Trivalent Glycoprotein Subunit Vaccine Prevents Neonatal Herpes Simplex Virus Mortality and Morbidity

Herpes simplex virus is among the most serious infections of newborns. Current antiviral therapies can prevent mortality if infection is recognized early and treated promptly. Most children who survive nHSV develop lifelong neurological and behavioral deficits, despite aggressive antiviral treatment. We propose that maternal immunization could provide protection against HSV for both mother and baby. To this end, we used a trivalent glycoprotein vaccine candidate to demonstrate that offspring are protected from nHSV following maternal immunization. Significantly, this approach protected offspring from long-term behavioral morbidity. Our results emphasize the importance of providing protective immunity to neonates during this window of vulnerability.

ABSTRACT

Herpes simplex virus (HSV) can cause severe infection in neonates leading to mortality and lifelong morbidity. Prophylactic approaches, such as maternal immunization, could prevent neonatal HSV (nHSV) infection by providing protective immunity and preventing perinatal transmission. We previously showed that maternal immunization with a replication-defective HSV vaccine candidate, dl5-29, leads to transfer of virus-specific antibodies into the neonatal circulation and protects against nHSV neurological sequela and mortality (C. D. Patel, I. M. Backes, S. A. Taylor, Y. Jiang, et al., Sci Transl Med, 11:eaau6039, 2019, https://doi.org/10.1126/scitranslmed.aau6039). In this study, we evaluated the efficacy of maternal immunization with an experimental trivalent (gC2, gD2, and gE2) subunit vaccine to protect against nHSV. Using a murine model of nHSV, we demonstrated that maternal immunization with the trivalent vaccine protected offspring against nHSV-disseminated disease and mortality. In addition, offspring of immunized dams were substantially protected from behavioral pathology following HSV infection. This study supports the idea that maternal immunization is a viable strategy for the prevention of neonatal infections.

IMPORTANCE Herpes simplex virus is among the most serious infections of newborns. Current antiviral therapies can prevent mortality if infection is recognized early and treated promptly. Most children who survive nHSV develop lifelong neurological and behavioral deficits, despite aggressive antiviral treatment. We propose that maternal immunization could provide protection against HSV for both mother and baby. To this end, we used a trivalent glycoprotein vaccine candidate to demonstrate that offspring are protected from nHSV following maternal immunization. Significantly, this approach protected offspring from long-term behavioral morbidity. Our results emphasize the importance of providing protective immunity to neonates during this window of vulnerability.

INTRODUCTION

Neonatal herpes simplex virus (nHSV) is a devastating disease that results in substantial morbidity and mortality, even with antiviral therapy (1). Infection in the neonate can present in three distinct ways: as lesions on the skin, eye, and mouth, as disseminated disease, and/or with central nervous system (CNS) involvement (2). Over 70% of neonates who survive CNS disease have lifelong neurological sequelae (3). Current treatment with antivirals such as acyclovir and other nucleoside analogs is effective in reducing rates of mortality but inadequate in reducing morbidity. In addition, intravenous acyclovir treatment can cause neutropenia in neonates (4). A high degree of clinical suspicion and early diagnosis are critical for successful treatment. This is challenging, as nHSV has a nonspecific presentation and timely diagnosis is often impeded due to initial suspicion of bacterial sepsis (5, 6).

Recent estimates suggest 14,000 cases worldwide annually, or an incidence of ∼1 in 2,000 births, which represents an increase from earlier studies (7, 8). Infection in the neonate is most often perinatally transmitted if the mother has contracted primary genital HSV type 1 or 2 (HSV-1/HSV-2) late during pregnancy (1, 9, 10). Clinical observations have suggested that infants of mothers with preexisting HSV immunity are largely protected from serious infection, likely due to vertical transfer of HSV-specific antibodies (11–13). nHSV transmission can also occur from close contact with caregivers who have an acute HSV infection (14, 15). When active maternal lesions prompt a cesarean delivery, HSV transmission can be prevented, but HSV shedding is often asymptomatic and antiviral treatment is unexploited (16–18). Therefore, new prophylactic avenues must be explored to reduce the global burden of this disease.

Vaccines for women during pregnancy can provide significant protection against infectious diseases for both the mother and newborn (19). In addition to preventing maternal infection, vaccines can also enhance the concentration of maternal antibodies transferred across the placenta and through breastmilk to directly protect infants who are too young to be immunized (20, 21). Influenza and tetanus toxoid, diphtheria toxoid, and acellular pertussis (Tdap) vaccines are routinely recommended for all pregnant women in the United States (19, 22, 23). While there are currently no licensed HSV vaccines, we propose that an HSV vaccine given to mothers could protect their infants from nHSV (24).

