Expression of the human cytomegalovirus UL97 gene in a chimeric guinea pig cytomegalovirus (GPCMV) results in viable virus with increased susceptibility to ganciclovir and maribavir
Introduction
Congenital infection with human cytomegalovirus (HCMV) is a major public health problem in the developed world. Congenital HCMV infection occurs in up to 2% of all deliveries, and is responsible for a variety of neurodevelopmental sequelae, including sensorineural hearing loss (SNHL) (Demmler, 1996). Although there is no licensed vaccine to HCMV, antiviral agents are available. A recent study of antiviral treatment of newborns with symptomatic congenital HCMV infection involving the central nervous system (CNS) indicated that ganciclovir therapy resulted in improved hearing outcomes (Kimberlin et al., 2003). Another recent study of passive immunotherapy, using high-titer HCMV antibody, suggested that treatment of HCMV infection in pregnancy resulted in improved neurodevelopmental outcomes in congenitally infected infants (Nigro et al., 2005). Thus, there is considerable interest in evaluating antivirals for use in treatment of the fetus and newborn infant, toward the goal of improving pregnancy outcomes.
Ideally, antiviral therapies would undergo pre-clinical evaluation in an animal model prior to licensure for human use. However, the species specificity of HCMV precludes study of this virus in animals, and necessitates the study of species-specific CMVs in their respective animal hosts. Among the small animal models, the guinea pig cytomegalovirus (GPCMV) model has unique advantages over other rodent models. These advantages include a disease pathogenesis similar to that observed in humans, including neurological injury and SNHL (Bia et al., 1983, Schleiss and Lacayo, 2006). Additionally, GPCMV is transmitted in utero producing congenital infection and disease in the newborn pup (Schleiss and Lacayo, 2006). This feature makes the GPCMV model particularly relevant to studies of antiviral therapy in pregnancy. However, there are disadvantages to the GPCMV model for the study of antivirals. GPCMV is resistant to medically relevant doses of the most commonly used antiviral for HCMV, ganciclovir (Matthews and Boehme, 1988, Williams et al., 2003), and attempts to study this agent in the guinea pig model have met with limited success (Fong et al., 1987, Woolf et al., 1988). A newer agent being explored in clinical trials, the benzimidazole riboside 1263W94 (Maribavir [MBV]; Biron et al., 2002) is similarly inactive against GPCMV (Williams et al., 2003). Other antivirals, such as cyclic cidofovir (HPMPC) and BAY 38-4766, are active against GPCMV in vitro and in vivo, but these agents are not yet licensed for use in HCMV infection (Bourne et al., 2000, Bravo et al., 2006, Schleiss et al., 2006a, Schleiss et al., 2006b, White et al., 2006, Kern, 2006).
The molecular basis for the decreased susceptibility of GPCMV to GCV and MBV, compared to HCMV, is unknown. To become active, GCV requires phosphorylation by the UL97 kinase protein prior to conversion into the active triphosphate form (Sullivan et al., 1992). GCV-resistant viral strains of HCMV can be generated in tissue culture under GCV selection or through prolonged drug therapy in patients, and resistant strains are characterized by mutations to genes UL54 or UL97 (Baldanti et al., 2004, Gilbert and Boivin, 2005, Chou et al., 1995). MBV-resistant strains of HCMV also have mutations in the UL97 gene, indicating a common target for both drugs, although the mode of action by MBV is different from that of GCV (Biron et al., 2002). All animal CMVs, including GPCMV, appear to encode a homolog to the HCMV UL97 gene (Fox and Schleiss, 1997, Michel and Mertens, 2004, Romaker et al., 2006). A comparison of the UL97 ORF of wild-type and GCV-resistant HCMV strains with GP97 demonstrated divergence at key amino acids that were hypothesized to play a role in resistance to GCV (Fox and Schleiss, 1997, Michel and Mertens, 2004). Therefore, we evaluated whether GPCMV could be rendered sensitive to GCV if the virus were engineered to encode the HCMV UL97 gene in the place of GP97. An infectious bacterial artificial chromosome (BAC) of the GPCMV genome in Escherichia coli was used to generate an intertypic, chimeric virus encoding the HCMV UL97 gene (GPCMV::UL97), and the antiviral susceptibility of this virus to GCV was compared to that of wild-type GPCMV. Additionally, a GP97 knockout (GPCMVdGP97) was generated to explore the role of the GP97 protein in the virus life cycle.
