Modulation of reactivation of latent herpes simplexvirus 1 in ganglionic organ cultures by p300/CBPand STAT3Te Du1, Guoying Zhou1, and Bernard Roizman2

Marjorie B. Kovler Viral Oncology Laboratories, The University of Chicago, Chicago, IL 60637

Contributed by Bernard Roizman, May 24, 2013 (sent for review May 9, 2013)

A key property of herpes simplex viruses (HSVs) is their ability toestablish latent infection in sensory or autonomic ganglia and toreactivate on physical, hormonal, or emotional stress. In latentlyinfected ganglia, HSV expresses a long noncoding RNA and a setof microRNAs, but viral proteins are not expressed. The mechanismby which latent HSV reactivates is unknown. A key question is,what is the mechanism of reactivation in the absence of tegumentproteins that enable gene expression in productive infection? Else-where we have reported the use of ganglionic organ cultures thatenable rapid reactivation in medium containing antibody to NGF ordelayed reactivation in medium containing NGF and EGF. We alsoreported that in the ganglionic organ cultures incubated in mediumcontaining antibody to NGF, all viral genes are derepressed at oncewithout requiring de novo protein synthesis within the time frameof a single replicative cycle. Here we report that latent HSV in gangliaimmersed in medium containing NGF and EGF is reactivated by(i) broad spectrum as well as specific histone deacetylase 1 or histonedeacetylase 4 inhibitors, (ii) activation of p300/CBP, and (iii) eitherSTAT3 carrying the substitution of tyrosine 705 to phenylalanine oran inhibitor of STAT3. Conversely, reactivation of latent HSV wasblocked by p300/CBP inhibitor in medium containing antibody toNGF. The results suggest that (i) STAT3 is required for the mainte-nance of the latent state and interference with its functions leadsto reactivation and (ii) p300/CBP is essential for HSV reactivation.

latency-associated transcript | nerve growth factor

Herpes simplex viruses 1 and 2 (HSV-1 and HSV-2) are com-mon human pathogens that are transmitted from person toperson by physical contact between infected and uninfected tis-sues. Characteristically, these viruses vigorously replicate at theportal of entry into the body. Concurrently they are transportedretrograde to neurons of ganglia innervating that site (1). In neu-rons, the viruses establish a latent infection (2, 3). In the courseof latent infection, the genes expressing the viral proteins arerepressed (4, 5), and only a long noncoding RNA designated asa latency-associated transcript (LAT) (6–9) and microRNAs areexpressed (10–12). Periodically, in response to physical hormonalor emotional stress, the virus replicates and is transported to asite at or near the portal of entry into the body where it can causea lesion. The virus made in the lesion can be transmitted to anoninfected individual (4).The fundamental problems associated with HSV infections

are multifold. HSV-1 is transmitted primarily by oral contact.Although in most individuals the recurrent lesions occur at themucocutaneous region of the lip or genitals (classical fever blis-ters), the virus is occasionally transported to the brain where itcauses encephalitis or to the eye where it causes herpes keratitis.Encephalitis typically occurs in 1 in 100,000 individuals per yearand can cause severe sequelae, including death (13). Recurrentherpes keratitis is a major cause of blindness in United States(14). The problems associated with HSV-2 genital infections canbe overwhelming, particularly if the recurrences are frequent,painful, or transmitted to a newborn.

Much of the morbidity of HSV-1 and HSV-2 infections is dueto the capacity of these viruses to establish latent infections inneurons and to reactivate. In most instances, the replication ofHSV-1 and HSV-2 can be controlled by existing antiviral drugs.Importantly, however, antiviral drugs have no effect on latentvirus, reactivation frequency, or shedding. Adding to the pressureto eliminate reactivating virus is evidence that individuals withrecurrent genital HSV infections are more susceptible to HIVinfection than those who are not infected (15, 16).The studies reported here are parts of an effort to define the

role played by specific regulatory pathways in the maintenance ofviral genes in a repressed state in trigeminal ganglia (TG) har-boring HSV-1 in a latent, silent state. The experimental design isas follows: Mice are inoculated by the corneal route with wild-type or mutant viruses. Wild-type HSV-1 replicates in the eye, istransported retrograde to TG, and replicates in some neuronsbut is silenced and retained in a latent state in other neurons(17). By day 30 after infection, the ganglia contain only silenced,latent virus. Our objective is to analyze the events that take placewithin a 24-h interval after induction of reactivation, that is, withinthe time frame of a single virus replicative cycle. To achieve thisobjective, ganglia are excised and incubated intact in mediumcontaining antibody to nerve growth factor (NGF) or in mediumcontaining both NGF and epidermal growth factor (EGF).Deprivation of NGF leads to activation of viral gene expressionand abrogation of latency (18). In the presence of both NGF andEGF, reactivation is delayed (19). The studies published to dateindicate the following:

i) Viral genes form several groups that are coordinate and se-quential on productive infection in cell culture and also atthe portal of entry into the body (20, 21). Thus, α-genes arederepressed with the involvement of viral protein 16 (VP16),a protein brought into cells during infection (22). At least oneα-protein designated infected cell protein 0 (ICP0) plays a

