Histone Deacetylase Inhibitors Induce Reactivation of Herpes Simplex Virus Type 1 in a Latency-Associated Transcript (LAT)-Independent Manner in Neuronal Cells

Abstract

Histone acetylation is implicated in the regulation of herpes simplex virus type 1 (HSV-1) latency. However, the role of histone acetylation in HSV-1 reactivation is less clear. In this study, the well established model system, quiescently-infected, neuronally-differentiated PC12 (QIF-PC12) cells, was used to address the participation of histone acetylation in HSV-1 reactivation. In this model, sodium butyrate and trichostatin A (TSA), two histone deacetylase inhibitors, stimulated production of infectious HSV-1 progeny from a quiescent state. To identify viral genes responsive to TSA, we analyzed representative α, β, and γ viral genes using quantitative real-time polymerase chain reaction. Only the latency-associated transcript (LAT) accumulated in response to TSA treatment, under culture conditions that restricted virus replication and spread. This led us to evaluate the importance of LAT expression on TSA-induced reactivation. In QIF-PC12 cells, the LAT deletion mutant virus dLAT2903 reactivated equivalently with its wild type parental strain (McKrae) after TSA treatment, as well as forskolin and heat stress treatment. Both viruses also reactivated equivalently from latently infected trigeminal ganglia explants from rabbits. In contrast, there was a marked reduction in the recovery of dLAT2903, as compared to wild type virus, from the eyes of latently infected rabbits following epinephrine iontophoresis. These combined in vitro, ex vivo and in vivo data suggest that LAT is not required for reactivation from latently infected neuronal cells per se, but may enhance processes that allow for the arrival of virus at, or close to, the site of original inoculation (i.e., recrudescence).

Introduction

Herpes simplex virus type 1 (HSV-1) is a large alpha herpesvirus capable of two types of infection. The initial lytic infection occurs in epithelium and results in the ultimate destruction of infected cells. Invasion of local peripheral nerve endings and retrograde axonal transport (Johnson, 1964) produces an infection of regional sensory ganglionic neurons (Hill et al 1972; Stevens and Cook, 1971). Neurons that survive retain the virus in a latent state. During latency, HSV-1 DNA forms an episome (Rock and Fraser, 1983) and the majority of the genome is transcriptionally repressed (reviewed in Preston, 2000). The region(s) responsible for regulating viral latency and periodic reactivation of HSV-1 remain largely unknown.

Evidence is accumulating that histone acetylation and derepression of viral gene expression may contribute to the regulation of the lytic and latent life cycles of HSV-1 (Hobbs and DeLuca, 1999; Poon et al, 2003; Kubat et al, 2004). Histones regulate gene activity via modulation of their interactions with DNA (Grunstein, 1997). Histones are post-translationally modified by acetylation, phosphorylation, and ubiquitination (Hong et al, 1993). Acetylation of the histone tail regions by acetyltransferases results in altered chromatin structure, making the nucleosome accessible to the RNA polymerase II complex and transcription initiation (Galasinski et al, 2000; Grunstein, 1997; Reeves, 1984). Recent evidence indicates that when expression/processing of the infected cell polypeptide ICP0 is perturbed and HSV-1 DNA has the opportunity to be silent in the nucleus, the viral DNA is in chromatin-like structures amenable to modification by histone acetylation (Poon et al 2003). Further, inhibition of histone deacetylases by trichostatin A (TSA) results in activation of the HSV-1 ICP0 promoter in neonatal dorsal root neurons in vitro (Arthur et al, 2001). However, whether infectious progeny result from latently infected neurons after TSA treatment is not known.

The latency-associated transcript (LAT) promoter of latent HSV-1 DNA has been found to be associated with acetylated histone H3 (Kubat et al, 2004). This suggests that the LAT promoter is euchromatic and a transcriptionally active structure. However, it is not known if LAT responds to histone deacetylation inhibition. In addition, the role of LAT during the induction phase of reactivation within the neuron after TSA treatment is not known.

In the neurally-derived PC12 cell line, several HSV-1 promoters are responsive to the histone deacetylase inhibitor, sodium butyrate (Frazier et al, 1996). Since histone deacetylase inhibitors enhance HSV gene expression in cells displaying a neuronal phenotype, it was important to determine if these agents could activate a cryptic HSV-1 genome and produce viral progeny in neuronally-differentiated PC12 cells. Rat pheochromocytoma-derived PC12 cells, following neuronal differentiation, support HSV-1 in a quiescent state that is reminiscent of latency (Danaher et al, 1999a; Su et al, 1999). These cells are advantageous in allowing for the establishment of quiescence (i.e., infectious virus not detected in culture supernatants and lysates) with wild-type HSV strains (Danaher et al, 1999a), and the maintenance of quiescence without the requirement of antiviral agents (Danaher et al, 1999a; 1999b). Moreover, reactivation of HSV from quiescently infected, neuronally-differentiated (QIF)-PC12 cells occurs following exposure to diverse stimuli such as physical (heat), chemical (forskolin), and biological (pituitary adenylate cyclase activating peptide) agents (Danaher et al, 1999a; 1999b; 2001; 2003). In addition, the proportion of QIF-PC12 cells that support reactivation is similar to that observed in vivo.

