INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Biological products derived from porcine plasma are currently licensed for clinical use. For example, hemophiliacs who develop antibodies to human factor VIII can effectively use porcine plasma-derived factor VIII (6, 19). In addition to porcine plasma, porcine organs have been considered for use in human recipients to alleviate the shortage of human organs for transplantation. Porcine cells, such as fetal neuronal cells, are now under clinical investigation for treatment of human disease (7). One problem with the safety of porcine organs and plasma is the potential to expose humans to pig-derived infectious agents. Of particular concern are the findings that the genomes of all pigs contain porcine endogenous retroviruses (PERVs) that are capable of giving rise to gammaretrovirus particles (2, 3, 12, 13, 22, 24). While recent reports have not found evidence for PERV infection of humans (10, 20, 21), little is known about levels of PERV expressed in vivo in pigs or the impact of exposure to PERV in different clinical settings, such as repeat injections of plasma derivatives or long-term exposure to transplanted pig cells.

The swine genome contains 20 to 50 copies of PERV, 10 to 20 copies of which correspond to full-length provirus (1, 12, 22). Sequence analysis of full-length clones indicates that PERV is most closely related to gibbon ape leukemia and murine leukemia virus (MuLV) genomes (1, 5), sharing approximately 60% amino acid identity with the gag and pol regions. In addition to there being multiple copies of PERV per genome, there are three known receptor classes based on sequence analysis, interference, and in vitro tropism studies, named PERV-A, PERV-B, and PERV-C (1, 12, 25). Retroviral pseudotypes bearing envelopes of class A or B are capable of infecting cell lines from several mammalian species, including human, while pseudotypes carrying PERV-C envelope are mostly restricted to cells of porcine origin (25).

PERVs derived from certain porcine cell lines (22), primary porcine endothelial cells (14), primary porcine islet cells (27), and activated porcine peripheral blood mononuclear cells (PBMC) (28) are capable of infecting various human cell lines. Recovery of infectious virus from both porcine PBMC (28) and porcine endothelial cells (14) suggests that one might expect PERV to be present in pig plasma. While reverse transcriptase (RT) activity and PERV-specific RNA have been found in porcine serum (10), it is not known whether PERV present in porcine plasma is infectious.

Over the last two decades, approximately half of the hemophiliac population contracted human immunodeficiency virus from contaminated plasma (4, 11). Given the use of porcine plasma derivatives and the current exploration of porcine cells and organs for human xenotransplantation, it is critical to ascertain if porcine plasma and its derivatives carry infectious PERV. In this study, porcine plasma and a plasma derivative, factor VIII (trade name, Hyate:C), were evaluated for the presence of RT activity and for PERV env sequences. In addition, we investigated whether infectious virus is present in either porcine plasma or the porcine plasma-derived factor VIII.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cells and viruses. Four different cell lines were used in infectivity assays: human embryonic kidney 293 cells (ATCC CRL-1573), swine testis ST-IOWA cells (obtained from Richard Fister, Tufts University), human fibrosarcoma HT1080 cells (ATCC CCL-121), and a feline line of glial cells (PG-4 cells; ATCC CRL 2032). ST-IOWA and 293 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (FBS), 2 mM glutamine, 1 mM sodium pyruvate, 100 U of penicillin per ml, and 100 µg of streptomycin per ml. HT1080 cells were maintained in Eagle's minimal essential medium supplemented with 10% FBS, 1× nonessential amino acids (Biofluids, Rockville, Md.), 0.75% sodium bicarbonate (BioWhittaker, Walkersville, Md.), 1 mM sodium pyruvate, 100 U of penicillin per ml, and 100 µg of streptomycin per ml. PG-4 cells were maintained in McCoy's 5A medium supplemented with 15% FBS, 2 mM glutamine, 1 mM sodium pyruvate, 100 U of penicillin per ml, and 100 µg of streptomycin per ml. As a positive control for infectivity studies, virus-containing supernatant was harvested from postconfluent 293 cells infected with PERV (PERV-NIH-3°) and passed through a 0.45-µm-pore-size filter to remove residual cells. This PERV was originally isolated from the supernatant of phytohemagglutinin- and phorbol myristate acetate-stimulated PBMC isolated from NIH minipigs and passaged onto 293 cells (28). The virus obtained from this culture (PERV-NIH-1°) was subsequently passaged two additional times in 293 cells to produce PERV-NIH-3°, which hereafter will be referred to as PERV-NIH. PERV-NIH has been shown to contain envelope sequences most similar to those of PERV-A (29).