Section snippets
Cells, viruses and oligonucleotides
GPCMV (strain 22122, ATCC VR682), GPCMV BAC derived virus and vAM403, an enhanced green fluorescent protein (eGFP)-tagged wild-type GPCMV (McGregor and Schleiss, 2001) were propagated on guinea pig fibroblast lung cells (GPL; ATCC CCL 158) in F-12 medium supplemented with 10% fetal calf serum (FCS; Gibco-BRl), 10,000 IU of penicillin/l, 10 mg of streptomycin/l (Gibco-BRL) and 7.5% NaHCO3 (Gibco-BRL). HCMV (strains Towne and Toledo) were propagated on human foreskin fibroblast (HFF) cells in DMEM
Transient expression of GP97 in GPL cells
The HCMV UL97 protein targets the cell nucleus, but in contrast the murine cytomegalovirus (MCMV) homolog, M97, is cytoplasmic (Michel et al., 1996, Michel et al., 1998, Wagner et al., 2000); moreover, although MCMV is sensitive to GCV, this antiviral effect is not dependent on M97 (Wagner et al., 2000). Hence, prior to any in-depth studies with GP97, we sought to determine if this protein targeted the nucleus. The full-length GP97 ORF (605 aa) was cloned into the C-terminal domain of an eGFP
Discussion
Although antiviral therapy requires further study as an interventional strategy for congenital HCMV infection, this virus cannot be effectively studied in animal models. The guinea pig is unique as the only small animal model for congenital CMV infection, likely due to the striking similarities in placenta structure and biology between guinea pigs and humans (Kaufmann, 2004, Carter, 2007, Mess, 2007). Since GPCMV congenital infection has a similar pathogenesis in guinea pigs as occurs in humans
Acknowledgements
We are grateful to Don Coen (Harvard University) and Bill Britt (University of Alabama/Birmingham) for providing anti-UL97 and anti-GPCMV gB antisera, respectively. We thank Jodi Anderson (University of Minnesota, Minneapolis, MN), Greg Stroup and Nanette Huey (Cincinnati, OH) for technical assistance. This work was supported by National Institute of Health AI-65289, DC008651, HD38416-01, and March of Dimes Basic Research Grants 6-FY98/99-0416 and FY01-226. A.M. is a recipient of a Minnesota
References (46)
- et al.
Structural changes in human cytomegalovirus cytoplasmic assembly sites in the absence of UL97 kinase activity
Virology
(2006) - et al.
Clinical and biologic aspects of human cytomegalovirus resistance to antiviral drugs
Hum. Immunol.
(2004) - et al.
Cyclic HPMPC is safe and effective against systemic guinea pig cytomegalovirus infection in immune compromised animals
Antiviral Res.
(2000) Animal models of human placentation—a review
Placenta
(2007)- et al.
Maribavir sensitivity of cytomegalovirus isolates resistant to ganciclovir, cidofovir or foscarnet
J. Clin. Virol.
(2006) - et al.
Antiviral effect of 9-(1,3-dihydroxy-2-propoxymethyl)guanine against cytomegalovirus infection in a guinea pig model
Antiviral Res.
(1987) Pivotal role of animal models in the development of new therapies for cytomegalovirus infections
Antiviral Res.
(2006)Effect of ganciclovir therapy on hearing in symptomatic congenital cytomegalovirus disease involving the central nervous system: a randomized, controlled trial
J. Pediatr.
(2003)- et al.
Molecular cloning of the guinea pig cytomegalovirus (GPCMV) genome as an infectious bacterial artificial chromosome (BAC) in Escherichia coli
Mol. Genet. Metab.
(2001) - et al.
Identification of essential and non-essential genes of the guinea pig cytomegalovirus (GPCMV) genome via transposome mutagenesis of an infectious BAC clone
Virus Res.
(2004)