Development of a Novel System To Study Hepatitis Delta Virus Genome Replication

Plasmid.

Beginning with a sequenced HDV cDNA (15), the ScaI-XbaI fragment containing the δAg open reading frame was inserted into pcDNA5/FRT/TO (Invitrogen). This vector has a hybrid human cytomegalovirus/TetO₂ promoter for tetracycline-regulated gene expression. It also has an Flp recombination target (FRT) site for Flp recombinase-mediated integration.

Generation of 293-δAg, a cell line expressing the small form of δAg under TET-on control.

Human embryonic kidney cells, Flp-In T-Rex (Invitrogen), were used to generate a δAg expression cell line. The genome of this cell line has been stably transfected with two plasmids, one containing the FRT site and the other expressing a TET repressor. Transfection of this cell line was performed with Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. The plasmid containing the HDV sequence was cotransfected with pOG44, which expresses Flp recombinase. This cotransfection led to the integration of HDV sequence through the FRT site with its expression under TET-on control. Two days after transfection, cells were reseeded at less than 25% confluence. Two to 3 h later, selection antibiotics were added to the medium. Selection antibiotics, hygromycin and blasticidin (Invitrogen), were used to maintain both the δAg and TET repressor genes. Two weeks later, separate foci appeared, and the pool of such cells was expanded to generate what we refer to as 293-δAg cells. After TET induction, immunoblot and immunocytochemistry were used to confirm the expression of δAg in 100% of these cells. 293-δAg cells were maintained in Dulbecco's modified Eagle's medium with high glucose and 10% fetal bovine serum along with the selection antibiotics mentioned above.

Transfection of 293-δAg cells with HDV RNA sequences to initiate HDV genome replication.

To initiate HDV replication, these 293-δAg cells under Tet⁻ conditions were transfected with a greater-than-full-length HDV RNA species. This RNA had a 2-nucleotide deletion at the unique EcoRI site, a change that interferes with the translation of a functional δAg (5). Using strategies described in the legend of Fig. 3, we consider that this transfection led to the replication and accumulation of HDV RNA in about 15% of cells. As one approach to achieve replication in more of the cells, total RNA was extracted from some of the cells and used to retransfect the culture. This cycle of retransfection was repeated as many as four times. The fraction of transfected cells was thus raised to ≈40%. As an additional approach, we carried out cell cloning under Tet⁻ conditions. We thus obtained clones, designated 293-HDV, with HDV replication occurring in all cells.

Quantitation of δAg and genomic RNA.

Tri Reagent (Molecular Research Center) was used to extract both protein and RNA. δAg was measured in protein samples by immunoblot followed by ¹²⁵I-staph A protein (Perkin-Elmer) and detection by Bio-imager (Fuji). Quantitation was measured relative to a standard of purified recombinant δAg provided by Harmon Zuccola and James Hogle (Harvard University), as previously described (13). Genomic HDV RNA was detected by Northern analysis using a ³²P-labeled specific RNA probe and quantitated relative to a standard of in vitro-transcribed HDV genomic RNA, as previously described (13). Detection and quantitation of radioactivity were done by using a Bio-imager (Fuji).

Immunocytochemistry.

The procedure was based on a method described previously (1, 14). As primary antibodies, for δAg, we used a 1:1,000 dilution of a rabbit polyclonal antibody; for SC35, we used a 1:2,000 dilution of a mouse monoclonal antibody (Abcam); and for fibrillarin, we used a 1:500 dilution of a mouse monoclonal antibody (Abcam). As secondary antibodies, we used a 1:500 dilution of Cy2-conjugated AffiniPure donkey anti-rabbit immunoglobulin G (heavy plus light chains) and rhodamine red-X-conjugated AffiniPure donkey anti-mouse immunoglobulin G (heavy plus light chains) (Jackson ImmunoResearch). Cellular DNA was stained with DAPI (4′,6′-diamidino-2-phenylindole) (Sigma). Samples were analyzed using an Inverted Nikon TE2000-U microscope with 40× Plan ELWD, WD 3.7-2.7 objective, and specific filter blocks. Images were acquired with a Cascade 650 monochrome camera (Photometrics) and MetaVue software (Universal Imaging). Subsequently, images were processed with Canvas 9.0 and Adobe Photoshop 7.0 software.

Cell cycle and apoptosis analysis.

