Research ArticleHerpes Simplex Virus

Inhibition of LSD1 reduces herpesvirus infection, shedding, and recurrence by promoting epigenetic suppression of viral genomes

See allHide authors and affiliations

Science Translational Medicine  03 Dec 2014:
Vol. 6, Issue 265, pp. 265ra169
DOI: 10.1126/scitranslmed.3010643

Abstract

Herpesviruses are highly prevalent and maintain lifelong latent reservoirs, thus posing challenges to the control of herpetic disease despite the availability of antiviral pharmaceuticals that target viral DNA replication. The initiation of herpes simplex virus infection and reactivation from latency is dependent on a transcriptional coactivator complex that contains two required histone demethylases, LSD1 (lysine-specific demethylase 1) and a member of the JMJD2 family (Jumonji C domain–containing protein 2). Inhibition of either of these enzymes results in heterochromatic suppression of the viral genome and blocks infection and reactivation in vitro. We demonstrate that viral infection can be epigenetically suppressed in three animal models of herpes simplex virus infection and disease. Treating animals with the monoamine oxidase inhibitor tranylcypromine to inhibit LSD1 suppressed viral lytic infection, subclinical shedding, and reactivation from latency in vivo. This phenotypic suppression was correlated with enhanced epigenetic suppression of the viral genome and suggests that, even during latency, the chromatin state of the virus is dynamic. Therefore, epi-pharmaceuticals may represent a promising approach to treat herpetic diseases.

INTRODUCTION

The worldwide prevalence of herpes simplex virus (HSV) in the human population is estimated to be 50 to 90%, with 70 to 80% of people serologically positive by adolescence and near 100% of adults >60 years of age positive for viral DNA in autopsied trigeminal ganglia (13). After the initial primary infection, HSV establishes lifelong latent infections in sensory neurons. Periodic reactivation from the latent viral reservoirs results in episodes of lytic infection and recurrent disease (4). In addition to clinical disease, asymptomatic viral shedding is the most prevalent means of HSV transmission.

In addition to oral and genital lesions, HSV ocular infections (primary and reactivations) remain the leading cause of virus-mediated blindness in the western world, with >300,000 cases of ocular infection diagnosed yearly (1, 3, 5). Neonatal infections also represent a critical clinical issue and can result in disseminated infection with mortality, ocular disease, and continued neurological or developmental issues (3). In addition to disease directly related to HSV infection and recurrence, the virus is also a co-pathogen and has been linked to increased rates of HIV transmission (69).

Current control of viral primary infection and reactivation from latency is with antiherpetic pharmaceuticals that target viral replication components to inhibit viral DNA synthesis during the late stage of the infectious cycle. Although an important advance in treatment of HSV disease, these compounds do not completely control reactivation, viral subclinical shedding, or the associated inflammation that contributes to pathologies such as ocular scarification. Additionally, because these inhibitors target virus-encoded replication components, resistant viral mutants have evolved that are particularly evident in immunosuppressed patients and those with recurrent herpetic keratitis (10). Thus, there is a need for new therapeutic approaches to control HSV infection, reactivation, and transmission.

Epigenetics represents an evolving frontier for the development of therapeutics for the control of cancers (1117) and other diseases (1820). Advances in the identification and small-molecule targeting of epigenetic components are rapidly driving emerging epi-pharmaceuticals (2125). However, recent recognition of the role of chromatin and chromatin modulation components in control of HSV lytic infection, latency, and reactivation suggests that epi-pharmaceuticals may also be used in approaches to antiherpes therapies.

Like the cellular genome, many viruses are subject to epigenetic regulation. Herpesviral genomes are encapsidated without histones, but rapidly acquire chromatin structures upon infection (2629). For HSV, this is a dynamic process in which repressive heterochromatin accumulates on the viral genome in the absence of critical viral and host factors (Fig. 1) (3034). This is countered by specific chromatin modulation components, including the transcriptional coactivator HCF-1 in concert with the LSD1 (lysine-specific demethylase 1; KDM1A family) and JMJD2 (Jumonji domain–containing protein 2; KDM4 family) histone demethylases, which reduce the levels of heterochromatic marks associated with the viral chromatin and promote the expression of the first wave of viral genes (32, 33, 35, 36).