In this report, we used in vitro (i.e., QIF-PC12 cells), ex vivo, and in vivo models of latency, the LAT deletion mutant dLAT2903, and inhibitors of histone deacetylation to further evaluate the participation of histone deacetylation and LAT in virus reactivation and recrudescence. Herein we report that 1) treatment of HSV-1 QIF-PC12 cells with inhibitors of histone deacetylases results in activation of the cryptic viral genome as indicated by the production of infectious progeny, 2) the major LAT accumulates in QIF-PC12 cells following TSA treatment, and 3) whereas the absence of LAT reduced measurable in vivo reactivation at the eye after epinephrine iontophoresis, LAT was not required for efficient reactivation of HSV-1 from explanted latently-infected trigeminal ganglia or from QIF-PC12 cells following induction with TSA and other known in vitro reactivation stimuli. These studies indicate that LAT is dispensable for reactivation, but enhances processes that allow for the arrival of virus at, or close to, the site of original inoculation (i.e., recrudescence).

Results

Histone deacetylase inhibitors induce HSV-1 reactivation from quiescence

The fact that the HSV-1 genome is maintained in association with nucleosomes during latency (Deshmane and Fraser, 1989), and the ICP0 promoter appears responsive to histone deacetylase inhibition in replication-defective HSV recombinants (Arthur et al, 2001), led us to determine whether inhibition of histone deacetylases can induce reactivation of replication-competent HSV-1 from a quiescent state. QIF-PC12 cells were used to examine the effect of histone deacetylase inhibitors on viral genomes harbored within neuronal cells per se, i.e. without the indirect effects due to glial cells, cells of the immune system, and other potential in vivo participants. QIF-PC12 cell cultures were established with HSV-1 strain 17⁺ and treated with sodium butyrate. Supernatants were monitored daily for the production of virus over the next 8 days. Treatment with sodium butyrate resulted in dramatic reactivation of HSV-1 (Fig. 1A). In contrast, virus did not reactivate significantly above background following treatment with a broad range of concentrations of 5′-azacytidine, an inhibitor of DNA methylation (data not shown). Sodium butyrate inhibits histone deacetylases (Candido et al, 1978), but also has pleiotropic activities (Birren et al, 1987; Kruh, 1982). To further support the involvement of inhibition of histone deacetylation in HSV-1 reactivation, QIF-PC12 cells were treated with TSA, a specific inhibitor of histone deacetylase (Van Lint et al, 1996; Yoshida et al, 1990). TSA treatment resulted in dose-dependent reactivation of the cryptic viral genome, with a maximal response observed at the highest concentration tested (400 ng/ml) (Fig. 1B). These findings suggest that inhibition of histone deacetylation results in altered cellular and/or viral gene expression that favors the production of infectious progeny from a cryptic HSV-1 genome.

Fig. 1.

Effect of sodium butyrate (NaB) and TSA on HSV-1 strain 17⁺ reactivation from QIF-PC12 cells. Cells, grown in duplicate 12-well plates per treatment group, were neuronally differentiated with nerve growth factor (NGF) then infected with HSV-1 strain 17⁺ at an MOI of 5 in the presence of acycloguanosine (ACV). Seven days after the withdrawal of ACV from the medium (day 17 postinfection), culture medium was changed to medium containing (A) NaB (solid bars), forskolin (striped bar), or no drug (clear bar), or (B) TSA at the indicated concentrations, as well as no treatment, control medium. Reactivation was monitored for 8 days using culture supernatants titered onto Vero cells. Histogram shows results of cumulative wells positive by day 8 post-induction. * P < 0.05 (mock vs. treatment with agent). For all treatment points, the standard deviations were < 10%. Similar results were obtained in duplicate NaB and TSA experiments.