Significance

HSVs initiate human infections with the aid of viral tegumentproteins brought along with viral DNA into cells. These virusesthen enter and establish latent, silent infection in sensoryganglia. Periodically, HSV reactivates from a latent state. A keyunresolved question is the mechanism by which the virusreactivates in the absence of the tegument proteins. Studies ofmurine trigeminal ganglia harboring latent virus and main-tained in organ cultures suggest that viral DNA is maintained ina dynamic equilibrium favoring gene repression. The equilib-rium shifts toward gene expression on inactivation of histonedeacetylases, inhibition of STAT3, or activation of p300/CBP.

Author contributions: T.D., G.Z., and B.R. designed research; T.D. and G.Z. performedresearch; T.D., G.Z., and B.R. analyzed data; and B.R. wrote the paper.

The authors declare no conflict of interest.1T.D. and G.Z. contributed equally to this work.2To whom correspondence should be addressed. E-mail: bernard.roizman@bsd.uchicago.edu.

www.pnas.org/cgi/doi/10.1073/pnas.1309906110 PNAS | Published online June 20, 2013 | E2621–E2628

MICRO

BIOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

July

1, 2

021

mailto:bernard.roizman@bsd.uchicago.eduwww.pnas.org/cgi/doi/10.1073/pnas.1309906110

key role in the derepression of β and at least a large fractionof γ-genes (23). Upon incubation of intact TG in mediumcontaining anti-NGF antibody, genes representative of all co-ordinately regulated viral genes are derepressed at once in theabsence of prior protein synthesis (19). In effect, the mecha-nism of reactivation does not involve VP16 or ICP0.

ii) One hypothesis that could explain the massive derepressionof all viral genes at once is that, in the absence of NGF, theneuron undergoes apoptosis (24). Indeed, exposure of trigem-inal organ cultures to proapoptotic drugs induced activationof viral genes in the presence of NGF and EGF. However,unlike the spontaneous reactivation in the absence of NGF,the reactivation of viral genes in the presence of at least oneproapoptotic drug required concurrent protein synthesis (18).

In this report, we detail the role of STAT3 and p300/CBPin reactivation of HSV-1 in latently infected ganglia. In thefirst series of experiments, we show that inhibition of histone-deacetylating enzymes and activation of histone acetyl transferasep300/CBP results in reactivation of latent virus whereas inhi-bition of p300/CBP suppresses reactivation. In the second se-ries of experiments, we show that STAT3 plays a key role in themaintenance of HSV-1 in a latent form inasmuch as interferencewith its functions induces reactivation.

ResultsHistone Deacetylase Inhibitors Activate Latent Virus in the GanglionicOrgan Culture Model. Earlier studies have shown that HSV-1 canbe reactivated from latently infected neurons by histone deace-tylase (HDAC) inhibitors (25). The purpose of these studies wasto verify that HDAC inhibitors, both broad spectrum and thosewith some degree of specificity, reactivate HSV-1 in latentlyinfected TG maintained in organ culture incubated in mediumcontaining NGF + EGF. In this series of experiments, we tested

three inhibitors. Trichostatin A (TSA) is a broad-spectrum HDACinhibitor (25, 26). Earlier studies have shown that it induces thereactivation of virus in neurons harboring latent virus (25). pyridin-3-ylmethyl 4-(2-aminophenylcarbamoyl)benzylcarbamate (MS275)and 3-[5-(3-(3-Fluorophenyl)-3-oxopropen-1-yl)-1-methyl-1H-pyrrol-2-yl]-N-hydroxy-2-propenamide (MC1568) at low concen-trations act as specific inhibitors of HDAC 1 and 4, respectively(27, 28). At high concentrations, their specificity decreases. Allthree HDACs inhibitors induced the reactivation of HSV in TGharvested 30 d after infection and incubated in medium containingNGF and EGF (Fig. 1 A and B). As in earlier studies, the resultsshow the geometric mean levels of viral mRNAs from six gangliaextracted at the time of excision and at 24 h after excision. Aspreviously reported, LAT and miRNAs decrease in amounts (Fig.1C) concurrent with accumulation of viral gene products (Fig. 1B).

Activation of p300/CBP Induces Reactivation of Latent Virus and,Conversely, Inhibition of p300/CBP Blocks Reactivation. We reporttwo series of experiments. In the first, ganglia excised from miceinoculated at least 30 d earlier were incubated in medium con-taining NGF+EGF and supplemented with N-(4-chloro-3-trifluoromethyl-phenyl)-2-ethoxy-benzamide (CTB). CTB isa powerful activator of histone acetyltransferase activity ofP300/CBP by binding and altering the structure of the enzyme(29). At 24 h after incubation, the ganglia were extracted andanalyzed for the presence of mRNA representative of the fourgroups of viral mRNAs and also for the presence of the LATand viral miRNAs (Fig. 2). Following exposure to the drug, allviral mRNAs increased ∼50-fold whereas miRNAs and theLAT decreased ∼10-fold.