Fig. 1. Multiple chromatin modulation enzymes drive the initial chromatin state of the HSV genome.

Model illustrating the initial heterochromatic chromatin state of viral genome during infection or latency and the coactivator complex required to modulate this to a euchromatic chromatin structure for productive infection or reactivation from latency. TCP inhibits the required activity of the histone demethylase LSD1.

Similarly, the recurrent cycle of viral latency reactivation in sensory neurons is modulated by chromatin and chromatin regulatory factors (26, 27, 37). During latency, nucleosomes bearing repressive histone H3 lysine 9 and lysine 27 methylation marks are associated with the silent viral lytic genes (26, 3841). Reactivation from latency is correlated with the transition to euchromatic marks (37, 42, 43) due, at least in part, to the HCF-1/LSD1/JMJD2 modulation complex (32, 35, 44). In cell culture models of HSV infection, inhibition of the activities of either LSD1 or the JMJD2s results in a block to viral gene expression and the progression of infection as well as a block to viral reactivation from latency in a mouse ganglia explant system. Thus, epigenetic therapies could provide approaches to suppression of the initiation of HSV infection and reactivation that are distinct from the present pharmaceuticals by promoting and maintaining epigenetic suppression of the virus.

Here, this concept is demonstrated in three animal model systems where pharmaceutical [tranylcypromine (TCP)] inhibition of the required histone demethylase LSD1 resulted in suppression of primary infection, a block to subclinical shedding, and reduction in recurrent lesions. Treated animals exhibited enhanced epigenetic suppression of the virus that correlated with the phenotypic inhibition of disease. The results represent a comprehensive study demonstrating the principle and the potential of epigenetic approaches to anti-HSV therapy.

RESULTS

Three animal model systems of HSV infection were used to investigate the effects of inhibition of the histone demethylase LSD1. Each animal model has distinct advantages in assessing the complex lifecycle of the virus. The mouse model is appropriate for the quantification of mortality rates and viral loads in tissues during ocular or intranasal infections. The rabbit eye model permits clinical assessment of herpetic keratitis, and, unlike the mouse model, these animals exhibit viral shedding and spontaneous viral reactivation from latency. Finally, the guinea pig genital model of HSV-2 infection is used to monitor clinical lesion recurrence. Together, these animal models comprehensively represent the spectrum of HSV-mediated disease.

Suppression of primary infection in the mouse model system

To initially assess the impact of LSD1 inhibition on HSV primary infection, mice were infected with HSV-1 through the ocular route and treated topically or intraperitoneally with the LSD1 inhibitor TCP (45), control ACV (acyclovir, viral DNA replication inhibitor), or vehicle. As shown in Fig. 2A, mice topically treated with TCP exhibited enhanced survival compared to vehicle-treated animals (vehicle, 37.5%; TCP, 75%; ACV, 85%). In a parallel manner, mice treated intraperitoneally with TCP had reduced viral loads in sensory ganglia (Fig. 2B).

Fig. 2. Inhibition of LSD1 reduces primary acute infection in the mouse model system.

(A and B) Ocular infection with 1 × 105 plaque-forming units (PFU) of HSV-1(F) per eye. (A) Survival of mice treated topically with vehicle, ACV (5 mM), or TCP (5 mM). (B) Viral loads in ganglia at 10 dpi of mice treated intraperitoneally with vehicle, ACV (100 mg/kg), or TCP (5 or 10 mg/kg). (C to E) Intranasal infection with 1.1 × 104 PFU (one LD50) of HSV-2(MS). Mice were treated orally with vehicle, ACV (100 mg/kg), or TCP (15 mg/kg). (C) Survival chart. (D) Viral loads in ganglia at 5 and 10 dpi. (E) Viral loads in select organs at 5 dpi. %T, % of total mice positive. (F to H) Intranasal infection with 1.1 × 105 PFU (one LD90) of HSV-2(MS). (F) Survival of mice treated orally with vehicle, ACV, or TCP at the indicated levels. (G) Viral loads in ganglia at the indicated dpi in mice treated orally with vehicle, ACV (100 mg/kg), or TCP (6 mg/kg). (H) Viral loads in ganglia of mice implanted with vehicle or TCP time-release pellets (ACV, oral, 100 mg/kg). (I) Survival of mice infected intranasally with 1.1 × 105 PFU (one LD90) of HSV-2(MS) and treated orally with vehicle, ACV, TCP, or the combination of ACV and TCP. Data are means ± SEM. *Statistically significant.