For 293-δAg and 293-HDV cells, both with and without TET induction, we used propidium iodide staining and analytical cell sorting with a Guava PCA apparatus (Guava Technologies) to determine the distribution of cells within the cell cycle. Data were analyzed using Flowjo software. Using the same apparatus together with a Nexin kit to stain for annexin V and 7-amino-actinomycin D (7-AAD) (Guava Technologies), we also assayed for various stages of cell death. Trypan blue was used as recommended by the supplier (Gibco).

Conditional expression of δAg in 293-δAg cells.

The first step in our strategy was to create a cell line conditionally expressing the small δAg. We used a commercially available engineered form of 293 cells, human embryonic kidney cells. These cells contain a single target site for the recombinase flp, into which vector sequences can be efficiently inserted. Furthermore, these cells express a protein that in the absence of TET will repress transcription from the transposed sequences. In this way, we transferred cDNA for the small δAg to create what we refer to as 293-δAg cells.

To test whether all cells were expressing δAg, we used immunocytochemistry with a δAg-specific rabbit polyclonal antibody. As shown in Fig. 1A, 2 days after TET was added at 1 μg/ml, all cells expressed δAg, with staining located predominantly in the nucleolus. This distribution was as expected from previous transient transfection studies (1).

Next, we used immunoblots to detect and quantitate δAg. As shown in Fig. 1B, in the absence of TET, there was a low amount of expression, about 40,000 copies per cell. However, after TET addition, the amount increased 60-fold, to 2.4 million copies. This high number approximates the abundance of host cell ribosomal components and many required translation factors (8).

As further characterization of these cells, Fig. 1C shows that 0.1 to 1 μg/ml of TET was sufficient to give maximal induction of δAg expression. In addition, Fig. 1D shows a time course of the induction for cells treated with 1 μg/ml of TET. Maximal levels were achieved within 2 days.

Initiation of HDV genome replication in 293-δAg cells.

The second step was to transfect 293-δAg cells with HDV RNA sequences. We specifically chose an HDV RNA with a 2-nucleotide deletion in a region that disrupts expression of δAg (4). That is, we wanted the integrated DNA copy of δAg cDNA to be the only and unchangeable source of this protein. At various times after transfection, we analyzed the RNA in these cells for replication and accumulation of HDV genomic RNA.

Consider first the situation, when these cells were maintained in the absence of TET. As shown in Fig. 2, the transfected cells achieved a low level of HDV genomic RNA, about 1,000 copies per average cell. As shown, this level was essentially stable for at least 1 year.

We next assessed HDV replication when TET was added to the medium. Figure 2 shows examples of when TET was added at 8 and 52 weeks after the initial RNA transfection. In both cases, a burst of HDV RNA accumulation was detected within 3 days. During the following weeks, we then observed significant detachment of cells from the monolayer. Furthermore, for the remaining cells, the amount of HDV RNA per average cell decreased dramatically so that by 2 to 3 weeks, no HDV could be detected. In contrast, for 293-δAg cells that had not been transfected with HDV RNA, we did not detect detachment of cells during 3 weeks of TET induction. Our preliminary interpretation of these data was that the initial HDV RNA transfection efficiency was less than 100%. Furthermore, it was considered that after TET addition, only those transfected cells that were replicating the HDV RNA were the ones that ultimately became detached from the monolayer. As now explained, this interpretation was supported by data obtained using immunocytochemistry.

As shown in Fig. 1A, we already knew that the 293-δAg cells all expressed δAg and that it was primarily localized to the nucleolus. However, when such cells transfected with the HDV RNA were subsequently induced with TET for 2 days, there were two different patterns of intranuclear distributions, as shown in Fig. 3A. More than half of the cells had the nucleolar distribution, while about 40% had a dramatic exclusion from the nucleolus together with a nucleoplasmic localization. This nucleoplasmic pattern included but was not limited to speckles, some of which colocalized with those that contain the splicing factor SC35 (16). This distribution is very much like that reported for replication of the HDV genome in Huh7 cells, where the staining was nucleoplasmic and included SC35 speckles (1). As shown in Fig. 3B, when we examined cells at 2 weeks after TET addition, the only cells staining for δAg had a nuclear pattern that was predominantly nucleolar. These data further support the interpretation that only a fraction of the cells were initially transfected with HDV RNA and that it was these cells that were being detached during the extensive TET induction.