Fig. 1. Effect of HDAC inhibitors (TSA, MS275, MC1568) on HSV-1(F) reac-tivation. Murine TG excised 30 d after infection with HSV-1(F) by the cornealroute were processed immediately after excision or after 24 h incubation inmedium containing anti-NGF antibody, NGF+EGF, NGF+EGF+TSA (400 nM)(A), NGF+EGF+MS275 (10 and 20 μM) (B and C), or NGF+EGF+MC1568 (10and 20 μM) (B and C). The figure shows the geometric mean amounts ofviral mRNAs normalized to 50 ng cellular RNA or viral miRNAs normalized to108 copies of cellular miRNA quantified by cellular miRNA standard Lethal 7family member a (Let-7a). The numbers shown hereafter are geometricmeans ± SEs based on assays of six TG per group.

Fig. 2. HSV-1(F) reactivation from latency induced by the p300/CBP activa-tor CTB. TG excised 30 d after inoculation of mice with HSV-1(F) were pro-cessed immediately after excision (columns 1 and 5) or after 24 h incubationin medium containing NGF+EGF (columns 2 and 6) or NGF+EGF+CTB (200and 500 μM, columns 3 and 7 and 4 and 8, respectively). Normalized copiesof viral mRNAs [ICP27 (a), TK (b), VP16 (c), UL41 (d), LAT (e)] and viral miRNAs[mir-H3 (f), mir-H5 (g), and mir-H6 (h)] are shown.

E2622 | www.pnas.org/cgi/doi/10.1073/pnas.1309906110 Du et al.

Dow

nloa

ded

by g

uest

on

July

1, 2

021

www.pnas.org/cgi/doi/10.1073/pnas.1309906110

Earlier studies have shown that the spice curcumin is a po-tent inhibitor of the acetyl transferase activity of p300 (30). Totest whether curcumin will block reactivation, we repeated theexperiment described above but with curcumin in place ofCTB. In this experiment, the ganglia were incubated either inmedium containing NGF+EGF or in medium containing an-tibody to NGF. As shown in Fig. 3A, curcumin blocked thereactivation of HSV and degradation of the LAT and miRNAsindependent of the medium in which the ganglia were in-cubated. It is noteworthy that, in ganglia incubated in mediumcontaining antibody to NGF, curcumin suppressed the accumu-lation of small amounts of mRNAs normally detected at the timeof excision of the ganglia. Compare, for example, the amounts ofmRNA detected in freshly excised ganglia (Fig. 3A, column 1)with those detected in ganglia incubated in medium containingcurcumin (Fig. 3A, columns 4 and 5). It is also noteworthy thatat the higher concentration (300 μM) curcumin suppressed theaccumulation of the LAT (Fig. 3B, columns 4 and 5).

Inhibition of STAT3 Phosphorylation Induces Activation of LatentHSV. In this series of experiments, we tested the effects of theSTAT3 inhibitor 2-hydroxy-4-(2-(tosyloxy)acetamido)benzoic acid(NSC74859) on reactivation of HSV from murine TG harboringlatent virus (31). The experimental design was similar to thatreported above. The excised ganglia were incubated in mediumcontaining NGF+EGF and one of two concentrations of thedrug. The results of analyses of the mRNAs extracted fromganglia 24 h after excision are shown in Fig. 4. In brief, theresults show that the representative of all four temporally regu-lated groups of viral genes was up-regulated by 24 h after exci-sion of the ganglia and incubation in the presence of the STAT3inhibitor. The results suggest that STAT3 plays a role in themaintenance of HSV-1 in the latent state in murine TG.

Construction and Testing of Recombinant HSV-1 Expressing a Wild-Type and Dominant-Negative STAT3. The objective of this series ofexperiments was to verify the conclusion that disruption of STAT3function induces reactivation of latent virus by constructionof viruses that deliver to the infected neurons a wild-type or adominant-negative STAT3.The procedures for construction of the recombinant viruses are

described in Materials and Methods. Fig. 5 shows the schematicdiagrams of the domains of STAT3 and the structure of thegenomes of the recombinant viruses. Specifically, the wild-typeSTAT3 coding sequences flanked by the SV40 early promoter atit 5′ terminus and the myc tag at its 3′ terminus were insertedbetween Unique Long 3 (UL3) and UL4 ORFs in recombinantnamed as R128 (Fig. 5, line 2). In R130 (Fig. 5, line 3), we insertedin the same location an identical construct except that tyrosine 705was replaced by phenylalanine. The objective of the Y705Fsubstitution was to block the phosphorylation of tyrosine705.