As a model of neonatal infection, mice infected intranasally with a median lethal dose (LD50) of HSV-2 and treated orally with TCP showed increased survival (Fig. 2C) and reduced viral loads in sensory ganglia (Fig. 2D) and select tissues (Fig. 2E and fig. S1). Similar results were obtained when mice were infected at a stringent LD90 (lethal dose for 90% of animals) and treated orally with TCP (Fig. 2, F and G) or through insertion of time-release TCP drug pellets (Fig. 2H).

Finally, because TCP blocks viral gene expression and the initiation of infection, whereas ACV functions at the later stage of viral DNA replication, the impact of combined therapy was investigated. As shown in Fig. 2I, combined TCP and ACV treatment increased the survival of animals infected with HSV-2 intranasally (LD90) over that achieved by treatment with either compound individually. Collectively, the results indicate that inhibition of the histone demethylase LSD1 suppresses HSV in vivo, even at high levels of infection, and that combinational therapy approaches can improve the effects of the present therapeutics.

Suppression of infection in the rabbit eye model of herpetic keratitis

Primary infection of the rabbit eye provides a model of herpetic keratitis in which defined parameters of infection and clinical pathology can be assessed. Therefore, rabbits were treated orally with TCP, vehicle, or control Valtrex (VCV) and subsequently infected with HSV-1 via the ocular route. Viral yields and parameters of ocular disease (fig. S2) were assessed from 7 to 14 days post-infection (dpi) (Fig. 3). Treatment with TCP and control VCV both reduced the levels of infectious virus detected in daily eye swabs relative to the vehicle control (Fig. 3A). Furthermore, in assessment of four clinical parameters of ocular disease (Fig. 3, B to E) in addition to slit lamp examination (Fig. 3F), TCP treatment improved the clinical status of treated animals (cumulative scores, Fig. 3G).

Fig. 3. Inhibition of LSD1 reduces viral yields and ocular disease during acute infection of rabbit corneas.

(A to G) Ocular infection with 2.5 × 105 PFU of HSV-1(17syn+). Rabbits were treated orally with vehicle, Valtrex (VCV, 30 mg/kg), or TCP (12 mg/kg) from –2 to 14 dpi. (A) Levels of infectious virus detected in eye swabs collected 3 to 8 dpi. (B to G) Parameters of ocular disease were assessed at 7 to 14 dpi. (B to F) Mean values of scores for corneal stroma, ocular neovascularization, corneal epithelium, scleral inflammation/injection, and slit-lamp examinations (SLE). (G) Cumulative scores of (B) to (F). Data are means ± SEM. *Statistically significant.

Suppression of viral reactivation and shedding

In contrast to the mouse model, rabbits readily exhibit spontaneous viral reactivation and subclinical viral shedding in vivo. Thus, the rabbit ophthalmic model has been used extensively to investigate the biology of HSV recurrent disease. Previously, inhibitors of LSD1 or JMJD2 protein activities suppressed HSV reactivation in a mouse latently infected ganglia ex vivo explant system (32, 35). To investigate the effects of TCP-mediated inhibition of LSD1 on viral reactivation in vivo, rabbits were infected with HSV-1 through the ocular route and maintained for 28 days to allow resolution of the primary infection and the establishment of viral latency. Subsequently, animals were treated orally with TCP, vehicle, or VCV. Eye swabs were taken daily to assess the level of viral reactivation and shedding by detection of infectious HSV-1 and viral DNA. As shown in Fig. 4 (A and B), 24.7% of swabs from the vehicle-treated group were positive for infectious virus (Fig. 4A), whereas 36.6% were positive for viral DNA (Fig. 4B). Strikingly, whereas VCV reduced these levels to 7.1 and 17.1%, respectively, TCP treatment abrogated the detection of infectious virus and reduced the number of swabs with detectable viral DNA to 3.3%. Additionally, those swabs positive for viral DNA in TCP-treated animals had significantly reduced copy number relative to both vehicle- and VCV-treated animals.