Induction pattern of α (immediate-early), β (early), and γ (late) gene transcription following TSA treatment

Based on the histone deacetylase inhibitor findings, it was important to identify viral genes responsive to TSA treatment. To investigate this, RNA isolated from QIF-PC12 cells established with strain 17⁺ was reverse transcribed, and the quantity of cDNA synthesized for select host and viral transcripts following TSA induction was determined by real-time PCR. The primers and probes used for the analyses of representative α, β and γ genes and the cellular control GAPDH are indicated in Table 1 and Figures 2A and 2B. In this experiment, the culture medium contained ACV from the day of infection until the day of harvest to maintain quiescence, limit viral gene expression dependent on DNA synthesis, and focus on those initial events that occur following induction without the complication of measuring subsequent events due to a spreading infection. In parallel, QIF-PC12 cells maintained in 12-well plates without ACV were analyzed and shown to reactivate HSV-1 as a control (Fig. 3A).

Table 1.

Real Time PCR Primers and Probes.

Primer/Probe	Sequence	Location	Source
ICP27 F	CGCCAAGAAAATTTCATCGA G	114,766 → 114,786	Cohrs et al 2000
ICP27 R	ACATCTTGCACCACGCCAG	114,811 ← 114,829
ICP27 Probe	CTGGCCTCCGCCGACGAGAC	114,790 → 114,809
Major LAT F	CCCACGTACTCCAAGAAGGC	119,721 → 119,740 6,631 ← 6,650	Cohrs et al 2000
Major LAT R	AGACCCAAGCATAGAGAGCC AG	6,579 → 6,600 119,771 ← 119,792
Major LAT Probe	CCCACCCCGCCTGTGTTTTTG TG	119,747 → 119,769 6,602 ← 6,624
α4 F	CCGTCCCTGTCCTTTTTCC	146,926 → 146,944 131,289 ← 131,307	*
α4 R	GTAGGTCACCTACGGACTCT CGAT	131,183 → 131,206 147,027 ← 147,050
α4 Probe	CCGCGCTAGTTCCGCGTCGA	146,964 → 146,983 131,250 ← 131,269
Minor LAT F	CGCATGCGCTGTGGTTT	119,287 → 119,303 7,068 ← 7,084	*
Minor LATR	AGAACAGGAAAGGCGATGG A	7,023 → 7,042 119,329 ← 119,348
Minor LAT Probe	CCCGGCAGAACACCGAGGAA AAA	7,044 → 7,066 119,305 ← 119,327
VP16 F	TCGACGACTTGGGCTTTAGC	104,864 ← 104,883	*
VP16 R	GAAAACAGATCCTCGTTCCA GGTA	104,811 → 104,834
VP16 Probe	CCCCGCGCTATGTACCATGC TCG	104,836 ← 104,858
UL39 F	ATAGCCAATCCATGACCCTG TATG	89,661 → 89,684	Cohrs et al 2000
UL39 R	GGGTGGAGGCTGGGAGG	89,707 ← 89,723
UL39 Probe	CACGGAGAAGGCGGACGGG A	89,686 → 89,705
Spliced α0 F	CCCTCCAGCCGCATACG	2,238 → 2,254 124,117 ← 124,133	*
Spliced α0 R	CCTCAGAGTCGCTGCTGTCC	123,225 → 123,244 3,127 ← 3,146
Spliced α0 Probe	CAGCGCGAGCCCGCCC	Probe spans intron
GAPDHS	GAACATCATCCCTGCATCCA		Medhurst et al 2000
GAPDHA	CCAGTGAGCTTCCCGTTCA
GAPDH Probe	CTTGCCCACAGCCTTGGCAG C

^*Designed using Primer Express software (PE Applied Biosystems).

Fig. 2.

Location of primers used to assess LAT cDNA levels. (A) Top line represents the region spanning from nucleotide position 117,000 to 124,000 of the prototype viral genome. Locations of the α0 transcripts and 8.3-kb minor LAT and 2.0-kb major LAT are displayed. (B) Nucleotide positions of regions assessed by real-time PCR represent the flanking boundaries of PCR products. The expression of LATs was assessed at three locations. The minor 8.3-kb LAT was assessed at the region 5′ of the major LAT to evaluate changes in expression from the LAP1 promoter (Dobson et al, 1989; Zwaagstra et al, 1990), and also 3′ of the major 2.0-kb LAT to include changes occurring from the LAP2 promoter (Chen et al, 1995; Goins et al, 1994; Nicosia et al, 1993; Wechsler et al, 1988). The third site was within the 2.0-kb major LAT (119,721 to 117,792). Note that primers used to assess levels of α0 cDNA span the first intron of the α0 gene transcript. Transcription from LAP1 initiates at nucleotide 118,801.

Figure 3.