Plasma virus isolation. Porcine peripheral blood (100 ml) was collected in acid citrate dextrose from NIH minipigs (NIH animal facility, Poolesville, Md.). Blood was centrifuged at 200 × g for 15 min to clarify plasma, and plasma was centrifuged at 1,000 × g for 15 min to remove any residual cells. To purify and concentrate PERV that might be present, 10 ml of freshly collected plasma was diluted with 19 ml of phosphate-buffered saline and ultracentrifuged over 7.5 ml of 20% (wt/vol) sucrose in TEN (10 mM Tris [pH 8.0], 1 mM EDTA, 100 mM NaCl) in a Beckman SW28 rotor (25,000 rpm, 3 h, 4°C). The pelleted material was resuspended in 1.5 ml of appropriate culture medium and immediately used in infectivity studies. Pilot experiments using analogous ultracentrifuge conditions resulted in reproducible recovery of infectious PERV-NIH, albeit with reductions in infectivity titer (data not shown).

Detection of PERV in porcine plasma and Hyate:C. Plasma samples and clinical lots of Hyate:C were assayed for RT activity by a product-enhanced RT assay that utilizes TaqMan technology for quantitative detection (TM-PERT), performed as previously described (17). Absolute RT activities were determined by running serial dilutions of known activities of either avian myeloblastosis virus or Moloney MuLV RT (Life Technologies, Rockville, Md.). All plasma samples were ultracentrifuged prior to use in the TM-PERT as described above (i.e., 10 ml was pelleted and resuspended in 1.5 ml of medium). In all cases, the test article was examined for inhibitory effects in the assay by including a sample of the test article spiked with the positive control RT. Spiked test articles resulted in values comparable with the positive control without the test article, indicating no inhibitory effects. Negative controls of buffer without sample were run with every TM-PERT assay, and no RT activity was observed in each case. The assay baseline was the lowest concentration of positive control RT that produced a fluorescence value above threshold levels in the TaqMan assay as previously described (17).

Infectivity assays using porcine plasma. Target cells were seeded into 24-well dishes at a concentration of 50,000 cells/well 1 day prior to exposure to virus. On the day of infection, PERV was harvested from postconfluent 293/PERV-NIH cultures for use as a positive control. Dilutions of PERV-NIH were prepared in culture medium, adjusted to a final Polybrene concentration of 6 µg/ml, and added to each target cell type. For plasma-derived material, ultracentrifuged pellets were resuspended in 1.5 ml of culture medium containing 6 µg of Polybrene/ml. One milliliter of either ultracentrifuged plasma or PERV-NIH-containing supernatant was added to the target cells. The following day, cells were washed with fresh medium. Negative controls were parallel cultures grown and maintained in a similar manner that were not exposed to PERV-NIH or porcine plasma.

Infectivity assays using Hyate:C. Six lots of lyophilized Hyate:C were reconstituted with 20 ml of sterile water for injection (USP grade) according to the manufacturer's instructions. Each lot of reconstituted Hyate:C was diluted fivefold in either culture medium or PERV-containing supernatant and adjusted to a final Polybrene concentration of 6 µg/ml (based on a previous analysis of cytotoxicity of Hyate:C, it was determined that a fivefold dilution was optimal). These Hyate: C preparations were then inoculated onto triplicate wells of 12-well dishes that had been seeded 1 day previously with porcine ST-IOWA, human 293, or feline PG-4 cells. Cultures were maintained for a total of 8 weeks, with passaging twice per week. Cell supernatants were sampled every 2 weeks and assayed for PERV production by either RT activity (ST-IOWA cultures) or RT-PCR for viral RNA (293 and PG-4 cultures) (29). All positive-control cultures were assessed for RT activity by the conventional assay (28). To determine whether latent virus may be present in the Hyate:C-inoculated cultures, genomic DNA was isolated at week 8 and examined for PERV sequences by PCR as previously described (28). Negative controls were parallel cultures not exposed to Hyate:C or PERV-NIH. Positive controls included cultures inoculated with various dilutions of PERV-NIH.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Analysis of porcine plasma for presence of PERV. Plasma was isolated from four different NIH minipigs and purified by ultracentrifugation through sucrose as described in Materials and Methods. TM-PERT was used to detect retrovirus in the ultracentrifuged plasma (17). As seen in Fig. 1, RT activity was present in all four minipig plasma samples. The RT activity of the NIH minipig plasmas ranged from 3.3 × 10⁶ to 2.4 × 10⁷ pU of RT activity per µl, while the RT activity of the positive control PERV-NIH supernatant was 8.8 × 10⁸ pU/µl. These values were more than 3 orders of magnitude over the assay baseline of 10³ pU/µl. In parallel, each sample was independently spiked with 3.3 × 10⁷ pU of RT enzyme. Levels of activity were comparable with the activity of a standard used as a spike, demonstrating lack of inhibitory effects (data not shown).