We previously demonstrated that maternal immunization with dl5-29, a replication-deficient HSV-2 vaccine candidate, can protect against nHSV morbidity and mortality in a murine model (25). dl5-29 lacks helicase primase and a DNA binding protein and has completed phase 1 clinical trials with Sanofi Pasteur (26). While vaccination is generally considered safe during pregnancy, live virus vaccines are often avoided due to potential risk to the fetus. Subunit vaccines could potentially offer a safer, more clinically feasible approach (27). To this end, our goal was to assess whether an HSV subunit vaccine could also be efficacious in preventing nHSV in a murine model. The trivalent subunit vaccine that was evaluated in this study consists of HSV-2 glycoproteins C, D, and E (gC2, gD2, and gE2) (28). HSV gC binds complement component C3b to inhibit complement activation, gD is essential for viral entry into host cells, and gE binds the IgG Fc domain of antibodies and can block IgG effector function, including complement activation and antibody-dependent cellular cytotoxicity (ADCC) (29–33). Antibodies produced to these antigens following immunization can neutralize virus, block cell-to-cell spread, and prevent immune evasion (34–36). Previous studies have shown that the trivalent subunit vaccine can induce protective antibody and CD4 T-cell responses in rhesus macaques and reduce genital lesions and viral shedding in mice and guinea pigs (28, 37).

The objective of this study was to evaluate whether maternal immunization with the trivalent subunit vaccine could not only protect against mortality but, most importantly, neurological morbidity following nHSV. Neurological impairment resulting from nHSV includes microcephaly, spastic quadriplegia, seizure disorder, blindness, and developmental delay (38). We have previously shown that nHSV morbidity in mice can be measured through behavioral assays such as the open field test (25). Mice that are challenged with HSV as neonates develop increased anxiety-like behavior compared to naive mice. We therefore hypothesized that nHSV-infected offspring of dams immunized with the trivalent vaccine would be protected from this behavioral morbidity. This research supports the notion that maternal immunity to HSV is critical to the outcome of nHSV and provides further evidence for maternal vaccination as a strategy to protect neonates from HSV infection.

RESULTS

Maternal immunization prevents disseminated infection and mortality.

We evaluated the efficacy of a maternal trivalent vaccination strategy by immunizing dams and challenging their offspring with HSV after birth. We immunized female C57BL/6 (B6) mice with 5 μg each of baculovirus-derived gC, gD, and gE subunit proteins with CpG and alum (Alhydrogel) as adjuvants or with CpG and alum alone (mock immunized) (Fig. 1A). Mice were boosted twice biweekly and were then bred. To understand the levels and transfer of vaccine-elicited antibodies, we analyzed sera collected from immunized dams and their offspring (Fig. 1B to D). Sera from offspring of immunized dams contained high titers of HSV glycoprotein-specific antibodies and neutralized HSV-1 and HSV-2 in vitro (Fig. 1E). Intriguingly, the HSV-specific antibody responses of immunized dams persisted to 21 months postimmunization. This was comparable to levels observed early after immunization (3 months), suggesting a remarkable durability of the vaccine response. We then sought to examine if maternal immunization with the trivalent vaccine could protect against neonatal disease. Offspring were challenged with 10⁴ PFU of HSV-1 strain 17 at postpartum days 1 to 2 (P1 to P2). We measured viral titers in various organs 3 days postinfection (dpi) to assess the extent of disseminated disease. Pups from dams that received the trivalent vaccine displayed no viral dissemination with the exception of 2 animals with minimal viral replication (Fig. 2A). However, offspring of mock-immunized dams had significant viral burden, most notably in the trigeminal ganglia (TG), brain, and lung. High viral titers in the lung were expected, given the intranasal route of infection. Pups from mock-immunized animals succumbed to infection by 8 dpi, whereas offspring of trivalent immunized dams were largely and significantly protected from lethal challenge (Fig. 2B). At a lower dose of 10³ PFU, offspring of trivalent immunized dams were not only protected from mortality but also from significant weight loss (Fig. 2C and D). Furthermore, we challenged the offspring of immunized dams with a low-passage clinical HSV-2 isolate (strain G) and found comparable protection (Fig. 2E and F). Thus, maternal immunization with the trivalent subunit vaccine led to the transfer of high-titer neutralizing antibodies and significantly protected offspring from disseminated disease, weight loss, and mortality following HSV challenge.

