Effects of famciclovir and valacyclovir on herpes simplex virus type 1 infection, latency, and reactivation in mice: how dissimilar are study results?

To the Editor—LeBlanc et al. [1] assessed the effects of famciclovir and valacyclovir on herpes simplex virus (HSV) type 1 infection, latency, and reactivation in mice but did not demonstrate superiority of famciclovir over valacyclovir on the establishment of HSV latency in mice. Furthermore, they assert in both the Introduction and Discussion sections of their article that their results “differ notably from the serial comparative mouse studies of famciclovir and valacyclovir” [1, p. 598] published by my colleagues and me [2–5]. We wish to draw attention to 4 important points that we believe explain the declared discrepancy between studies in our 2 laboratories. First, we infected a strain of moderate virulence (HSV-1 SC16) into the skin of the murine ear pinna, at a dose of 10 pfu per mouse. The experiments described by LeBlanc et al. [1] involved bilateral application of 10 pfu of the neurovirulent McKrae strain of virus to the scarified corneas. We believe that the latter method of inoculation would favor direct uptake of virus from the inoculum into the axons and rapid transfer to the ganglionic neurons. This then would lead to the establishment of unamplified latency within 24 h, as described by Simmons et al. [6]. A significant number of neurons would, therefore, contain latent HSV before the commencement of therapy on postinfection day 1. We would anticipate the effects of therapy under such conditions to be minimal. Furthermore, subtle differences between the 2 drugs may have been obscured. Second, the experiments by LeBlanc et al. [1] resulted in 100% mortality, with deaths occurring on postinoculaton days 4 and 5, suggesting that a severe neurologic infection was established very quickly in this model. Because there were no survivors in the untreated group, there were no positive control animals with which to compare therapy with either drug. By contrast, our experiments showed ∼50% mortality without treatment (which usually occurred 6–10 days after infection). Mortality was reduced to 0 when treatment with either drug was started within 2 days after infection. Under these conditions, survivors were available for comparison with treatment groups [2–5]. Third, LeBlanc et al. state, “In contrast to prior studies that used the ear pinna model [2, 5], we did not observe a rebound of virus titers after cessation of valacyclovir therapy” [1, p. 596–7]. This statement appears to be at variance with the results reported by LeBlanc et al., since positive virus titers in trigeminal ganglia and brains are evident on day 11, the last day on which samples were tested for infectious virus. Therefore, the infection was not cleared by the end of the observation period, and, in any case, the animals were not tested daily, which in our experience is necessary for detection of transient recurrences of infectious virus. We question the statement that no “rebound” of infectious virus was seen, since samples were tested intermittently (days 2, 4, 7, 9, and 11 after infection) and before the acute infection was completely cleared. Indeed, some of our results would look similar to those published by LeBlanc et al. [1] if our data had been recorded only on alternate days. Fourth, LeBlanc et al. state in their introductory text that Thackray et al. [2–5] claim that “famciclovir is better than valacyclovir in terminating ganglionic HSV infection and thereby limits subsequent reactivation” [1, p. 594]. We did not claim that either compound was superior in terminating ganglionic HSV infection; however, we reported a reduction in the amount of detectable latent virus measured by disaggregation, long-term culture of whole ganglia, in situ hybridization for latency-associated transcripts, and reduced rates of reactivation by explant culture [4]. In all cases, the reduction was greater for famciclovir, but neither compound resulted in the “termination” of latency. For these reasons, we contend that the assertion by LeBlanc et al. that their data differ from ours should be judged cautiously. We see little justification for the proposal that the “McKrae-induced infection and disease closely resemble that seen in humans” [1, p. 598]. Indeed, this may not be a good model in which to compare the effects of nucleoside analogues on establishment of latency. We believe that further studies in a variety of laboratory-infection models are urgently required in order to fully interpret clinical data and to optimize treatment strategies using the 2 oral prodrugs, valacyclovir and famciclovir.