HSV-1 strain 17⁺ reactivates after TSA induction, and expression pattern of viral genes following TSA-induced reactivation. (A) QIF-PC12 cells were established and maintained in 12-well plates and assayed for reactivation as described in Figure 1, as a positive control for experiment shown in B. Ten days after ACV was removed, cultures were induced with TSA (400 ng/ml), heat stress (HS, 43°C x 3 h, positive control), or mock medium on day 17 postinfection. (B) QIF-PC12 cells, maintained in medium containing ACV, in parallel 6-well plates, were harvested from duplicate wells and RNA isolated at the times indicated following mock or TSA treatment. RNA was reverse transcribed and the quantity of cDNA synthesized for selected host and viral transcripts was determined by real-time PCR. The average numbers of cDNA copies were determined from triplicate reactions and are presented relative to the levels detected 6 h after mock treatment. The 6 h mock treatment was set at 1 for each transcript. Sufficient amounts of cDNA were used in all PCRs such that all representatives of α, β, and γ class genes examined were readily detected. Signal was rarely detected in controls lacking Rtase (-Rtase). When detected in –Rtase controls, the level was substantially below the levels detected in experimental samples. * P < 0.05 (LAT vs. mock treatment). For all cDNA results, the mean standard deviations were < 17%.

Figure 3B shows that the level of the cellular GAPDH transcript remained constant at 6, 12, and 24 h post-mock treatment and 6 and 12 h post-TSA treatment, then decreased by approximately 40 to 50% by 24 and 48 h post-TSA treatment. During this time frame of quiescence, expression of all viral transcripts examined was detected and decreased to varying degrees after mock treatment. Following TSA treatment on day 17 postinfection, the amount of the major LAT more than doubled by 24 h and increased nearly 5-fold by 48 h post treatment (P < 0.05 compared with mock treatment). Similar increases were observed in the minor LAT in most assays (data not shown). In contrast, increases were not detected in any of the representative α, β or γ genes.

LAT is not required for reactivation of a cryptic HSV-1 genome in neuronal cells

Because LAT accumulated in response to TSA treatment and the role of LAT during the induction phase of reactivation has not been characterized, we investigated whether LAT plays a direct role in reactivation. A correlative approach using in vitro, ex vivo, and in vivo models of latency was employed (Miller et al, 2003). This strategy allowed us to observe the role of LAT during HSV-1 reactivation from latency at the cellular and tissue levels and in the animal.

In the in vitro analysis, neuronally-differentiated PC12 cultures were infected with the HSV-1 LAT deletion mutant dLAT2903 or its parent strain McKrae at a multiplicity of infection (MOI) of 5 as described in the Materials and Methods. The quiescent phase was established and maintained in 99% of cultures until induction on day 17 pi. Activation of HSV-1 from QIF-PC12 cells was measured using daily aliquots of culture supernatants that were titered onto Vero cells. In replicate experiments, HSV-1 reactivated from dLAT2903-infected cultures after induction with TSA or forskolin (Fig. 4). For each treatment, the level and kinetics of reactivation from dLAT2903-infected QIF-PC12 cells was equivalent with that of cells infected with its parent McKrae strain. Similar results were observed following heat stress (data not shown).

Fig. 4.

Reactivation of dLAT2903 and McKrae from QIF-PC12 cells is similar. Cells were grown in duplicate 12-well plates per treatment group, neuronally differentiated with NGF, and infected with the indicated strain at an MOI of 5. On day 17 postinfection, cultures were exposed to maintenance medium (mock), medium containing TSA (400 ng/ml) or forskolin (50 μM). Reactivation was monitored daily for 8 days using culture supernatants titered onto Vero cells. Data represents means from at least two independent experiments. Spontaneous reactivation from mock induced cultures was less than 12%. The mean standard deviations for each time point were less than 15%. Similar results were observed when TSA treatment was performed on day 31 postinfection.

To determine if viral genome copy number contributed to the observed response, the amount of viral DNA present in QIF-PC12 cells was measured from triplicate cultures by quantitative real-time PCR. The viral DNA copy was virtually equivalent for cultures infected with the two strains (McKrae = 3.85 ± 0.12 x 10⁶ genome equivalents vs. dLAT2903 = 4.23 ± 0.4 x 10⁶ genome equivalents per 100 ng PC12 DNA).

The explant and in vivo studies were performed by infecting the eyes of New Zealand white rabbits with HSV-1 McKrae, dLAT2903, and its rescue (dLAT2903R). These strains were selected because the absence of LAT in dLAT2903 has been well documented to reduce reactivation efficiency in vivo (Perng et al, 1994). The acute infection was monitored by slit lamp examination (SLE). No differences in the severity of corneal lesions were observed among infections by the three strains. After the ocular infections resolved, latency was established after 4 weeks post-infection in rabbit trigeminal ganglia (TG). With this model, we evaluated reactivation in vivo using detection of ocular shedding of HSV-1 induced by transcorneal adrenergic iontophoresis, and ex vivo using TGs that were dissected and cultured as explants in medium containing serum. Aliquots of supernatant explant medium were removed daily and tested for cytopathic effects on primary rabbit kidney (PRK) cells beginning on day 7 post-explant (PE) through day 30 PE. In the ex vivo model, dLAT2903 reactivated equivalently and within a similar time frame as the McKrae and rescue strain (Table 2). The findings indicate that LAT is not required for efficient reactivation of these strains ex vivo.