FIG 1.

Vaccine-elicited antibodies are transferred from immunized dams to offspring. (A) Schematic of the experimental workflow. C57BL6 (B6) female mice were immunized with gC2, gD2, and gE2 with CpG and alum (trivalent) or CpG and alum alone (mock) intramuscularly (i.m.). Mice were boosted twice, 14 days apart, prior to breeding with a naive B6 male. Sera was collected from uninfected pups (postpartum day 3) and immunized dams (3 and 21 months postimmunization) and analyzed for gC2- (B), gD2- (C), and gE2- (D) specific antibody titers by ELISA. Endpoint titers reflect serum dilution. Sera from uninfected pups of immunized dams and naive age-matched animals were assayed for neutralizing antibodies to HSV-1 and HSV-2 via serum neutralization assay (E). Data are represented as individual animals, and statistical significance was determined by unpaired t test (B to D) and two-way analysis of variance (ANOVA) (E). *P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. Time points for sera collection from dams represent separate immunizations.

FIG 2.

Maternal immunization with the trivalent subunit vaccine protects against nHSV. B6 females were immunized and boosted as shown in Fig. 1. Neonates P1 to P2 from trivalent immunized dams (red) and mock-immunized dams (blue) were challenged with HSV. (A) Viral titers in perfused organs at 3 days postinfection (dpi) and (B) survival of neonates challenged with 10⁴ PFU HSV-1. Baseline corrected percent weight change (C and E) and survival (D and F) following a 10³ PFU HSV-1 or HSV-2 challenge. Statistical significance was determined by multiple t test (A and C) or log-rank test (B, D, and F). *, P < 0.05; **, P, < 0.01; ***, P < 0.001; ****, P < 0.0001. Data are representative of at least two independent experiments.

Maternal immunization prevents behavioral morbidity.

Despite antiviral treatment in the clinical setting, infants who survive nHSV infection can be left with debilitating lifelong neurological morbidity (2). Having shown that maternal immunization with the trivalent subunit vaccine prevented disseminated disease and mortality, we wanted to monitor long-term neurological sequela of infection in these offspring. We previously demonstrated that mice challenged with low-dose, nonlethal HSV exhibit anxiety-like behavior (25). When mice are introduced into a novel environment, they normally explore their surroundings. Mice with anxiety-like behavior, however, tend to remain close to the periphery or walls, a behavior known as thigmotaxis (39). Anxiety-like behavior can be measured via the open field test (OFT) and elevated plus maze (EPM) (40, 41). For the OFT, the animal is placed in an arena and allowed to freely move for 10 min while being recorded by an overhead camera. The footage is then analyzed by an automated tracking system to quantify time spent in predefined zones. The EPM maze consists of intersecting open and closed (wall-sheltered) walkways in the shape of a “plus” sign (40). Mice can choose between the open or closed walkways while their movements are recorded. Both of these tests allow measurement of the amount of time that a rodent spends in open or closed areas and generate an index of thigmotaxis. To assess whether offspring of immunized dams exhibit any behavioral morbidity, we challenged pups with 10² PFU of HSV-1 and monitored them until 5 weeks of age. At this low dose, mice of mock-immunized and immunized dams displayed no differences in mortality or weight. However, in the OFT, offspring of mock-immunized dams exhibited a preference (P < 0.05) for the outer perimeter, while offspring of trivalent immunized dams spent similar amounts of time exploring the outer and central areas of the open field (Fig. 3A and B). In order to validate these findings, we tested these mice in the EPM. Although not reaching full statistical significance (P = 0.066), comparable results were found using EPM in that mice from trivalent immunized dams spent more time in the open areas than offspring of mock-immunized animals (Fig. 3C and D). This was despite no differences in ambulatory activity as measured by overall distance traveled (Fig. 3E). Both male and female mice were used for these studies, and we found no differences in behavioral activity between males and females. Interestingly, at this infectious dose, we did not detect latent viral DNA in the brains or trigeminal ganglia of these mice (Fig. 4). As a positive control, DNA extracted from brains and trigeminal ganglia of mice infected with 10³ PFU showed a modest but detectable level of latent viral DNA. This implies that infection with 10² PFU results in establishment of latency that is just under the threshold of DNA detection by quantitative real-time PCR (qPCR). Together, these findings suggest that maternal trivalent immunization can prevent substantial anxiety-like behavior that results from nHSV and that this behavioral phenotype may be independent of latent infection.