Characterization of the Course of R128 and R130 Recombinant VirusInfection in Mice. In this series of experiments, we characterized theaccumulation of viral DNA, mRNAs representative of the majorkinetic classes of viral mRNAs, the LAT, and the representativemiRNAs in mice following intracorneal inoculation of mice. As inall studies reported, each experimental point represents the results

Fig. 3. Viral gene, LAT, and miRNA expressions during virus reactivation fromlatently infected ganglia treated by p300 inhibitor (curcumin). TG excised 30 d af-ter inoculation of HSV-1(F) were processed immediately after excision (columns 1)or after 24 h of incubation in medium containing anti-NGF antibody (columns 2),anti-NGF antibody + curcumin (100 and 300 μM, columns 4 and 6, respectively),NGF+EGF (columns 3), or NGF+EGF+curcumin (100 and 300 μM, columns 5 and 7,respectively). ViralmRNAs (ICP27, TK, VP16, UL41) (PanelA), LAT andmiRNAs (mir-H3,mir-H5, andmir-H6) (Panel B) weremeasured and normalized to cellular RNA.

Fig. 4. HSV-1(F) reactivation from latency induced by STAT3 inhibitor(NSC74859). TG excised on 30 d after inoculation of HSV-1(F) were processedimmediately (column 1) or after 24 h of incubation in medium containingNGF+EGF (column 2) or NGF+EGF+NSC74859 (10 and 50 μM, columns 3 and 4,respectively). Normalized copies of ICP27 (a), TK (b), VP16 (c), and UL41 (d)are shown.

Du et al. PNAS | Published online June 20, 2013 | E2623

MICRO

BIOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

July

1, 2

021

obtained on six ganglia. The salient features of the results arerelated below.Fig. 6A shows the accumulation of viral DNAs in TG har-

vested 1, 3, 7, and 14 d after infection with wild-type HSV-1(Fstrain) recombinant viruses. Consistent with earlier studies, HSVDNA reached peak levels on day 7 and declined thereafter. In thisexperiment, the amounts of viral DNA accumulating in gangliaof mice infected with R130 were consistently lower than thoseof mice infected with wild-type virus or recombinant carryingwild-type STAT3.Fig. 7 shows the patterns of accumulation of infected cell

protein (ICP27), thymidine kinase (TK), VP16, or UL41 mRNAsrepresentative of viral α-, β-, γ1-, and γ2-genes, respectively, andof the LAT and selected miRNAs. In the experiment shown, wehave observed significant fluctuation in the levels of mRNAs onday 3 but less at later time points. In general, mRNAs attainedpeak levels on days 3 or 7 and decreased thereafter. The dataindicate that the patterns of accumulation of mRNAs of wild-type virus were either similar or slightly elevated with respect tothose of the recombinant viruses. In a similar vein, the patterns ofaccumulation of the LAT and miRNAs were either similar orslightly elevated with respect to the accumulation of the corre-sponding products of the recombinant viruses.

Interference with STAT3 Functions by a Virus Expressing a GeneEncoding a Dominant-Negative STAT3. The objective of the studiespresented here was to verify the conclusion based on the STAT3inhibitor studies that STAT3 plays a role in defining the status oflatent virus and that interference with the function of STAT3

would lead to virus reactivation. To test this hypothesis, TG wereexcised 30 d after infection with wild-type virus or with R128 orR130 recombinant viruses. One set of ganglia were immediatelyextracted and assayed for the amounts of viral DNA and mRNAsrepresenting the major kinetic classes of viral genes. The remainingganglia were incubated in medium containing antibody to NGFor both NGF+EGF for 24 h and then assayed for the amounts ofviral mRNAs, the LAT, or representative miRNAs. The resultswere as related below.The relative amounts of viral DNA recovered from TG at the

time of excision of the ganglia are shown in Fig. 6B. The resultsindicate that there was significantly less viral DNA in TG ofmice exposed to the recombinant virus carrying the Y705F sub-stitution than in mice infected with wild-type virus. Conversely,

Fig. 5. Schematic representation of the recombinant viruses carrying wild-type or dominant-negative human STAT3 in the HSV-1(F) genome. (Line 1) DNA sequencearrangement of HSV-1 DNA. (Line 2) Schematic representation of HSV-1 recombinant virus expressing wild-type STAT3 designated as R128 and arrangement of STAT3functional domains. (Line 3) Schematic representation of HSV-1 recombinant virus expressing dominant-negative STAT3 (with single substitution of Y705F) desig-nated as R130. Wild-type or dominant-negative STAT3 withmyc tag flanked by the SV40 promoter and poly(A) sequence were inserted between UL3 and UL4 genes.

Fig. 6. Accumulation of viral DNA in TG of mice infected with wild-typeor recombinant viruses. On indicated days after infection with HSV-1(F),R128, or R130 viruses, the TG were removed and extracted. The DNA copynumbers normalized to 50 ng cellular DNA are shown as a function of daysafter infection. (A) DNA copy number in murine TG at days 1–3, 7, and 14after infection. (B) DNA copy number in TG 30 d after infection.