Fig. 4. Inhibition of LSD1 blocks viral reactivation and shedding in the rabbit eye model system.

(A to H) Ocular infection with 3 × 105 PFU of HSV-1(17syn+). (A and B) Eye swabs, collected at 33 to 53 dpi, were assayed for the presence of infectious HSV-1 and viral DNA in animals treated orally with vehicle, Valtrex (250 mg/kg), or TCP (12 mg/kg). (C and D) Latently infected rabbits were implanted with placebo (vehicle) or TCP (9.5 mg/kg) time-release pellets at 48 dpi. Eye swabs collected at 52 to 68 dpi were assayed for the presence of infectious HSV-1 and represented as positive per group (C) and per rabbit (D). (E and F) Latently infected rabbits were treated with placebo or selegiline (EMSAM) time-release topical patch (5.5 mg/kg) at 17 to 47 dpi. Eye swabs collected at 20 to 47 dpi were assayed for infectious HSV-1 and represented as positive per treatment group (E) and per eye (F). (G) Viral loads in ganglia of rabbits treated with vehicle and TCP time-release pellets at 68 dpi. (H) Viral loads in ganglia of rabbits treated with vehicle or EMSAM transdermal patches at 47 dpi. *Statistically significant.

Similar results showing significant reductions in the levels of viral reactivation were obtained from animals treated by insertion of time-release TCP drug pellets (Fig. 4, C and D) or through transdermal delivery of a second LSD1 inhibitor (selegiline/EMSAM patch) (Fig. 4, E and F). In each case, the inhibitor-treated group had significantly reduced numbers of HSV-1–positive eye swabs, reduced number of recurrences (episodes), and reduced times of shedding (Fig. 4, E and F). No differences were seen in the viral loads in trigeminal ganglia of animals from the treated groups relative to the control groups (Fig. 4, G and H), indicating that TCP-mediated reductions in viral shedding and reactivation were not due to inherent variance in the ganglia loads.

Enhanced epigenetic suppression of the latent HSV-1 genome

Because inhibition of LSD1 results in enhanced epigenetic suppression of HSV-1 in vitro (32, 44), chromatin immunoprecipitation (ChIP) assays were performed on trigeminal ganglia removed from vehicle-, TCP (time-release drug pellet)–, and EMSAM-treated rabbits at the conclusion of the treatment periods. As shown, viral genomes in TCP-treated (Fig. 5A) and EMSAM-treated (Fig. 5B) animals exhibited increased levels of histones (H3) and highly enriched levels of the repressive histone H3 lysine 9 methylation (H3K9-me) associated with the viral immediate-early gene regions. In contrast, no significant changes in the levels of H3 or H3K9-me were seen on the control cellular glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene. The results illustrate that inhibition of LSD1 results in enhanced epigenetic suppression of the HSV-1 genome that translates into reduced viral shedding and reactivation in vivo.

Fig. 5. Inhibition of LSD1 results in enhanced epigenetic suppression of HSV-1.

(A and B) ChIP assays showing increased levels of histone H3 and H3K9-me at viral genomes in ganglia of time-release TCP–treated (A) and transdermal EMSAM–treated rabbits (B). (C to E) ChIP assays showing (C) increased levels of histone H3 and H3K9-me3 and (D and E) decreased levels of H3K9-me2 and H3K27-me3 on the viral genome of latently infected mice treated with time-release TCP pellets. Data are means ± SEM.