FIG 3.

Maternal trivalent immunization can prevent anxiety-like behavior. Mice challenged with 10² PFU of HSV-1 as neonates (P1 to P2) were analyzed in the OFT and EPM. Open field behavior in offspring of mock (A, Left) and trivalent (A, Right) immunized dams is shown by movement tracking. (B) Movements in the OFT can be quantified as thigmotaxis, a ratio of time spent in the outer perimeter relative to total time in the test. Representative EPM behavior in offspring of mock (C, Left) and trivalent (C, Right) immunized dams is shown by heat map. Movements can be quantified as a ratio of the time spent in each location of the maze relative to total time (D) and by total distance moved in the EPM (E). Data are represented as individual animals, and statistical significance was determined by unpaired t test. Error bars represent standard deviation (SD). *, P < 0.05. Data represent two independent experiments.

FIG 4.

Measurement of HSV genome copy number following HSV-1 infection. Viral (thymidine kinase [TK]) and cellular (adipsin) DNA in nanograms per microliter were quantified from brains (A) and trigeminal ganglia (TG) (B) harvested from offspring of mock- and trivalent immunized mice 18 months after neonatal infection with 10² PFU of HSV-1. From these data, HSV genome copy numbers were calculated and are shown for brains (C) and TG (D). As a positive control, tissue from mice infected with 10³ PFU was assessed for HSV DNA at 28 dpi. Standard curves for limit of detection (LOD) (dotted line) were created with naive neural tissue into which known amounts of HSV BAC DNA were added. Data are represented as tissue from individual animals.

DISCUSSION

Globally, more than 67% of the population is infected with HSV-1 (42). Despite this high prevalence, a licensed vaccine does not exist. HSV vaccines have been unsuccessful in preventing adult-to-adult transmission, which was the primary endpoint of previous clinical trials (43, 44). Reducing the risk of HSV transmission to partners is an important goal for a genital herpes vaccine, and such an outcome would also reduce the incidence of nHSV. That said, nHSV has never been included as an outcome measure in clinical trials of HSV vaccines, likely due to its relative infrequency. nHSV is, however, the most life-threatening consequence of HSV-1 and HSV-2 infections (45). Infections range from skin lesions to severe neurologic disease and multiorgan failure, leading to substantial morbidity and mortality (2). The human and economic burden of nHSV is significant (8). While the tragedy of fatal nHSV is apparent, the lifelong neurological consequences frequently observed in survivors have been less well characterized. nHSV is becoming increasingly prevalent in high-income countries as HSV seroprevalence among adults is declining (46). The higher number of seronegative women of reproductive age thereby presents an increased risk of nHSV to the unprotected neonate (47).

We propose that preventing HSV transmission to neonates does not require a vaccine that induces sterilizing immunity. The only licensed herpesvirus vaccine is for varicella-zoster, which diminishes disease but does not induce sterilizing immunity (48, 49). For HSV, seropositive mothers rarely transmit HSV to their offspring, as HSV-specific antibodies are transferred via cord blood to their offspring and are highly neutralizing (25, 50–53). This suggests that higher levels of transferred HSV-specific antibodies may result in greater protection against nHSV (13, 54–57). HSV-specific IgG has also been detected in human fetal neural tissue, implicating a role for maternal antibody-mediated protection of the nervous system even in utero (58). These antibodies could also play an important role in neonates with CNS involvement, in which HSV can often be detected in the cerebrospinal fluid (CSF) (59). In a mouse model, we and others have shown that maternal antibodies from seropositive mice can be transferred to and protect pups from nHSV challenge (58, 60, 61). Heretofore, a maternal immunization approach for nHSV has only been conducted with replication-defective and attenuated vaccine candidates (25, 62, 63). These approaches have been successful in inducing HSV-specific antibodies that are transferred to and protect offspring in rodent models. In this study, we demonstrate that a subunit vaccine can also prevent nHSV mortality and behavioral morbidity.

