of the late HPV-31 promoter requires commitment to S phase while filaggrin expression occurs in G₀ or early G₁. In raft cultures of CIN-612 cells, we could not distinguish between cells expressing filaggrin and those expressing E1^∧E4, and it is possible that this occurred in separate populations of cells. A report demonstrating that E1^∧E4-positive cells in the granular layer of a low-grade HPV-16 lesion failed to express filaggrin supports this hypothesis (8). Finally, we were not able to detect the synthesis of significant levels of L1 protein in methylcellulose, indicating that more complete or terminal differentiation may be required for this process. Studies are in progress to determine if the addition of activators of PKC to cells in methylcellulose induces synthesis of significant amounts of L1 protein and whether this is coincident with filaggrin expression in these same cells.

Our data demonstrate that while a significant proportion of LKP-31 or CIN-612 cells induce late functions after only 16 h of suspension in methylcellulose, the majority of cells, roughly 75%, still remain E1^∧E4 negative and do not contain amplified HPV DNA. This observation could explain why we saw only a 3.5-fold increase in total viral DNA by Southern analysis. If the entire population of cells were to amplify DNA, we would expect a more dramatic induction of viral DNA by Southern analysis. Interestingly, after suspension in methylcellulose for periods longer than 24 h, we did not observe an increase in the number of cells expressing E1^∧E4 or amplifying DNA but rather observed that the intensity of the signals in the positive cells increased. This was surprising given that all cells in the culture were exposed to the same signals following suspension in methylcellulose. It is interesting that in organotypic raft cultures, a similarly sized subset of HPV-31-positive cells induced E1^∧E4 synthesis and amplification of viral DNA (11). Several explanations for this observation are possible. Since amplification of HPV DNA requires the DNA replication machinery of the host cell, these cells must be in S phase or have the capacity to express S-phase-specific genes. Only a subset of cells may be competent to express the cellular proteins required for DNA synthesis, and these may be the cells which express E1^∧E4 and amplify DNA. Alternatively, it is possible that the cell cycle distribution of the cells when they were placed in methylcellulose determines whether they are capable of reentering S phase upon differentiation. A final possibility is that cells in methylcellulose differentiate coordinately only through the early stages and that only a subset continue on to terminal differentiation. Since our data show that those cells which express the granular layer marker filaggrin do not induce E1^∧E4, it is also possible that terminal differentiation and induction of HPV late functions are incompatible. All of these possibilities can now be addressed by using this system.

In conclusion, we have described the use of a simple system of keratinocyte differentiation for the study of the HPV life cycle. Our studies support the hypothesis that viral DNA amplification and late gene expression are interdependent processes. While the late functions studied here clearly required differentiation and were compatible with the induction of the early markers involucrin and transglutaminase, further differentiation was not required and even absent in these permissive cells. Further use of this system will allow us to address both the differentiation and cell cycle signals required for HPV replication and late functions.