Fig. 7. Gene expressions in murine TG after infection with R128 or R130recombinant viruses. On indicated days after infection with HSV-1(F), R128,or R130 viruses, mouse TG were removed, extracted, and subjected to RNAassay. The relative number of copies of mRNAs encoding ICP27 (a), TK (b),VP16 (c), UL41 (d), or LAT (e) normalized to cellular RNA and mir-H3 (f), mir-H5 (g), and mir-H6 (h) normalized with respect to cellular miRNA are shownas a function of days after infection.

E2624 | www.pnas.org/cgi/doi/10.1073/pnas.1309906110 Du et al.

Dow

nloa

ded

by g

uest

on

July

1, 2

021

www.pnas.org/cgi/doi/10.1073/pnas.1309906110

TG harboring the R128 recombinant viral genome expressing awild-type STAT3 gene contained at least as much viral DNA asthe ganglia carrying latent wild-type virus and significantly morethan ganglia harboring latent R130 recombinant virus carryingthe Y705F substitution.Fig. 8 shows the relative amounts of mRNAs transcribed from

representative genes sequentially derepressed during productiveinfection. Fig. 8, column 1, shows the basal levels of mRNAdetected in ganglia processed immediately after excision. As pre-viously described, during the time frame of a productive infectionthe amounts of viral transcripts increase ∼50-fold on incubation inmedium containing anti-NGF antibody and ∼2- to 3-fold in me-dium containing NGF+EGF. We detected fewer copies of viralmRNA in ganglia harboring latent recombinant viruses at 0 h(compare columns 4 and 7 with column 1 in Fig. 8). Comparisonof columns 1–3 with 7–9 shows that recombinant viruses R128and R130 and wild type HSV-1(F) exhibit similar patterns, in-dicating that in ganglia harboring R128 the pattern of activationof latent virus is similar to that of HSV-1(F). In contrast, gangliaharboring latent R130 exhibited higher levels of expression andaccumulation of mRNAs in medium containing NGF+EGF

than in medium containing antibody to NGF (compare columns1–3 or 7–9 with 4–6).We conclude that the results of this study are congruent with

those obtained with the STAT3 inhibitor and suggest that STAT3defines the status of latent HSV-1 in murine TG.

Induction of Reactivation of Latent HSV-1(F) by the STAT3 Inhibitor inContrast to the Induction of Reactivation by p300/CBP ActivationRequires de Novo Protein Synthesis. The impetus for these studieswas the observation that TG harboring the R130 recombinantvirus accumulated higher levels of viral mRNAs only upon in-cubation in medium containing NGF+EGF but not in mediumcontaining antibody to NGF. One interpretation of the results isthat the dominant-negative STAT3 carrying the Y705F substitutionhad to perform or induce a function that was partly or totallyblocked in ganglia incubated in medium containing antibody toNGF. To test the hypothesis that this function required de novoprotein synthesis, ganglia excised 30 d after infection wereincubated in medium containing NGF+EGF and the STAT3inhibitor alone or inhibitor plus cycloheximide. As a control du-plicate ganglia harboring latent wild-type virus were incubatedin medium containing inhibitory concentrations of TSA alone orboth TSA and cycloheximide. The rationale in designing this studywas based on the expectation that induction of reactivation byHDAC inhibitors would not require de novo protein synthesis.The results shown in Fig. 9 are as follows: The accumulation of

mRNAs in ganglia exposed to TSA and cycloheximide (columns

Fig. 8. Reactivation of latent HSV-1(F), R128, and R130 viruses in TG at24 h after excision and incubation in medium containing anti-NGF antibodyor NGF+EGF. At 30 d postinoculation, TG excised from the wild-type or re-combinant viruses were processed immediately (columns 1, 4, 7) or after 24 hof incubation in medium containing anti-NGF antibody (columns 2, 5, 8) orNGF+EGF (columns 3, 6, 9). Copy numbers of mRNAs encoding ICP27 (a), TK(b), VP16 (c), and UL41 (d) normalized to cellular RNA are shown.

Fig. 9. Reactivation of latent HSV-1(F) in murine ganglia incubated in me-dium containing the HDAC inhibitor TSA or STAT3 inhibitor NSC74859 inthe presence or absence of cycloheximide. TG excised on 30 d after infec-tion with HSV-1(F) were processed immediately (columns 1 and 6), or after24 h of incubation in medium containing NGF+EGF+TSA (400 nM) (columns 2and 7), NGF+EGF+TSA + cycloheximide (CHX, 150 μg/mL) (columns 3 and 8),NGF+EGF+NSC74859 (10 μM) (columns 4 and 9) or NGF+EGF+NSC74859+CHX(columns 5 and 10). Normalized copies of viral mRNAs (ICP27, TK, VP16, UL41)(Panel A) LAT, mir-H3, mir-H5, and mir-H6 (Panel B) are shown.