While all candidate HSV glycoprotein vaccines have been protective in rodent models, these vaccines have not induced long-lasting immune responses in humans (64). Immunization with adjuvanted gD subunit protein protected seronegative women from HSV-1 but not HSV-2 (65–67). Importantly, immunization did not have any adverse effects on pregnancy, but unfortunately, postnatal HSV status was not reported (68). These results spurred the development of vaccines that combine other HSV-2 glycoproteins with gD (36, 69, 70). Beyond providing additional immunogens, this approach would be expected to induce antibodies that could block viral immune evasion. Studies have demonstrated that the addition of gC and gE to gD significantly outperforms gD alone in preventing genital lesions following infection in rodent models (28, 36). In rhesus macaques, this trivalent vaccine also protected against vaginal infection and pathology (37).

In this study, we challenged neonatal pups intranasally following maternal vaccination with trivalent glycoprotein subunits. Our previous study utilized this approach with a replication-defective HSV-2 vaccine candidate, dl5-29 (25). Intriguingly, the trivalent subunit vaccine may offer especially strong protection in the lung during disseminated neonatal infection (Fig. 2A) relative to dl5-29 (25). A focus of future studies will be to further and directly compare these vaccine candidates. Ultimately, the trivalent and dl5-29 vaccines were both protective when pups were challenged with HSV at postpartum days 1 to 2. This differs from a ΔgD vaccine candidate which was not as effective in preventing neonatal mortality unless the challenge was at postpartum day 7 or later (62). The ΔgD vaccine does not induce strong neutralizing titers but does induce robust ADCC responses. These are perhaps more beneficial once neonatal immune cells have become further developed (62). Thus, a possible explanation is that high neutralizing antibody titers are critical for protecting newborns whose immune systems are still developing, but further studies are needed to address correlates of protection (56, 57, 71). Our study also showed that maternal trivalent immunization protected offspring from long-term behavioral consequences. Behavioral pathology following 10² PFU infection does not seem to result from a robust establishment of latency, as we were unable to detect HSV DNA in the brain (<1,000 copies/brain) or trigeminal ganglia (<100 copies/TG). Both the OFT and EPM measure anxiety-like behavior in rodents and have been validated in humans (72, 73). These results suggest that anxiety could manifest from nHSV infection and warrants longitudinal studies in surviving infants.

Despite the multiple failures in the HSV vaccine field, we propose that maternal vaccination aimed at preventing neonatal herpes is feasible and necessary. It is also possible that the seronegative women of childbearing age who were enrolled in these clinical trials became protective of their offspring as a result of the vaccine. This, however, was not examined, and perhaps the repurposing of failed HSV vaccine candidates for prevention of nHSV should be explored. A maternal vaccine would not only stimulate seroconversion of HSV-seronegative women but could also boost the HSV-specific antibody titers of seropositive expectant mothers. A vaccine booster dose could also be important for subsequent pregnancies. While this study adds to the growing body of literature supporting maternal immunization for HSV, other promising therapeutics, including nucleotide and recombinant vaccines and monoclonal antibodies, should be assessed for efficacy in preventing nHSV (74–81). Beyond HSV, vertical transmission of toxoplasma, Zika, rubella, and cytomegalovirus can cause severe neurological sequela (82, 83). While there are several challenges in developing vaccines for pregnant women and neonates, vaccination of women of reproductive age should continue to be explored as an approach to protect these vulnerable populations (84–86). Importantly, maternal immunization could prevent long-term neurological pathologies associated with neonatal infection.

MATERIALS AND METHODS

Cells and viruses.

The strains used in this study were HSV-1 strain 17syn+ (87), HSV-1 17eGFP, HSV-2 333 ZAG, HSV-2 G (88). HSV-2 333 ZAG expresses green fluorescent protein (GFP) under the control of a cytomegalovirus (CMV) promoter inserted in an intergenic region between UL3 and UL4 (gift from B. Herold with permission from P. Spear). HSV-1 17eGFP was created by homologous recombination using the plasmid pKUL37eGFP, which contains enhanced GFP (eGFP)-tagged strain KOS UL37 (89). This was linearized with restriction endonuclease (EcoRI) and cotransfected with HSV-1 strain 17 genomic DNA into Vero cells. The virus was purified by the selection of green fluorescent plaques. Virus stock preparation and the plaque assay were performed using Vero cells as described previously (90, 91).

Mice and animal procedures.

All procedures were performed in accordance with federal and university policies. C57BL/6 (B6) mice were purchased from the Jackson Laboratory and bred in the barrier facility in the Center for Comparative Medicine and Research at the Geisel School of Medicine at Dartmouth.

Immunization.

