Regulation of Porcine Endogenous Retrovirus Release by Porcine and Human Tetherins

Abstract

The risk of transmission of porcine endogenous retrovirus (PERV) is one of the major safety issues in xenotransplantation. Human tetherin, recently described as an antiviral protein able to inhibit the release of enveloped viruses, and its porcine homologue were shown to inhibit PERV release from producer cells, establishing themselves as candidate molecules to suppress PERV production in porcine xenografts by animal engineering.

The potential risk of pig-to-human transmission of porcine endogenous retroviruses (PERV) is a major safety issue for xenotransplantation (4, 17, 24). Genetic engineering of pigs can be applied to reduce this risk by creating pig lines unable, or with reduced ability, to produce PERV. Two groups have adopted this strategy and produced transgenic pigs expressing small interfering RNAs for the inhibition of PERV expression (5, 6, 22). Other molecules have been investigated and could potentially be employed in the development of transgenic pigs: intracellularly expressed single domain antibodies directed against PERV Gag (3), sugar-modifying enzymes to remodel PERV envelope glycoprotein (19), and the restriction factor human APOBEC3G (7, 13).

Tetherin (also named BST-2, CD317, HM1.24) has recently been described as a restriction factor in human cells able to block the release of retroviruses (14, 21), filoviruses (14, 23), and arenavirus (23). Tetherin expression varies between different cell types and can be induced by type I interferon (IFN) (2, 21). To overcome this restriction, certain lentiviruses and filovirus have developed various antitetherin countermeasures (11, 12, 15, 16, 21, 25). However, no antitetherin activity has been reported for gammaretroviruses, including murine leukemia virus (MLV) and PERV.

Here, to explore the feasibility of using tetherin to generate transgenic pigs with reduced PERV production, in addition to human tetherin, we have cloned and characterized the porcine homologue of tetherin and have tested the ability of both to inhibit PERV production.

Human tetherin mRNA sequence (GenBank accession number NM_004335.2) was submitted as a query in the pig expressed sequence tag (EST) database using basic local alignment search tool software (BLAST; http://www.ncbi.nlm.nih.gov/projects/genome/seq/BlastGen/BlastGen.cgi?taxid=9823). Among the first 18 identical hits, sequence EW580921.2 was arbitrarily chosen with which to design primers that anneal to the hypothetical translational start and end of the porcine tetherin candidate. Since the human tetherin gene is reported to be differentially expressed in human cell lines (21), we used three pig cell lines, PK15, MPK, and ST-IOWA cells, as a potential cDNA source for the pig homologue. The expected 533-bp cDNA fragments were obtained by reverse transcription-PCR (RT-PCR) using HotStart polymerase (Qiagen): MPK and ST-IOWA cDNAs revealed a perfect match with the EST EW580921.2; PK15 cDNA had two nonsynonymous substitutions, at nucleotides 351 and 427, probably reflecting the highly polymorphic nature of tetherin (10, 18) (Fig. 1A).

FIG. 1.

Sequence and expression of porcine tetherin. (A) Porcine tetherin cDNA was obtained by RT-PCR of the total RNA from PK15, ST-IOWA (IOWA), and MPK cells. The sequences were aligned against human tetherin using the ClustalW program (amino acids that are identical [*], conserved [:], and semiconserved [.] between porcine and human tetherins are indicated). The identity percentage between human and porcine tetherin amino acid sequences is 44%. The two amino acid differences between PK15 and IOWA/MPK tetherin sequences are boxed. Direct sequencing of the PK15 tetherin cDNA showed that PK15 cells express two types of cDNA: one the same as EW580921.2 and the other with changes at positions 351 and 427 as in the cloned PK15 fragment (data not shown). (B) One-eighth of the volume of cDNA obtained by reverse transcription of 1 μg of total RNA was processed in a SYBR green-based quantitative RT-PCR. Samples were run in triplicate. The number of copies for each gene was extrapolated from an analysis of the standard curves. Histograms represent porcine and human tetherin (THN) copies normalized to one 18S rRNA copy.

Human cell lines 293T and HeLa express significantly different levels of human tetherin, reflecting their differing ability to release retroviral particles (21). The level of porcine tetherin expressed in pig cells was determined by SYBR green-based quantitative PCR and compared to the level of human tetherin mRNA in 293T and HeLa cells (Fig. 1B). Tetherin mRNA expression in pig cells was estimated to be about 35 times higher than that in human 293T cells but five times lower than that in HeLa cells.