Du et al. PNAS | Published online June 20, 2013 | E2625

MICRO

BIOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

July

1, 2

021

2 and 3) were similar to those obtained in the absence of cy-cloheximide and 10-fold higher than those obtained in freshlyexcised ganglia (column 1). As expected from earlier studies(19), the amounts of the LAT and miRNAs were higher in cy-cloheximide-treated ganglia (column 8) than in ganglia treatedwith TSA only (column 7).In contrast to the results of studies with the HDAC inhibitor, the

levels of accumulated mRNAs were consistently lower in gangliatreated with both STAT3 inhibitor and cycloheximide (column 5)than with the inhibitor alone. In contrast, the levels of the LATand miRNAs were higher in cycloheximide-treated ganglia.We conclude that induction of viral gene expression in latently

infected ganglia by the inhibitor of STAT3 requires de novo proteinsynthesis. In contrast, activation of viral gene expression by HDACinhibitor does not require prior protein synthesis.

DiscussionThe focus of this report is on the reactivation of latent HSV-1 inthe TG organ culture model. The results of the studies presentedin this report are shown schematically in Fig. 10. The salientfeatures of the results and their significance are summarized asfollows:

i) In an earlier study we reported that, in organ culture contain-ing antibody to NGF, ganglia harboring latent HSV reacti-vated the virus within the time frame of a single replicativecycle after excision from mice (18). We also reported thatgenes representing distinct kinetic classes of viral genes areexpressed concurrently. The hypothesis that we advanced toexplain this phenomenon is that deprivation of NGF inducesapoptosis and causes a massive derepression of all viral genesat once. To test the hypothesis, we exposed ganglia harboringlatent virus to proapoptotic drugs in medium containing NGFand EGF. We did indeed observe activation of viral genes inmedium containing dexamethasone, but the induction wasblocked by inhibition of protein synthesis as depicted sche-matically in Fig. 10B (24). The key conclusion of that study isthat the mechanism by which the absence of NGF triggersreactivation of the latent virus is different from that inducedby proapoptotic drugs (18, 24).

ii) One focus of the current report is on the role of the STAT3transcriptional factor in reactivation of latent HSV-1.STAT3 transcriptional factor has been shown to reside in thecytoplasm of neurons (32). Upon induction by cytokines(e.g., IL6) or stress, STAT3 is phosphorylated at residue705 and is translocated to the nucleus where it interactswith and is acetylated by p300/CBP (Fig. 10A) (33–35).STAT3 then activates the transcription of neuroprotectivegenes (e.g., bcl-2, bcl-xl, IAP2, Pim-1, Reg), neurodegener-ative genes (e.g., GAP-43), and neurodevelopmental genes(e.g., GFPAF) (36–41). In this report the inhibitor of STAT3blocks its posttranslational modification. The presumed mech-anism of reactivation induced by STAT3 carrying the Y705Fsubstitution is that once it is translocated into the nucleus, itcompetes with endogenous wild-type STAT3.We do not have direct evidence that activated, nuclear

STAT3 blocks reactivation of latent virus. The available evi-dence is twofold. Foremost, overexpression of wild-typeSTAT3 did not alter the pattern of accumulation of viraltranscripts in ganglia maintained in medium containingNGF+EGF. More significantly, interference with STAT3function either by an inhibitor of STAT3 function or by adominant-negative STAT3 resulted in reactivation of latentvirus. The fundamental conclusion of this report is that STAT3defines the status of latent HSV-1 in TG. One hypothesis thatis consistent with the data but does not prove it is that activatedSTAT3 in turn can activate the expression of either neuronprotective anti-apoptotic genes or proapoptotic genes. Thenature of the STAT3 function may depend on the STAT3activator, whether a cytokine or stress signaling.

iii) The second focus of our report is on the role of p300/CBP.We show that curcumin, a potent inhibitor of p300/CBP, blocksreactivation of ganglia immersed in medium lacking NGF.These results suggest that reactivation requires p300/CBP his-tone acetyl transferase activity. As cited above, reactivation ofviral genes in ganglia harboring latent virus in medium con-taining antibody against NGF can take place in the absenceof de novo protein synthesis. Implicit in this observation isthat, in ganglia harboring latent virus, p300/CBP is availableand its function in facilitating viral gene expression does

Fig. 10. Schematic representation of the roles of STAT3, HDAC, and p300/CBP in the transcriptional activation of latent HSV-1 as a function of STAT3, HDACs,and p300/CBP. (A) STAT3 activated by stress or cytokines is translocated to the nucleus where it forms dimers, recruits p300/CBP, and induces the expression ofgenes involved in neuronal protection, neural degeneration, or neural development. The results shown schematically in A suggest that activated STAT3 blocksreactivation of latent virus in ganglia maintained in the presence of NGF and EGF. (B) Reactivation of latent virus is induced by maintaining ganglia in mediumdeprived of NGF in the absence of de novo protein synthesis. Reactivation in the absence of de novo protein synthesis is induced by inhibition of HDACs or byactivation of p300/CBP. Reactivation is also induced by inhibitors of STAT3 or by a dominant-negative STAT3 or by proapoptotic drugs, but in these instancesreactivation requires de novo protein synthesis.

E2626 | www.pnas.org/cgi/doi/10.1073/pnas.1309906110 Du et al.

