Is PERV transfer across hollow fiber membranes relevant to bioartificial liver model?

Abstract

We read with interest the article from Nyberg et al. regarding transfer of porcine endogenous retrovirus (PERV) across hollow-fiber (HF) membranes, published in a recent issue of Transplantation (1). The authors showed that pore size, membrane composition, and duration of exposure might influence the transfer of PERV across hollow-fiber membranes. HF were loaded with PK15 cells, which produce PERV. Moreover, the use of porcine hepatocytes, as the source of PERV, might be more significant to a bioartificial liver. We performed in vitro bioartificial liver procedures, in triplicate, using porcine hepatocytes, previously positive for PERV DNA, and RNA, respectively, by PCR and reverse transcriptase-polymerase chain reaction (RT-PCR) (2). HF, made of polypropylene, were loaded with 1.5×1010 hepatocytes of porcine origin. Porosity of the hollow fiber membranes was 600 nm. Human plasma recirculated through the HF for 3 hr at a flow rate of 200 ml/min. Plasma samples were collected before recirculation (T0) and at the end of the procedure (T3) and stored at −80°C. Infectivity assay was performed incubating plasma with MRC5 cells (not susceptible to PERV) and human embryonic kidney 293 cells (susceptible to PERV) for 6 and 24 hr (3). Supernatants were collected and tested for PERV RNA by RT-PCR, and cells were tested for PERV DNA by PCR (4, 5). PK15 cells were used as positive controls. Plasma samples collected at T0 and T3 were negative for both PERV DNA and RNA by PCR and RT-PCR, respectively. MRC5 and HK293 cells were negative for both PERV DNA and RNA after 6 and 24 hr of incubation with plasma collected at T0 and T3. Our results showed lack of transfer of PERV from porcine hepatocytes to human plasma recirculated through polypropylene HF of 600 nm of porosity. Moreover, we did not find any evidence of infection by PERV in MRC5 and HK293 cells incubated with human plasma previously recirculated through HF loaded with porcine hepatocytes. We consider that in both studies, pore size diameter was too large to be selective for viral particles. Further studies will be needed to better define whether porcine hepatocytes are able to produce PERV particles during the restricted time of clinical BAL application. We finally agree with Nyberg et al. when they suggest that pore size, membrane composition, and duration of exposure might influence the transfer of PERV across HF membranes. Elisabetta Falasca1 Umberto Baccarani Corrado Pipan Alberto Degrassi Annibale Donini