Table 2.

Reactivation of HSV-1 from explanted, latently-infected rabbit trigeminal ganglia.

Virus	Time to Positive for Reactivation in Days (Range)	Explants Positive for Reactivationa/Total Explants (%)	P valuesb
McKrae	15.50 (10–21)	7/8 (88%)
dLAT2903	14.25 (9–20)	13/15 (87%)	0.9549
dLAT2903R	15.33 (9–21)	10/11 (91%)	0.8111

^aReactivation determined as described in the Material and methods.

^bCompared with McKrae.

In the in vivo latency model, which served as a control, reactivation induced by epinephrine iontophoresis was high in rabbits latently infected with McKrae and dLAT2903R (Table 3). In contrast, reactivation of the LAT deletion mutant dLAT2903 was significantly less as determined by the frequency of eyes and tear-film swabs positive for virus (i.e., recrudescence). To determine whether viral DNA copy number contributed to the observed outcome, the amount of HSV-1 DNA per TG was quantified by real-time PCR. The copy number per TG (9 or 10 TGs per virus group) for the LAT null virus dLAT2903 (2095 ± 545 [SEM]) was virtually equivalent to that of McKrae (1920 ± 517) and dLAT2903R (1808 ± 517).

Table 3.

Ocular reactivation in HSV-1 latently infected rabbits.

Virus	Rabbits Positive for Reactivationa/Total Rabbits (%)	Eyes Positive for Reactivation/Total Eyes (%)	Eye Swabs Positive for Reactivation/Total Swabs (%)	P valuesb
McKrae	8/8 (100%)	15/16 (94%)	47/112 (42%)
dLAT2903	2/8 (25%)	2/16 (13%)	6/112 (5%)	<0.0001
dLAT2903R	7/8 (88%)	14/16 (88%)	41/112 (37%)	0.4941

^aDetermined by recrudescence in the eye.

^bP values are a statistical analysis comparing McKrae to dLAT2903 and dLAT2903R for swabs positive per total swabs.

Collectively, these data indicate that LAT is not required for efficient reactivation from the neuronal cells or TG explants. However, LAT appears to contribute to our ability to detect reactivation in vivo when measured as recrudescence at the eye. In addition, the observed difference in reactivation efficiency of the two strains in vivo was not attributable to different viral genome copy number in cells harboring cryptic HSV-1 genomes.

Discussion

In the present study, we used QIF-PC12 cells, TG explants, and the rabbit eye model of latency to examine the effect of histone deacetylase inhibitors and LAT expression on HSV-1 reactivation. The QIF-PC12 model was used because it provides a useful system similar to latency in vivo (Miller et al, 2003; Millhouse and Wigdahl, 2000), with the distinct advantages that all cells have a neuronal phenotype and can be infected with wild-type HSV-1 strains, and the predictable disruption of quiescence and the resulting activation of virus production following exposure to physical (heat), chemical (forskolin), and biological (pituitary adenylate cyclase activating peptide) stimuli (Danaher et al, 1999a; Danaher et al, 1999b; Danaher et al, 2001). This culture system allows one to analyze events that regulate the transition of a cryptic genome to an active genome without the complications of immunological processes that can limit one’s ability to distinguish between reactivation (a neuronal event) and recrudescence (the subsequent epithelial event) in vivo.