To determine whether similar epigenetic suppression would be evident in ganglia of TCP-treated mice, latently infected animals were implanted with time-release TCP drug pellets delivering different doses of compound. At 21 days after implantation, trigeminal ganglia were removed and ChIP assays were performed. As shown (Fig. 5C), TCP-treated mice exhibited increased levels of histones and H3K9-me3 that correlated with the level of TCP dosage, further indicating that TCP treatment results in enhanced epigenetic suppression of latent HSV-1 in vivo. Concomitant with these increases, TCP treatment also resulted in reduced levels of H3K9-me2 (Fig. 5D), likely due to the conversion of the dimethyl mark to the H3K9-me3 state. Strikingly, the “facultative heterochromatin” mark H3K27-me3 was also significantly reduced in TCP-treated mice (Fig. 5E), suggesting that treatment resulted in a shift in the chromatin dynamics of the latent viral pool.

Suppression of HSV-2 viral shedding and lesion recurrence in the guinea pig vaginal model

The guinea pig has been used extensively as a model of recurrent HSV genital lesions and viral shedding. Animals are infected intravaginally with HSV-2, and recurrent lesions are scored after recovery from primary disease according to defined criteria (fig. S3A). Here, infected guinea pigs were treated orally with TCP, vehicle, or ACV control from 15 to 35 dpi. Recurrent lesions were scored by severity during the treatment period and extending for an additional 16 days. As shown in Fig. 6, A to C, TCP reduced the cumulative HSV disease and recurrent lesion scores of infected animals in a manner comparable to that of the control ACV-treated group. More strikingly, vaginal swabs assessed for HSV-2 DNA demonstrated a clear reduction in the number of positive swabs and the level of viral genomes in the TCP-treated animals relative to either vehicle or ACV (Fig. 6, D and E). This difference was not due to variance in the viral loads in the dorsal root ganglia (DRGs) or spinal cords (Fig. 6F).

Fig. 6. Inhibition of LSD1 reduces recurrent disease in the guinea pig model.

(A to F) Vaginal infection with 1 × 106 PFU of HSV-2(MS). Guinea pigs were treated orally with vehicle, ACV (5 mg/ml ad libitum in water), or TCP (15 mg/kg per day) from 15 to 35 dpi. (A to C) Lesions were scored (26 to 52 dpi) and shown as (A) cumulative average disease and recurrent lesion scores per day as defined in fig. S3A, (B) average lesion scores per day, and (C) cumulative lesion scores per guinea pig. (D and E) HSV DNA detected in vaginal swabs taken from 21 to 35 dpi. (F) Viral loads in DRGs and spinal cords at 52 dpi. (G to J) Vaginal infection with 2 × 105 PFU of HSV-2(MS). Guinea pigs were treated intraperitoneally with ACV (6 mg/kg per day), TCP (5 mg/kg per day), or the combination for 22 to 43 dpi. (G to I) Lesion recurrences were scored as defined in fig. S3B and are shown as (G) cumulative average recurrences, (H) total recurrences per guinea pig, and (I) average recurrences per group per day. (J) Viral loads in DRGs at 43 dpi. (K and L) ChIP assays showing increased levels of histone H3 and H3K9-me at viral promoters in dorsal root ganglia of guinea pigs (K) treated orally with vehicle or TCP (30 mg/kg per day) and (L) treated intraperitoneally with ACV (6 mg/kg per day) or TCP (5 mg/kg per day). Data are means ± SEM. *Statistically significant.

As demonstrated in the mouse model system (Fig. 2I), the combination of TCP and ACV treatment increased the survival rate to a greater extent than did either compound alone. In a similar manner and on the basis of the previous experiment (Fig. 6A), combined ACV and TCP treatment in the guinea pig vaginal model further reduced the cumulative HSV-2 recurrences as compared to either compound alone (Fig. 6, G to I, and fig. S3B). No differences in the viral loads between the groups were detected in the dorsal root ganglia (Fig. 6J).