View larger version (17K): [in this window] [in a new window]   FIG. 1.   RT activity in porcine plasma. All plasmas were ultracentrifuged through 20% sucrose prior to testing as described in Materials and Methods. Samples were assayed for RT activity by TM-PERT (17). Values were standardized using a curve generated from assay of serial dilutions of a stock of Moloney MuLV RT with known enzymatic activity. The assay baseline was 103 pU/µl.

Susceptibility of target cells to PERV-NIH. To determine whether the porcine plasma contained infectious PERV particles, three different cell lines were used to allow detection of any of the three receptor classes of PERV. Previous experiments have shown ST-IOWA cells to be permissive to retroviral pseudotypes containing each of the three env classes, while 293 and PG-4 cells are permissive to virus containing class A or class B env (25, 29). Plasmas 1 to 4 were tested in independent culture experiments with positive and negative controls included for each experiment. Positive controls were serial dilutions of PERV-NIH supernatant added to each target cell culture in parallel to exposure to pig plasma to demonstrate that culture conditions were able to support PERV replication. Infection of target cells was monitored by detection of RT activity. All cell lines tested in each of the four assays were permissive for PERV-NIH, although to varying degrees (Table 1).

Analysis of porcine plasma for infectious PERV. Plasmas from the four NIH minipigs were isolated and then ultracentrifuged over 20% sucrose to purify viral particles. Pelleted material was resuspended in culture medium and added to duplicate wells of target cells. To detect PERV replication, we assayed culture supernatants for RT activity and for viral pol sequences by RT-PCR for a period of 8 weeks. PERV replication was not detected in PG-4 and 293 cells exposed to NIH minipig plasmas. However, the ST-IOWA cells became positive for RT activity after exposure to each of the ultracentrifuged plasma samples. In Fig. 2, the levels of [³H]TTP incorporation measured in supernatants from the ST-IOWA/plasma cultures over time are shown. The average RT activity of the plasma cultures at week 8 was increased 17.1 ± 4.8 (average ± standard deviation) times the values obtained from parallel cultures of uninfected ST-IOWA cells. This consistent high increase in RT over time activity indicates that the ST-IOWA cells were infected with plasma-derived virus and that virus spread throughout the culture.

View larger version (18K): [in this window] [in a new window]   FIG. 2.   Time course of ST-IOWA infection with porcine plasma-derived virus. Culture supernatants were assayed for RT activity in a conventional RT assay (28) at weekly intervals. All control uninfected ST-IOWA cells had consistently low endogenous RT activity, and values from one representative uninfected control are shown.

Envelope classification of infectious plasma virus. Of the three cell lines used in this experiment, only ST-IOWA cells are permissive to infection by pseudotypes of PERV expressing class C envelope (25). To determine the envelope class of the infectious PERV isolated from the pig plasma, RT-PCR analysis was done on the supernatants from the ST-IOWA cells exposed to plasma 8 weeks previously, using envelope primers specific for class C (1). Since ST-IOWA cells express endogenous envelope A and B sequences (25), only expression of PERV class C sequences could be analyzed. As seen in Fig. 3, class C envelope sequences were detected in each of the four ST-IOWA cultures exposed to ultracentrifuged plasma samples, while these sequences were not detected in any of the control uninfected ST-IOWA cells or in the cultures exposed to PERV-NIH. Therefore, infection of ST-IOWA cells exposed to porcine plasma was likely the result of plasma-derived PERV that has a type C envelope.

View larger version (43K): [in this window] [in a new window]   FIG. 3.   Analysis of viral env RNA from ST-IOWA cell cultures after exposure to NIH minipig plasmas. RT-PCR analysis was done on supernatants from replicate cultures of ST-IOWA cells after exposure to each of the NIH minipig plasma samples 1 to 4, PERV-NIH, or uninfected control cells. Primers specific for class C PERV envelope sequences were used to amplify PERV-C env from these samples (25), and DNA products were detected by Southern blot analysis using a class C envelope-specific probe.