In the current study, treatment of QIF-PC12 cells with inhibitors of histone deacetylases resulted in the production of infectious HSV-1 progeny. Our findings regarding sodium butyrate and TSA are consistent with observations on herpesvirus DNA (Poon et al, 2003; Kubat et al, 2004) and the role of histone deacetylases in the regulation of gene expression (Van Lint et al, 1996). During latency, HSV-1 DNA has been detected in association with nucleosome-like structures (Deshmane and Fraser, 1989). Thus, a post-translational change in the histone state could alter the expression of herpesvirus genes. Arthur et al 2001 found that cytomegalovirus IE1 and HSV-1 ICP0 promoters are responsive to TSA during quiescence in neonatal dorsal root neurons. A product of the ICP0 gene appears to be operationally similar to inhibitors of histone deacetylases (Hobbs and DeLuca, 1999; Poon et al, 2003), and the LAT promoter has been shown to be associated with hyperacetylated H3 histones during latency (Kubat et al, 2004). In human γ-herpesviruses, a correlation between repression by histone deacetylases and maintenance of latency and, conversely, histone acetylation and viral reactivation has been demonstrated (Bryant and Farrell, 2002; Gwack et al, 2001; Jenkins et al, 2000). Thus, it is plausible that the HSV-1 genome is silenced during latency by a mechanism involving histones, and those processes that allow for acetylation of histones result in derepression and transcription of host and/or viral genes required of reactivation (Hsia and Shi, 2002; Poon et al, 2003). It is important to note that regulation of viral quiescence does not appear to involve demethylation, as the HSV-1 genome is reported to be unmethylated (Dressler et al, 1987; Low et al, 1969; Lundberg et al, 2003; Kubat et al, 2004) and virus production was not induced from QIF-PC12 cells following 5′-azacytidine treatment, over a broad range of concentrations tested (data not shown).

The accumulation of the major LAT following TSA induction, in the absence of obvious increased expression from representative genes of α (ICP0, 4, 27), β (U_L39), and γ (VP16, a γ₁ gene) classes, was unexpected. HSV-1 immediate early promoters are responsive to sodium butyrate or TSA in neonatal neurons (Arthur et al, 2001) and PC12 cells (Frazier et al, 1996), and infectious progeny are generally detectable in QIF-PC12 cultures by day 3 post-TSA treatment. Thus, if viral transcription levels increased, 48 h would have been sufficient time to detect such increases. One possible explanation is that viral transcription initiates in only a few cells in the latently infected population (Arthur et al, 2001; Danaher et al, 1999a; Sawtell and Thompson, 1992), and when assessed among the majority of cells that do not permit reactivation, the increase in transcription may be masked. The presence of ACV may also have restricted the occurrence of an event (i.e., DNA replication) critical for increased viral gene transcription in neurons (Kosz-Vnenchak et al, 1993; Nichol et al, 1996). Alternatively, the changes induced by TSA could act at a post-transcriptional level or on cellular genes important for reactivation.

Although LAT accumulates in QIF-PC12 cells after TSA treatment, the biological significance of this event is less clear. LAT is not required for HSV-1 reactivation in vitro (Fig. 4). Similarly, LAT is not required for reactivation from latently-infected ganglion explants of rabbits (Table 2) or BALB/c mice (Perng et al, 2001), or from outbred Swiss Webster mice following establishment with wild-type levels of viral latency (Thompson and Sawtell, 1997). In contrast, LAT is important for efficient reactivation in vivo (Bloom et al, 1997; Hill et al, 1997; Hill et al, 1990; Thompson and Sawtell, 1997; Trousdale et al, 1991). Evidence supports the fact that in vivo neuroprotection (Thompson and Sawtell, 2001) and down regulation of productive viral gene expression (Chen et al, 1997; Garber, Schaffer, and Knipe, 1997) could be involved. TSA induces neuronal apoptosis in vitro (Boutillier et al, 2002; Boutillier et al, 2003). Thus, LAT could be a viral response that attempts to protect the neuron during an event that induces apoptosis and/or reactivation (Gill and Windebank, 1998; Perng et al, 2000; Perng et al, 2001; Thompson and Sawtell, 2001). If LAT is serving a protective role during reactivation, a logical explanation would be the accumulation of LAT in cells that become protected and do not support reactivation (i.e., remain latently infected), whereas LAT would fail to accumulate to appreciable levels in cells supporting reactivation (Colgin et al, 2001; Halford et al, 1996). Because our analysis involved RT-PCR, it is not certain if the high levels of LAT are accumulating in the select few cells that reactivate virus, or lower levels are produced in the majority of cells that do not.

To our knowledge, this study provides the first evidence that treatment of neuronal cells harboring cryptic HSV-1 genomes with histone deacetylase inhibitors results in the production of infectious progeny. Interestingly, although LAT accumulates in response to TSA treatment and the LAT-ICP0 locus appears important in the histone regulatory state of latency, the absence of LAT did not affect the ability of the virus to reactivate in neuronal cells. Thus, LAT is dispensable for reactivation, but appears to enhance processes subsequent to reactivation that allow for the arrival of virus at, or close to, the site of original inoculation (i.e., recrudescence).