Because of the two amino acid difference in their sequences, both PK15 tetherin and MPK/ST-IOWA tetherin (IOWA tetherin), along with human tetherin, were tested for their ability to block release of gammaretrovirus particles (Fig. 2A and B). Recombinant PERV subgroup A (PERV-A) and amphotropic MLV (MLV-A) particles coding for enhanced green fluorescent protein (EGFP) were produced by transfection of 293T cells with tetherin-expressing plasmid or empty plasmid, and EGFP virus released to the culture supernatant was titrated as previously described (11). Both human and porcine tetherin reduced PERV-A and MLV-A titers 30- to 80-fold in comparison to those with an empty plasmid (pcDNA3) (Fig. 2A). Immunoblots using a polyclonal rabbit anti-PERV CA antibody (1) showed that expression of human and porcine tetherin reduced the level of mature virions (processed capsid protein p30) in the supernatant, but not cell-associated Gag (Fig. 2B). These results confirm that both porcine tetherins were able to inhibit the release of retroviral particles from human 293T cells to the same extent as did human tetherin. The two amino acid mutations in the PK15 tetherin sequence did not affect its restriction function.

FIG. 2.

Porcine tetherin blocks PERV and MLV release. EGFP-expressing viruses, PERV-A and MLV-A, were produced by plasmid transfection together with an expression plasmid for human tetherin (huTHN) PK15 tetherin (PK15THN), IOWA tetherin (IOWATHN), or an equal amount of empty plasmid (E). An expression plasmid encoding HIV-1 Vpu was also added where indicated. (A) Viral titers were determined by monitoring of EGFP expression by flow cytometry. Histograms represent the averages from at least two independent experiments (± standard errors of the means). (B) Cell lysate and supernatant were harvested 2 days after transfection and subjected to immunoblotting using rabbit polyclonal anti-PERV capsid antibody. In the cell lysates, capsid precursor (p60), the intermediate forms, and the processed capsid (p30) were visible. In the supernatant (SN), p30 was the main form represented.

Block of PERV release by PK15 tetherin and IOWA tetherin was also examined for sensitivity to an antitetherin countermeasure, human immunodeficiency virus type 1 (HIV-1) Vpu. Vpu expression rescued the PERV-A titer reduction by human tetherin, but not that by porcine PK15 tetherin and IOWA tetherin (Fig. 2B, compare results for PERV-A with and without HIV-1 Vpu). Resistance of porcine tetherin to the viral antitetherin countermeasures, HIV-1 Vpu and Env of simian immunodeficiency virus tantalus (SIVtan), has been shown elsewhere using vesicular stomatitis virus G protein (VSV-G)-pseudotyped HIV particles as the assay virus (11).

Human tetherin is type I IFN inducible. Upon treatment of 293T cells with alpha IFN (IFN-α), the human tetherin mRNA quantity increases more than 20-fold and reduces the yield of Vpu-deleted HIV-1 particles released in the supernatant about 10-fold (20, 21). To test whether porcine tetherin can reduce the amount of PERV particles released from pig cells, we first addressed whether endogenous tetherin was type I IFN responsive and then examined the effect of IFN treatment on PERV production in PK15 cells. Pig PK15 cells were treated for 24 h in the presence of 2,000 U/ml of IFN-β. Cells were lysed, and the amount of porcine tetherin mRNA was quantified by SYBR green-based quantitative RT-PCR. IFN-β induced a 15-fold increase in the porcine tetherin mRNA level compared to that of untreated cells (Fig. 3A). Serial dilutions of the supernatant from untreated and IFN-β-treated PK15 cells were employed to infect 293T cells. After 72 h, the PERV titer was determined by in situ immunostaining for PERV CA. The PERV titer from IFN-β-treated cells was 4-fold lower than that obtained from untreated cells (Fig. 3B). Consistently, the release of processed Gag to the supernatant by IFN-β-treated PK15 cells was significantly reduced compared to that by untreated PK15 cells (Fig. 3C). These results show that porcine tetherin, like human tetherin, is type I IFN inducible and that IFN-β treatment of PK15 cells reduces PERV release, possibly via pig tetherin induction.

FIG. 3.

Induction of tetherin and reduction of PERV particle release by IFN treatment of pig cells. PK15 cells were treated for 24 h with 2,000 U/ml of IFN-β. (A) Total RNA was analyzed by a SYBR green-based quantitative RT-PCR. Samples were run in triplicate, and the number of porcine tetherin (THN) copies were normalized to one 18S rRNA copy. Histograms represent the averages from two independent experiments (± standard errors of the means). (B) Supernatant from IFN-β-treated, and untreated, PK15 cells was titrated on 293T cells by in situ immunostaining using rabbit anti-PERV CA antibody. Histograms represent the averages from three independent experiments (± standard errors of the means). The reduction in the PERV titer by IFN treatment was found to be significant by t test (P = 0.002). (C) PK15 cells and their supernatants (SN) were processed by Western blotting, and PERV proteins were detected using anti-PERV CA antibody as described for Fig. 2.