Detection of PERV in Hyate:C. To determine whether PERV would be present in a manufactured product derived from pig plasma, we analyzed six clinical lots of Hyate:C (porcine factor VIII) for the presence of PERV RNA by RT-PCR for pol sequences and for evidence of retroviral particles by measuring RT activity using TM-PERT. As shown in Fig. 4A, PERV pol RNA could be detected by RT-PCR in all lots examined. To assess whether the viral RNA may be associated with viral particles, we next tested the same lots for the presence of retroviral RT activity by TM-PERT (17). Using this method, all six lots analyzed had measurable levels of RT activity in the range of 10⁵ to 10⁶ pU of RT activity per µl of Hyate:C analyzed (Fig. 4B). By comparison, PERV-NIH-containing supernatant was shown to have activity 4 to 5 logs higher in the same assay (RT activity of approximately 3.3 × 10¹⁰ pU/µl).

View larger version (28K): [in this window] [in a new window]   FIG. 4.   Detection of PERV in Hyate:C. Six lots of Hyate:C, labeled A through F, were examined for PERV by RT-PCR for PERV RNA pol sequences (A) or for RT activity by TM-PERT (B). The values for RT activity were derived from a standard curve generated using dilutions of a quantitated stock of avian myeloblastosis virus RT measured in the same assay. The assay baseline was 103 pU/ml.

View larger version (57K): [in this window] [in a new window]   FIG. 5.   (A) DNA PCR for PERV pol in 293 and PG-4 cells exposed to Hyate:C. Human 293 and feline PG-4 cells were exposed to each of six lots of Hyate:C in triplicate (labeled A through F). After 8 weeks of culture, DNA was isolated from each culture, and PERV pol sequences were amplified and hybridized to a pol-specific probe as described in Materials and Methods. P represents the PERV-NIH positive control. (B) RT activity in cell lines exposed to Hyate:C inoculated with PERV. Human 293, feline PG-4, or swine ST-IOWA cells were exposed to 1:2 dilution of PERV-NIH-containing supernatant alone (no Hyate:C) or to one of six lots of Hyate:C (labeled A through F) diluted 1:5 in the diluted PERV-NIH-containing supernatant. Cells not exposed to virus were negative controls for the experiment. The y axis represents [3H]TTP incorporated in a conventional RT assay as described in Materials and Methods; values for the negative controls were all less than 1,000 cpm.

Analysis of Speywood plasma infectivity. To determine whether the porcine plasma that is used as the starting material to manufacture Hyate:C contains infectious PERV, we obtained a sample of plasma from a pool of approximately 60 pigs used by Speywood BioPharm. Measurable RT activity was demonstrated by TM-PERT in this plasma sample, as shown in Fig. 1. Although RT levels were lower than those observed for NIH minipig plasmas (4.1 × 10⁴ versus 3.3 × 10⁶ pU/µl for the lowest NIH plasma sample), the activity detected was over 1 log above the assay baseline (10³ pU/µl). Additionally, RT-PCR for PERV envelope sequences revealed that the Speywood plasma contains all three envelope classes (data not shown).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In this report, we evaluated five porcine plasmas for the presence of PERV. All plasma samples examined had measurable levels of RT activity as detected by TM-PERT and PERV RNA as detected by RT-PCR. Four different cell substrates were used in this study to facilitate detection of PERVs with different in vitro tropisms. Viral RNA expression was detected in the culture supernatants of ST-IOWA cells that were exposed to each of four NIH minipig plasmas. In addition, these cultures were monitored for RT activity. The kinetics and amplitude of the RT activity measured indicated that the ST-IOWA cells were productively infected with virus derived from the porcine plasma.