Material and methods

Viruses and cells

HSV-1 strain 17⁺ was obtained from N. Fraser of the Wistar Institute, Philadelphia, PA. The McKrae, dLAT2903, and dLAT2903-R strains were obtained from S. Wechsler, UCLA School of Medicine, Los Angeles, CA. Viral stocks were prepared in PRK cells for in vivo studies and in Vero cells for in vitro studies and maintained at −85°C. Rat pheochromocytoma (PC12) and African green monkey kidney (Vero cells) were grown as previously described (Danaher et al, 1999a). PRK cells were propagated and maintained at 37ºC in 2% CO₂ as monolayers as previously described (Miller et al, 2003). All cells were obtained from American Type Culture Collection (Rockville, MD).

Neuronal differentiation

PC12 cells were dissociated by passage through a 22-gauge needle and plated in RPMI 1640 containing 0.1% fraction V bovine serum albumin (BSA) in 12-well tissue culture dishes coated with rat tail collagen type 1 (Becton Dickinson, Franklin Lakes, NJ) at 2.2 X 10⁵ cells/well. Cells were differentiated and maintained in RPMI 1640 supplemented with 0.1% BSA and 50 ng/ml of 2.5S mouse nerve growth factor (NGF) (Becton Dickinson) (maintenance medium) beginning on the day of plating. Morphologic differentiation was confirmed by microscopic visualization of dendritic processes.

Establishment of a quiescent infection and reactivation of HSV-1

Quiescent HSV-1 infections in neuronally differentiated PC12 cells were established in 12-well plates using 100 μM ACV (Sigma; St. Louis, MO) as previously described (Danaher et al, 1999a; 2001; Miller et al, 2003). After ACV withdrawal, a quiescent state (i.e., free of detectable infectious virus in culture supernatants) was maintained for 7 days prior to induction. Quiescent cultures, that were free of detectable infectious virus, were induced to activate virus on day 17 postinfection, unless otherwise indicated, by subjecting the cells to heat stress (43°C for 3 h), as previously described (Danaher et al, 1999a), or by culturing the cells in maintenance medium supplemented with 50 μM forskolin (Sigma), 0.2 to 20 μM 5′-azacytidine (Sigma), 0.5 to 50 mM sodium butyrate (Sigma), or 0.4 ng/ml to 400 ng/ml TSA (Sigma). Sodium butyrate and TSA were prepared in DMSO. Treatment of QIF-PC12 cells with DMSO at the concentrations used for solubilization of these compounds does not induce detectable HSV-1 reactivation (Danaher et al, 1999b). Culture supernatants were monitored for virus production by plaque forming assays on monolayers of Vero cells as previously described (Miller et al, 2003).

Rabbit corneal inoculation

New Zealand white rabbits (2–3 kg) used in these studies were housed in American Association for Laboratory Animal Care-approved quarters. The corneas of the rabbits were topically anesthetized with proparacaine hydrochloride and inoculated with 2 x 10⁵ PFU of virus following light scarification with a 27-gauge needle. To maximize viral adherence, the eyelids were gently rubbed. Corneal infection was monitored by SLE from day 3 to 14 post-inoculation and scored based on lesions characteristic of HSV-1 infection as previously described (Miller et al, 2003).

In vivo induced viral reactivation

Viral reactivation was induced at least 28 days post-inoculation, by which time viral latency was established. Only eyes of rabbits that had a primary infection confirmed by virus isolation and were clear of epithelial and stromal defects at the time of iontophoresis and negative for HSV-1 by eye swab on the day of harvest were used. Rabbits received transcorneal iontophoresis of 0.015% epinephrine (0.8 mA for 8 min) once a day for 3 days (Gebhardt et al, 1999; Hill et al, 1998; Kwon et al, 1981). The frequency of viral reactivation was determined by detecting viral shedding in the tear film collected on Dacron^TM swabs for 7 days following the first of three treatments of epinephrine iontophoresis. Swabs were placed onto PRK cells, and cells were monitored for cytopathic effects indicative of infectious virus.

Explant (cocultivation) assay

Following recovery from adrenergic induction, the trigeminal ganglia (TG) of randomly selected rabbits were removed, and the tissues were separated into four or five pieces and placed in Eagles’ modified essential medium containing 10% fetal bovine serum. Beginning 7 days after removal of the ganglia, the culture supernatants were assayed daily for infectious virus on PRK cells through day 20 post-removal. From day 20 to 30, the medium was assessed every other day.