Consistent with the effects of TCP on the chromatin status of viral genomes in both the mouse and rabbit trigeminal ganglia, treatment of infected guinea pigs with TCP (30 mg/kg per day) resulted in enhanced histone levels and repressive heterochromatic marks associated with the viral genomes in the dorsal root ganglia (Fig. 6K). In a second study, qualitatively comparable results were also seen in animals treated with a lower (5 mg/kg per day) TCP dose (Fig. 6L) when compared to the control ACV.

DISCUSSION

Despite the availability and widespread use of pharmaceutical inhibitors for the treatment of herpesvirus infections, HSV remains a clinically important pathogen. A high percentage of the human population is infected and carries the virus in a latent state in sensory neurons, whereas asymptomatic shedding of the virus maintains the transmission rate. Neonatal HSV infections can result in mortality or continued developmental-neurological issues, and HSV ocular infections and recurrences are the leading virus-mediated cause of blindness or the requirement for corneal transplants. In addition, along with other herpesviruses, HSV is a complicating factor in immunosuppressed individuals and is a cofactor in HIV transmission. Presently, the primary antiherpetic pharmaceuticals target virus-encoded DNA replication machinery during a late stage in the infection cycle. However, these compounds do not efficiently suppress viral shedding, subclinical reactivation, and lytic recurrence. Additionally, although the incidence of resistant virus populations is low, the frequency is increased in immunosuppressed patients and those with herpetic keratitis. Furthermore, recent studies indicate that prophylactic ACV treatment does not reduce the HSV-mediated enhancement of HIV transmission (46, 47). Thus, it is clear that new approaches to address deficiencies in the prevention of transmission and treatment of recurrence are needed.

Together, the recognition that viruses such as HSV are subject to epigenetic control (27) and the emerging development of epigenetic pharmaceuticals open new avenues for developing antivirals. Inhibitors of chromatin modification enzymes required for the initiation of viral infection can prevent viral gene expression, reduce the potential for inflammation, and reduce the evolution of drug-resistant viral strains. Here, epigenetic suppression of HSV infection, shedding, reactivation, and clinical recurrence was demonstrated in the primary animal models of HSV disease, using pharmaceutical inhibition of the histone demethylase LSD1.

LSD1 is a component of the regulatory paradigm that mediates the initiation of HSV infection by modulating the epigenetic state of the virus to allow the first wave of viral gene expression. The enzyme is specifically recruited to viral gene promoters upon infection, and the catalytic demethylation activity is required to drive the chromatin-mediated repressed state of the genome to an active euchromatic state. TCP and other monoamine oxidase inhibitors (MAOIs) inhibit the activity of LSD1, and these compounds block infection in vitro and viral reactivation in the latently infected mouse ganglia explant model.

Mechanistically, inhibition of LSD1 results in accumulation of repressive methylation on histones associated with the viral genome and a block to viral gene expression. Notably, a specific LSD1 inhibitor (OG-L002) that does not inhibit the cellular monoamine oxidase A (MAO-A) or monoamine oxidase B (MAO-B) was also effective in reducing HSV lytic infection and reactivation (44). In contrast, a compound that targets the cellular MAOA and MAOB but does not inhibit LSD1 had no effect on viral infection. Thus, LSD1 represents a potentially important therapeutic target for antiherpetic therapeutics.

Here, TCP treatment of HSV-infected mice resulted in reduced pathology as observed by reduced mortality, reduced accumulation of genomes in sensory ganglia, and reduced viral loads in tissues. Reduction in viral pathology was evident in animals infected with high doses (LD50 or LD90 doses) and using a variety of inhibitor treatments (oral, time-release, topical, intraperitoneal). Indeed, in support of these data, Yao et al. (48) reported the use of TCP to suppress HSV infection in the mouse model system. Similarly, in the rabbit eye model of herpetic keratitis, treatment of primary infection reduced the critical clinical indicators of ocular disease.

Although suppression of HSV primary infection in these experiments was effective, TCP treatment resulted in a marked reduction in viral shedding and reactivation or recurrences. This was clearly evident in the rabbit model of spontaneous reactivation and shedding. In this model, TCP was significantly more effective than VCV treatment.