For intramuscular immunization, 5 μg of gC2, gD2, or gE2 baculovirus subunit proteins (provided by G. Cohen) were individually incubated with 16.7 μg of CpG oligonucleotide 5′-TCCATGACGTTCCTGACGTT-3′ (Coley Pharmaceutical) and 125 μg of alum (Alhydrogel; Accurate Chemical and Scientific Corp.) at room temperature for 2 h. The three immunogens were combined before immunization in a total volume of 50 μl. Mock immunizations were performed with CpG and alum in saline without proteins. Six- to 8-week-old female B6 mice were intramuscularly immunized 3 times, 14 days apart (Fig. 1A), while mice were under isoflurane anesthesia. Following the third immunization, mice were bred.

Sera analyses.

For sera preparation, mock- or trivalent immunized dams were anesthetized with isoflurane, and the submandibular vein was punctured with a 5-mm lancet (Medipoint). Blood was collected from uninfected postpartum day 3 pups by decapitation. Blood was clotted by stasis for 15 to 30 min at room temperature and then centrifuged at 2,000 × g for 10 min at 4°C. Serum neutralization assays were conducted by heat-inactivating serum samples at 56°C for 30 min and incubating serial dilutions with 50 PFU of HSV-1 17eGFP or HSV-2 333 ZAG for 1 h at 37°C. Samples were then incubated on a monolayer of Vero cells for 48 h and assessed for cytopathic effect (CPE) using GFP signaling from the reporter viruses. The neutralization titer was determined by the mean highest dilution without CPE. Enzyme-linked immunosorbent assays (ELISAs) for anti-gC2, anti-gD2, and anti-gE2 were conducted as previously described (28). Briefly, ELISA plates were coated individually with 100 ng of gC2, gD2, or gE2 subunit antigen and incubated with serial dilutions (1:250 to 1:256,000) of sera from dams and pups, followed by horseradish peroxidase (HRP)-conjugated anti-mouse IgG at a dilution of 1:2,000. The endpoint titer was defined as the highest dilution at which the optical density (OD) is greater than 0.1 and twice the value from naive serum.

Viral challenge.

Neonatal mice postpartum days 1 to 2 were infected intranasally (i.n.) with 10² to 10⁴ PFU of HSV-1 strain 17 or HSV-2 G in a volume of 5 μl under isoflurane anesthesia. For viral titers of organs, tissue was harvested 3 dpi after cardiac perfusion with at least 5 ml of cold phosphate-buffered saline (PBS) per animal. All organs were collected in 1.5-ml tubes containing ∼100 μl of 1-mm-diameter glass beads and 1 ml of Dulbecco’s modified Eagle’s medium (DMEM) containing 1% fetal bovine serum (FBS), 1% penicillin-streptomycin, and 1% amphotericin B. Sample homogenates were prepared by mechanical disruption in Mini-Beadbeater-8 (BioSpec Products) and then sonicated. Titers of homogenates were determined by plaque assay on Vero cells.

Behavioral analysis.

Breeding and infection were conducted in a room that was separate from the behavioral studies room. All animals were housed in the behavioral room for at least 1 week prior to testing. Environmental conditions (test room lighting, temperature, and noise levels) were kept consistent, and both male and female mice were included in the study. Animal movements were recorded (Canon Vixia HFM52), and videos were coded and analyzed with open-source software that had been manually validated (92). The open field test (OFT) was conducted as previously described (25). Briefly, 5- to 7-week-old B6 mice were placed into the open field arena (30 cm by 30 cm) and allowed to habituate for 10 min followed by recording for 10 min. Animals were tested in the elevated plus maze (graciously provided by V. Galton and the late D. Bucci) at least 2 weeks after the OFT. Mice were placed on an open arm facing away from the center, and their movements were recorded for 5 min.

Quantitative real-time PCR.

Brains and trigeminal ganglia (TG) were harvested from 10² PFU infected mice 18 months after infection and extracted using Qiagen DNeasy blood and tissue kit. Amounts of viral (thymidine kinase) and cellular (adipsin) DNA in samples were quantified by real-time PCR as previously described (93). The quantification standards of viral genomes were prepared by reconstituting known amounts (copies) of HSV-1 bacterial artificial chromosome (BAC) DNA with homogenates of tissues from uninfected mice. The limit of detection for this assay was 11 copies of HSV DNA per reaction. Known amounts of tissue DNA prepared from mouse neural tissues were used as the quantification standards for cellular DNA. As a positive control for detecting HSV genomes in latently infected mice, DNA was also extracted from the brains and TG of mice infected with 10³ PFU and analyzed in parallel.