The above results suggest the possibility that overexpressing tetherin in porcine cells could inhibit the release of continuously produced PERV particles. Human or pig tetherin genes were delivered into porcine PK15 cells by HIV-based retroviral particles also carrying the hygromycin B resistance gene (8), and cell populations which were hygromycin B resistant were obtained. First, to confirm the expression of tetherin in PK15 cells, total cellular RNA was processed in SYBR green-based quantitative PCR to measure the tetherin mRNA level. Bulk populations of PK15 tetherin and IOWA tetherin-transduced cells expressed, on average, 10-fold more pig tetherin mRNA than did parental cells (Fig. 4A). Human tetherin-transduced cells had a tetherin level similar to that of HeLa cells (Fig. 1B).

FIG. 4.

Exogenous expression of tetherin in PK15 cells decreases PERV release. PK15 cells were stably transduced by HIV-based vector encoding both human or porcine tetherin and hygromycin B resistance gene products. (A) Total RNA extracted from these cells was processed in a SYBR green-based quantitative RT-PCR. Samples were run in triplicate. Histograms represent the averages of the means from two independent experiments (± standard errors of the means) with porcine (poTHN) and human (huTHN) tetherin copies normalized to one 18S rRNA copy. (B) The supernatant from tetherin-transduced PK15 cells was titrated on 293T cells by in situ immunostaining using an anti-PERV CA antibody and calculated as percentage of the titer from parental cells. Histograms represent the averages from four independent experiments (± standard errors of the means). The statistical validity of PERV titer reduction in transduced-PK15 cells was assessed by t test. P values were 0.001 for huTHN, 0.028 for PK15THN, and 0.013 for IOWATHN. (C) Cell lysate and supernatant (SN) from tetherin-transduced PK15 or parental cells were immunoblotted using an anti-PERV CA antibody. Band intensities were determined by analysis with the Kodak 1D program, and the percentage of the parental cells was reported. wt, wild type.

Next, we analyzed how this affected PERV particle release to the supernatant. The day prior to infection, tetherin-expressing PK15 and parental cells were seeded in equal numbers. Serial dilutions of their supernatants were employed to infect 293T cells, and titers were determined by in situ immunostaining for PERV CA. The PERV titers from cells stably expressing PK15 tetherin and IOWA tetherin showed a 60% reduction compared to the titer from parental cells. Expression of human tetherin reduced the PERV titer to 23% of that of nontransduced cells (Fig. 4B). These data were supported by a Western blot analysis of the cell lysates and supernatants from tetherin-transduced PK15 cells. While the amount of Gag in the cell lysate appeared to be similar between all samples (Fig. 4C, cell lysate), the presence of mature particles in the supernatant of PK15 overexpressing tetherin was reduced to 16% of that of parental cells (Fig. 4C, SN).

This study has demonstrated that human tetherin and its porcine homologue can inhibit PERV release from producer cells. Our results suggest that constitutive overexpression of exogenously introduced tetherin (either human or porcine) in pig xenografts could reduce PERV dissemination in xenotransplantation settings. Human tetherin showed slightly greater effects than pig tetherin in our assays, but the difference was not statistically significant. Tetherins from different species have been shown to have different sensitivities to various viral countermeasures (12, 15, 25); for example, human, but not porcine, tetherin is sensitive to HIV-1 Vpu and SIVtan Env (11). It is therefore possible that expression of different tetherins could confer an antiviral function against a range of viruses other than PERV.

Considering the potential benefit of generating tetherin transgenic pigs as xenograft sources, the 5-fold reduction in PERV release from pig cells appears modest. However, further optimization of the tetherin expression system may improve this. Alternatively, the effect of exogenously expressed tetherin may be more pronounced in cells with lower levels of endogenous tetherin expression, such as 293T cells, than in PK15 cells, whose tetherin mRNA level is higher but still moderate (Fig. 1B). It is likely that endogenous tetherin expression is IFN regulated and not prevalent in vivo in pigs as in humans (9). In such circumstances, constitutive tetherin expression may reduce PERV dissemination substantially and represent a pig-engineering strategy to combat the PERV threat. It can be used in combination with other anti-PERV strategies, such as short hairpin RNA (shRNA) and APOBEC (6, 13, 22).