Dow

nloa

ded

by g

uest

on

July

1, 2

021

www.pnas.org/cgi/doi/10.1073/pnas.1309906110

not require prior or concurrent protein synthesis. This con-clusion is reinforced by the observation that CTB, a drugthat binds to p300/CBP and activates it by altering its con-formation effectively, induced viral gene expression in gangliaincubated in medium containing NGF and EGF without prioror concurrent protein synthesis.

iv) Studies reported elsewhere have shown that wide-spectrumHDAC inhibitor induced HSV reactivation (25). In thisstudy we reaffirm the expectation that neurons harboring latentvirus by inhibition of the HDACs do not require prior or con-current protein synthesis. We also show that HDAC inhib-itors acting on a more narrow spectrum of HDACs also ac-tivate latent HSV-1. Implicit in the observation that HDACinhibitors can induce reactivation is the hypothesis that viralgenes are in a dynamic equilibrium with respect to histoneacetylation and deacetylation and that the equilibrium shiftsas the supply of functional HDACs is depleted. The hypoth-esis predicts that massive activation of p300/CBP may alsoact to shift the equilibrium, a hypothesis supported by theactivation of latent virus gene expression by the p300 activa-tor CTB.

Finally, two issues are relevant here. First, a fundamental ob-jective of the studies reported here is to find the means to inducethe reactivation of HSV in the presence of antiviral drugs to de-crease and ultimately eliminate reactivable virus from patientswith recurrent herpetic lesions. The results of the studies presentedhere point to several avenues by which this could be accomplished.In brief, we have shown that inhibition of STAT3 or HDACs oractivation of p300/CBP can be predicted to initiate the sequenceof events leading to controlled virus reactivation.Second, to remain latent, viral DNA must be extensively re-

modeled into facultative heterochromatin (42–44). The resultsthat we present are that inhibition of STAT3 or activation ofp300/CBP is capable of modifying latent viral DNA to the pointwhere it can replicate and be susceptible to antiviral drugs. It couldbe expected that the same process may enable the activation ofthe DNAs of other viruses that remain latent with humans forlife. The list includes major human pathogens such as humancytomegalovirus, Epstein–Barr virus, varicella-zoster virus, andhuman immunodeficiency virus.

Materials and MethodsVirus Strains and Cells. Vero cells originally obtained from the American TypeCulture Collection were grown in DMEM supplemented with 5% (vol/vol)FBS. The BAC encoding the HSV-1(F) DNA was reported elsewhere (18).

Plasmid Construction of STAT3 and dnSTAT3. The plasmid (pMXs-STAT3) con-taining wild-type STAT3 was a kind of gift S. Yamanaka (Kyoto University,Japan) (45). The mutant STAT3 with single substitution of Y705F(dnSTAT3) was obtained by use of the QuikChange XL Site-Directed Mu-tagenesis Kit (Stratagene) with two oligo primers: ggtagtgctgccccgtTcct-gaagaccaagttc and gaacttggtcttcaggaacggggcagcactacc).

Construction of STAT3 and dnSTAT3 Recombinant Viruses. The recombinantviruses, the gene encoding human STAT3, or dnSTAT3 was inserted betweenthe genes encoding UL3 and UL4 under the SV40 promoter. The strategy ofvirus construction has been reported elsewhere (17, 18).

Murine Model of Virus Infection. Four-week-old inbred female CBA/J mice(Jackson Labs) received unrestricted access to food and water. All animalstudies were done according to protocols approved by the Institutional AnimalCare and Use Committee of the University of Chicago. Following light scari-fication of the cornea, 1 × 105 pfu of virus were applied in a dropwise mannerin a volume of 5 μL to each cornea of the mice. TG were excised on indicateddays and subjected to DNA replication and viral gene expressions assays.

Murine Model of Virus Reactivation and Drug Treatment. TG were removed30 d after infection and incubated at 37 °C with 5% (vol/vol) CO2 inmedium 199V supplemented with anti-NGF antibody (1 mg/mL; Abcam) for24 h. To temporarily block virus reactivation, TG were incubated in mediumcontaining 300 ng/mL NGF+EGF (Invitrogen). TG were treated by drugs asdescribed in the text. HDAC inhibitor TSA, P300 inhibitor curcumin, and P300activator CTB were purchased from Sigma. HDAC class I specific inhibitorMS275 and HDAC class II specific inhibitor MC1568 were purchased from SelleckChemicals. STAT3 inhibitor NSC74859 was purchased from EMD Millipore.

DNA Copy-Number Assays. Total DNA was extracted from murine TG as re-ported previously (17). The quantification of viral DNA copy numbers in TGwas performed by SYBR green real-time PCR technology (StepOnePlussystem, ABI) using viral TK gene primers and murine adipsin gene primersas internal control.

RNA Isolation and Assays. RNAs depleted and enriched in small RNAs (

23. Gu H, Roizman B (2007) Herpes simplex virus-infected cell protein 0 blocks thesilencing of viral DNA by dissociating histone deacetylases from the CoREST-RESTcomplex. Proc Natl Acad Sci USA 104(43):17134–17139.