To further establish these interpretations, we subcloned cells that had been RNA transfected and maintained for a long period in the absence of TET. Seventeen such clones were generated and tested by Northern analysis for HDV replication. Only six clones gave extensive HDV RNA accumulation when induced for 2 days with TET (data not shown). The other clones were negative for HDV RNA. Furthermore, when the induced clones were examined by immunocytochemistry, it was only with the HDV RNA-positive clones that we detected a pattern of δAg in the nucleoplasm and excluded from the nucleolus. Data for such a clone, designated 293-HDV, are shown in Fig. 3C. Note that all the cells had this nucleoplasmic distribution of δAg after a brief TET induction. (Without the induction, the distribution was the same, except that the signal was less intense [data not shown].)

An additional immunostaining experiment was used to compare the original 293-δAg cells with the clone 293-HDV interpreted as replicating the HDV. As shown in Fig. 4A, at 1 day after TET induction, all the nuclear δAg staining colocalized with that for fibrillarin, a marker for the nucleoli. In contrast, for 293-HDV cells, the δAg staining was nucleoplasmic but excluded from the sites of fibrillarin staining (Fig. 4B).

As a further characterization, we examined two of the 293-HDV clones for the accumulation of HDV RNA after TET induction. Figure 5 shows the average HDV RNA values for these two clones as a function of the time after TET induction, with the error bars representing the range of values. In other words, both clones were essentially the same in terms of HDV RNA accumulation. The HDV RNA levels reached 4.3 × 10⁴ copies per average adherent cell, a value about twice that achieved prior to subcloning (Fig. 2). Also, the accumulation was prompt, reaching half-maximal values in less than 1 day. If we compare these data with the time course of accumulation of δAg during TET induction as presented in Fig. 1D, the results are very similar; that is, the levels of δAg and HDV RNA increase in parallel during TET induction.

Next, we examined the consequences of TET induction on the viability of 293-HDV cells. As mentioned earlier, prior to subcloning, we did observe detachment of some cells following several days of TET induction (Fig. 2 and 3). Furthermore, we interpreted the 293-HDV cells but not the 293-δAg cells as being susceptible to such detachment. We confirmed this interpretation by observation of the cloned 293-HDV cells relative to 293-δAg at day 6 of culture with or without TET induction. As shown in Fig. 6, extensive detachment occurred for the TET-induced 293-HDV cells (Fig. 6B). In contrast, the 293-δAg cells showed no detachment (Fig. 6D). There was no detachment in the absence of TET induction (Fig. 6A and C).

Finally, we used several procedures in parallel to better compare the effects of TET induction on both 293-δAg and 293-HDV cells. Cultures were studied at days 1 to 4, as well as at day 6, when TET induction led to extensive cell detachment of 293-HDV cells. We counted adherent and nonadherent cells. Trypan blue staining was used to detect dead cells. Staining for annexin V and 7-AAD followed by cell sorting was used to detect stages of apoptosis. Finally, propidium iodide staining followed by cell sorting was used to detect stages of the cell cycle.

Overall, TET induction of 293-δAg cells had no effect detectable by these assays. However, for 293-HDV cells, TET induction began the following series of events. (i) Even by day 1, we could detect a reduction in the S- and G₂/M-phase cells relative to G₁/G₀ by cell cycle analysis. (ii) By day 2, these cell cycle effects were dramatic, as shown in Fig. 7, with a fivefold reduction in S- and G₂/M-phase cells relative to G₁/G₀. There was also a significant increase in cell debris. (iii) By day 3, the cell cycle blockage continued, and in addition, we could detect a 60% decrease in the total cell number relative to that of uninduced 293-HDV cultures. Also, about 6% of the total cells were no longer adherent. (iv) By day 4, the proportion of nonadherent cells increased to about 20%. This was at least 20 times more than that for uninduced 293-HDV cultures. (v) By day 6, virtually 100% of the cells were no longer adherent.

The adherent and nonadherent cells were also examined by staining for annexin V, 7-AAD, and trypan blue. Typical data for day 4 are summarized in Table 1. The data show that for all four culture conditions, the adherent cells were not significantly different in terms of these viability markers. Similarly, the nonadherent cells were also not significantly different. Thus, by these particular viability assays, the only specific difference for the TET-induced 293-HDV culture was in the dramatic increase in the fraction of cells that were nonadherent.

In summary, from these studies, we conclude that HDV genome replication in TET-induced 293-HDV cells was specifically interfering with cell growth, cell cycling, and finally, cell adherence. It was secondarily leading to cell death, some but not necessarily all of which was via apoptotic mechanisms; that is, HDV genome replication was indirectly cytotoxic. Others have previously shown that the loss of cell adhesion can lead to the onset of apoptotic cell death (17).