Quantification of HSV DNA from QIF-PC12 cells

DNA was isolated from QIF-PC12 cells harvested on day 25 postinfection using the QIAamp DNA Mini Kit (Qiagen, Valencia, CA) according to the manufacturer’s directions. Real time PCR was performed on an ABI Prism 7700 Sequence Detection system (PE Applied Biosystems, Foster City, CA) in a 50 xl reaction volume consisting of final concentrations of 1X TaqMan Universal PCR master mix (PE Applied Biosystems), 900 nM primers, and 250 nM TaqMan probe. Analysis of host DNA was performed with rat GAPDH gene primers and probe (GAPDH sense 5′-GAACATCATCCCTGCATCCA-3′; GAPDH antisense 5′-CCAGTGAGCTTCCCGTTCA-3′; GAPDH probe 5′-CTTGCCCACAGCCTTGGCAGC-3′) as described elsewhere (Medhurst et al, 2000). Standards consisting of 10-fold serial dilutions of PC12 DNA ranging from 2 to 20,000 haploid genome equivalents per reaction were performed in triplicate. Primers and a probe specific to the HSV-1 gene encoding VP16 (forward 5′-TCGACGACTTGGGCTTTAGC-3′; reverse 5′-GAAAACAGATCCTCGTTCCAGGTA-3′; probe 5′-CCCCGCGCTATGTACCATGCTCG-3′) were designed with the use of the Primer Express 1.0 software (PE Applied Biosystems). Standards consisting of 10-fold serial dilutions of a plasmid containing the HSV-1 VP16 gene ranging from 2 X 10² to 2 X 10⁶ genome equivalents per reaction were performed in triplicate. Probes were labeled at the 5′ end with the reporter fluorochrome, 6-carboxy-fluorescein (6-FAM), and at the 3′ end with quencher fluorochrome, 6-carboxy-tetramethyl-rhodamine (TAMRA) (PE Applied Biosystems).

Quantification of HSV DNA from TG

To quantitate the latent viral genome copies in rabbits, DNA was isolated from TG using a commercial extraction method (DNeasy Tissue Kit; Qiagen). All real-time PCR reactions were performed in a 50 xl mixture containing 100 ng of TG DNA (2 xl), 1X PCR buffer (Invitrogen, Carlsbad, CA), 4 mM MgCl₂, 0.2 xM of each primer, 0.2 xM of probe, 0.2 mM dNTPs mix and 0.025 Unit of PlatinumTaq thermostable DNA polymerase (Invitrogen). Real-time quantitations were performed using the BIO-RAD iCycler iQ system (BioRad, Hercules, CA). The fluorescence threshold value was calculated using the iCycle iQ system software. Relative copy number was calculated using a standard curve generated from purified plasmid DNA (a generous gift from Dr. D.C. Bloom, University of Florida, Gainesville) carrying HSV-1 polymerase gene that had been serially diluted in water to contain from 1 copy to 1 million copies in 2 μl. The forward primer was 5′-CATCACCGACCCGAGGAGGAC-3′, the reverse primer was 5′-GGGCCAGGCGCTTGTTGGTGTA-3′, and the probe was 5′-CCGCCGAACTGAGCAGACACCCGCGC-3′. Primers were purchased from LSU Health Sciences Center Core Laboratories, New Orleans, LA. Probes were labeled at the 5′ end with the reporter fluorochrome, 6-FAM, and at the 3′ end with fluorochrome black hole quencher (BHQ) (Integrated DNA Technologies, Coralville, IA). Standard curves of the Ct values plotted against the logarithm of the copy number were found to be linear from 10 to 10⁶ copies per well.

Quantification of RNA levels

PC12 cells were seeded at 8 X 10⁵ cells/well in 6-well collagen-coated plates, infected with strain 17⁺ at an MOI of 5, and maintained in the presence of ACV for the duration of the experiment. Cultures were induced with TSA or mock-treated (maintenance medium), as described above. RNA was isolated at the indicated time points using the RNeasy Mini Kit (Qiagen) as recommended and quantified spectrophotometrically. Total RNA (2.5 μg) was treated with DNAse using DNA-free^TM (Ambion, Austin, TX) as recommended. DNase-treated RNA was reverse transcribed using random primers and SuperScript II (Invitrogen) as recommended. The quantity of cDNA synthesized for selected host and viral transcripts was determined by real-time PCR. Genes assessed included the 2.0-kb major LAT, 8.3- kb minor LAT, α0 (ICP0), α4 (ICP4), α27 (ICP27), U_L39 (ribonucleotide reductase), and U_L48 (VP16). Primers and probes are summarized in Table 1 and include those described by Cohrs et al 2000 for the major LAT and U_L39, Medhurst et al 2000 for GAPDH, and others designed using Primer Express software (PE Applied Biosystems). There was comparable sensitivity for all of the gene sets and primers used. All PCR assays of negative control (−Rtase) and experimental (+RTase) samples were performed individually and in triplicate, respectively, from duplicate samples. Each real-time PCR assay included five standards examined in triplicate.

Statistical analysis

Data are presented as means. Results were analyzed with the two-tailed Student’s t test, using the statistical package in Microsoft Excel (Microsoft Corp., Redmond, WA).