Interventions are urgently needed for the problem of congenital CMV infection. Although CMV vaccines are in various stages of preclinical development, none are currently licensed for human use. Antiviral agents for CMV, in contrast, are available, and have already been shown to be useful in the management of congenital and perinatal CMV infections [
2,
13]. However, there is no conclusive evidence that antiviral treatment of pregnant women is capable of interrupting CMV transmission. Since the guinea-pig model of CMV infection is the most authentic non-primate animal model for study of interventions against congenital CMV infection [
14], these studies were undertaken to test the hypothesis that an effective antiviral agent might be capable of reducing transmission of CMV during pregnancy.
The intrinsic resistance of GPCMV to the most commonly used CMV antiviral, ganciclovir, precludes evaluation of this agent for efficacy against congenital GPCMV infection [
7,
8]. However, the cyclic cogener of cidofovir is highly active against GPCMV
in vitro and, therapeutic against disseminated GPCVM disease, and against viral labyrinthitis and SNHL,
in vivo [
8,
9]. The molecular basis for the improved activity of this agent against GPCMV, compared to ganciclovir, is not clear, but it may derive from the fact that cidofovir does not require phosphorylation by the CMV U
L97 kinase for antiviral activity, in contrast to ganciclovir. Since the GPCMV U
L97 homolog has significant divergence (compared to HCMV) of amino acid sequence in catalytic domains of the protein important for ganciclovir phosphorylation, this may provide a molecular explanation for the decreased susceptibility of GPCMV to this nucleoside antiviral [
15]. In any case, cyclic cidofovir has been found to be not only highly effective against GPCMV, but also non-toxic to animals [
8]. The study described in this report found that, following challenge with an eGFP-tagged strain of GPCMV (vAM403), guinea pigs that received cyclic cidofovir therapy (20 mg/kg, via intraperitoneal route) had reduced overall maternal and pup mortality, compared to placebo-treated controls. This difference in overall mortality was statistically significant, with a reduction from 5/25 (20%) in the placebo group, to 0/21 (0%) animals in the cyclic cidofovir-treated group (p = 0.05, Fisher's exact test). Furthermore, these studies found that antiviral therapy resulted in elimination of congenital GPCMV transmission. In untreated controls, congenital GPCMV infection was demonstrated in 5/19 pups (26%), versus 0/16 pups in the cyclic cidofovir group (p < 0.05, Fisher's exact test). These are the first data to demonstrate that antiviral therapy during pregnancy can prevent congenital CMV transmission. These observations provide support for more detailed examination of cyclic cidofovir, and other antiviral agents, in this small animal model of CMV transmission. These data may also provide indirect support for potential clinical trials of antiviral interventions in pregnant women at high risk of transmission of CMV to the fetus, although caution should be exercised in human studies, given the potential for toxicity of these agents when used during pregnancy [
13]. A recent study of passive immunization of women at high risk for giving birth to infants with symptomatic congenital CMV infection suggested that antibody therapy
in utero resulted in improved pregnancy outcomes [
16]. Future controlled clinical trials will be required to further assess the value as well as the potential toxicities of antiviral interventions targeting CMV transmission in the pregnant patient.
An important observation from these studies was the demonstration of the utility of using an eGFP-tagged, recombinant CMV in the experimental modelling of congenital CMV infection. Previous studies using an eGFP-tagged rhesus macaque CMV have demonstrated that the resultant recombinant virus retained full pathogenicity in the experimentally inoculated fetus
in utero, but this required direct inoculation of the fetus [
17]. It was therefore of considerable interest to determine if the recombinant GPCMV used in these studies was capable of resulting in transplacental infection and conferring maternal and fetal disease in pregnant guinea pigs without direct fetal inoculation. Characterization of the eGFP-tagged GPCMV recombinant, vAM403, indicated that insertion of the eGFP/
gpt cassette into the GPCMV genome, in the nonessential
Hin d III "N" locus, had minimal impact on viral protein expression compared to wild-type virus, as assessed by SDS-PAGE. This observation, coupled with the apparent absence of any conserved betaherpesvirus proteins encoded by this region of the genome, as ascertained by DNA sequence analyses [
6], suggested that this virus would have the full potential for virulence in animals. Consistent with this prediction, this virus was found to be fully capable of transplacental transmission, with attendant maternal and pup mortality, following subcutaneous inoculation of pregnant guinea pigs. The future use of other eGFP-tagged viruses should facilitate study of congenital GPCMV infection in this model. Targeted mutagenesis of the GPCMV genome cloned in
E. coli as an infectious bacterial artificial chromosome (BAC) [
18] will facilitate future, more detailed evaluation of the role of viral genes in pathogenesis of maternal-fetal infection.