RT-PCR analysis for envelope in culture supernatants of ST-IOWA cells exposed to porcine plasma revealed that the envelope was class C in all four samples. The absence of PERV-C env detection in uninfected ST-IOWA cells or in control cultures of ST cells exposed to PERV-NIH demonstrates that the PERV-C envelope sequences must have originated from the minipig plasma samples. In vitro tropism studies using retroviral pseudotypes demonstrate that PERV-C is mostly restricted to cells of porcine origin, with the one exception of the human fibrosarcoma cell line HT1080 (25). Although PERV-C does not appear to be infectious for most human cell lines examined to date, there may nonetheless be human cells in vivo that are susceptible to PERV-C infection. Furthermore, it is important to note that previous studies of MuLV have shown that as few as two to three amino acid substitutions in the envelope protein can change viral tropism (8, 15, 16). For example, Han et al. demonstrated that three amino acid changes in an amphotropic MuLV envelope allowed pseudotypes bearing this mutated envelope to infect Chinese hamster ovary cells; this represents an expanded species tropism for amphotropic MuLV (8). We do not know whether a similarly small change in PERV-C could alter the species tropism, but with mutations accumulating during viral replication, it is possible that a humantropic variant could arise.

In four of four plasma samples, neither human 293 nor feline PG-4 cell cultures were permissive for plasma-derived PERV, although both cell lines were permissive for PERV-NIH. Based on previous findings, it was expected that 293 and PG-4 cells were permissive for PERV expressing class A and class B envelopes (25, 29). Interestingly, sequences for all three PERV env classes were detected in the source plasma samples. These results suggest that the viral genomes encoding class A and class B env either were not generating infectious particles or were present at a level too low to be detected in our culture assay. Alternatively, we cannot exclude the possibility that PERV-A and PERV-B may be more labile than PERV-C upon ultracentrifugation though a sucrose cushion. Detection of infectious virus only in ST cells correlates with the previously reported tropism of PERV-C env pseudotypes that could not infect 293 cells or a feline cell line (25).

The finding that infectious virus could be isolated from porcine plasma suggested that PERV may be present in biologics manufactured using porcine plasma as a source material. To examine this possibility, we analyzed one pig plasma derivative, porcine factor VIII (Hyate:C), and showed it to contain PERV-specific RNA sequences and RT activity. These findings confirm observations made by Heneine and coworkers (9). Here, we extended these findings by looking specifically for evidence of infectious virus in the clinical lots of Hyate:C. Infectious PERV was not detected in any of the six lots analyzed despite extensive cultures with cell lines permissive of each of the three known PERV envelope classes. These infectivity data suggest either that the PERV associated with Hyate:C is noninfectious or that infectious PERV is present at a level below the current limit of detection. In this experiment, the 1:1,000 dilution of PERV-NIH could be detected in 293 and ST-IOWA cells but not PG-4 cells. The activity of RT detected in clinical lots of Hyate:C was approximately 4 to 5 logs lower than that of a typical PERV-NIH-containing tissue culture supernatant. One could infer that the amount of virus our culture system can detect is 10- to 100-fold higher than the amount of PERV in Hyate:C.

To investigate whether the source plasma used to make porcine factor VIII contains infectious PERV, we analyzed a pooled plasma sample obtained from Speywood BioPharm. Infectious virus was not detected, although the sample was positive for both PERV RNA and RT activity. Absence of infectious virus in this sample could be due to the fact that the RT activity detected in the Speywood sample was 80 times lower than that in the lowest-activity NIH minipig plasma, representing a level that could be below our assay limit of detection. Alternatively, other differences between the Speywood plasma and the NIH minipig plasmas could contribute to the different results. The Speywood sample had been previously filter sterilized and frozen. This process may have removed or inactivated the infectious PERV to below detectable levels. Additionally, breed-specific differences may influence the amount of infectious PERV found in plasma. With the data currently available, it cannot be concluded which of these factors affected the result or whether the result was due to combined effects of all of them.

The absence of detectable infectious virus in both the Speywood plasma and the clinical lots of Hyate:C provides support for the continued use of Hyate:C in the treatment of individuals with hemophilia who have developed inhibitors to human factor VIII. Based on this analysis, it is unclear whether the pigs used to produce Hyate:C may contain a lower level of PERV in their plasma or whether the manufacturing process may contribute to a reduction in infectious virus in the clinical lots. Lyophilization has been shown to reduce the infectious titer of retroviruses (23, 26).

In summary, isolation of infectious PERV from porcine plasma of NIH minipigs demonstrates that in vivo there are porcine cells that constitutively express infectious PERV virions, reinforcing the inherent risk of exposure to PERV when porcine cells are used for human xenotransplantation. In addition, we show here that porcine plasma derivatives have the potential to carry virus, although it may be noninfectious. Our analyses suggest that differences in pig breeds and the incorporation of manufacturing steps that remove or decrease the load of infectious virus in the final product will reduce the risk to patients exposed to these products.