24. Du T, Zhou G, Roizman B (2012) Induction of apoptosis accelerates reactivation oflatent HSV-1 in ganglionic organ cultures and replication in cell cultures. Proc NatlAcad Sci USA 109(36):14616–14621.

25. Danaher RJ, Jacob RJ, Steiner MR, Allen WR, Hill JM, Miller CS (2005) Histone deacetylaseinhibitors induce reactivation of herpes simplex virus type 1 in a latency-associatedtranscript (LAT)-independent manner in neuronal cells. J Neurovirol 11(3):306–317.

26. Kazantsev AG, Thompson LM (2008) Therapeutic application of histone deacetylaseinhibitors for central nervous system disorders. Nat Rev Drug Discov 7(10):854–868.

27. Nebbioso A, et al. (2009) Selective class II HDAC inhibitors impair myogenesis bymodulating the stability and activity of HDAC-MEF2 complexes. EMBO Rep 10(7):776–782.

28. Khan N, et al. (2008) Determination of the class and isoform selectivity of small-molecule histone deacetylase inhibitors. Biochem J 409(2):581–589.

29. Balasubramanyam K, Swaminathan V, Ranganathan A, Kundu TK (2003) Small moleculemodulators of histone acetyltransferase p300. J Biol Chem 278(21):19134–19140.

30. Balasubramanyam K, et al. (2004) Curcumin, a novel p300/CREB-binding protein-specific inhibitor of acetyltransferase, represses the acetylation of histone/nonhistoneproteins and histone acetyltransferase-dependent chromatin transcription. J BiolChem 279(49):51163–51171.

31. Siddiquee K, et al. (2007) Selective chemical probe inhibitor of Stat3, identified throughstructure-based virtual screening, induces antitumor activity. Proc Natl Acad Sci USA104(18):7391–7396.

32. Kriesel JD, Jones BB, Hwang IP, Dahms KM, Spruance SL (2001) Signal transducers andactivators of transcription (Stat) are detectable in mouse trigeminal ganglion neurons.J Interferon Cytokine Res 21(6):445–450.

33. Nadiminty N, et al. (2006) Stat3 activation of NF-kappaB p100 processing involvesCBP/p300-mediated acetylation. Proc Natl Acad Sci USA 103(19):7264–7269.

34. Cheng PY, et al. (2011) Interplay between SIN3A and STAT3 mediates chromatinconformational changes and GFAP expression during cellular differentiation. PLoSONE 6(7):e22018.

35. Al Zaid Siddiquee K, Turkson J (2008) STAT3 as a target for inducing apoptosis in solidand hematological tumors. Cell Res 18(2):254–267.

36. Catlett-Falcone R, et al. (1999) Constitutive activation of Stat3 signaling confersresistance to apoptosis in human U266 myeloma cells. Immunity 10(1):105–115.

37. Aoki Y, Feldman GM, Tosato G (2003) Inhibition of STAT3 signaling induces apoptosisand decreases survivin expression in primary effusion lymphoma. Blood 101(4):1535–1542.

38. Li GH, et al. (2009) STAT3 silencing with lentivirus inhibits growth and inducesapoptosis and differentiation of U251 cells. J Neurooncol 91(2):165–174.

39. Lee N, Neitzel KL, Devlin BK, MacLennan AJ (2004) STAT3 phosphorylation in injuredaxons before sensory and motor neuron nuclei: Potential role for STAT3 as a retrogradesignaling transcription factor. J Comp Neurol 474(4):535–545.

40. Dziennis S, Alkayed NJ (2008) Role of signal transducer and activator of transcription3 in neuronal survival and regeneration. Rev Neurosci 19(4–5):341–361.

41. Pranada AL, Metz S, Herrmann A, Heinrich PC, Müller-Newen G (2004) Real timeanalysis of STAT3 nucleocytoplasmic shuttling. J Biol Chem 279(15):15114–15123.

42. Kubat NJ, Tran RK, McAnany P, Bloom DC (2004) Specific histone tail modificationand not DNA methylation is a determinant of herpes simplex virus type 1 latent geneexpression. J Virol 78(3):1139–1149.

43. Kwiatkowski DL, Thompson HW, Bloom DC (2009) The polycomb group protein Bmi1binds to the herpes simplex virus 1 latent genome and maintains repressive histonemarks during latency. J Virol 83(16):8173–8181.

44. Bloom DC, Giordani NV, Kwiatkowski DL (2010) Epigenetic regulation of latent HSV-1gene expression. Biochim Biophys Acta 1799(3–4):246–256.

45. Takahashi K, Yamanaka S (2006) Induction of pluripotent stem cells from mouseembryonic and adult fibroblast cultures by defined factors. Cell 126(4):663–676.

E2628 | www.pnas.org/cgi/doi/10.1073/pnas.1309906110 Du et al.

Dow

nloa

ded

by g

uest

on

July

1, 2

021

www.pnas.org/cgi/doi/10.1073/pnas.1309906110