Finally, in the guinea pig vaginal model of HSV-2 recurrent disease, inhibition of LSD1 resulted in suppression of recurrent lesions as effectively as the control ACV and suppression of recurrent shedding more effectively than ACV. Together, these studies demonstrate suppression of HSV lytic infection, shedding, and reactivation that correlated with enhanced epigenetic suppression of the viral genome.

TCP inhibition of LSD1 targets the initiation stage of viral lytic infection and reactivation. This is distinct from the predominant antiherpetic pharmaceuticals, nucleoside analogs that target the later stage of viral DNA replication. Consistent with these two distinct mechanisms, combined treatment (TCP and ACV) resulted in enhanced suppression over the use of either compound alone. The results are more striking because the combination treatment used lower concentrations of the two compounds relative to the individual compound treatments. The results suggest that combination therapies targeting distinct stages of the HSV infectious cycle may be clinically advantageous with respect to the present antiviral therapies and may also reduce the propensity for the generation of ACV-resistant viruses in the population.

The increased epigenetic silencing of the HSV genome, in sensory ganglia where the viral genome is latent, is consistent with the antiviral effects of the LSD1 inhibitors. Strikingly, this argues that the epigenetic status of latent viral genomes is dynamic, with histone marks being continuously modulated. Whereas the levels of repressive H3K9-me3 increased upon TCP treatment, the levels of the “facultative heterochromatin mark” H3K27-me3 decreased, suggesting a shift in the heterochromatin status of the population of viral genomes. This dynamic nature of HSV chromatin is consistent with recent studies of HIV where latent infection involves epigenetic silencing (49), and reactivation is a stochastic process (50, 51), involving “bursts” of reactivation that may proceed forward to a full reactivation (52).

The sensitivity of the state of the viral genome to LSD1 inhibition, as compared to the host GAPDH gene, may reflect the critical nature of this enzyme for HSV gene expression relative to other LSD1 targets. LSD1 demethylates both histone H3K4me1/2 and H3K9me1/2. However, the target selectivity and specificity is determined by cofactors. Additionally, multiple enzymes modulate H3K9-me levels in the cell, and thus, LSD1 inhibition would not be expected to have global effects. The specific effects on HSV gene expression may be conceptually similar to the regulation of cellular super-enhancers where it has been demonstrated that specific loci are surprisingly hypersensitive to a critical regulatory factor (5355).

Investigation of the role of chromatin and specific chromatin modulation components in the control of viral infection, states of latency, and reactivation offers targets for antiviral therapies. In addition to HSV, human cytomegalovirus and adenovirus have also been shown to rely on the HCF-1/LSD1/JMJD2 transcriptional coactivator complex (35, 44). Therefore, compounds targeting these components may have activity against other viral pathogens in addition to HSV.

Epigenetic pharmaceuticals represent potential antivirals to control viral infection and reactivation of latent genomes. Inhibitors of LSD1 (OG-L002) as well as inhibitors of the JMJD2 demethylases (ML324) have also been shown to suppress HSV infection and reactivation in vitro (35, 44). Thus, there is potential to target multiple components to increase efficacy.

In the in vivo studies presented here, the MAOIs TCP and selegiline were chosen to investigate the effects of LSD1 inhibition on HSV infection, shedding, and reactivation in vivo on the basis of multiple parameters. These compounds are very well-characterized clinical pharmaceuticals and are available in multiple formats (oral compound, time-release subcutaneous drug administration pellets, and time-release transdermal patches). The characteristic TCP clinical doses range from 40 to 60 mg/day to 170 mg/day for refractive depression in patients. Although the dose range used in these studies exceeds the levels used for the treatment of depression, this was anticipated because the IC50 (half-maximal inhibitory concentration) for LSD1 is higher than for the targets MAOA and MAOB. In the studies presented here, no toxicity was exhibited in the three animal model systems even at the highest dosage, and lower doses also reduced viral infection and reactivation. Furthermore, advances in MAOI therapy, such as the transdermal or topical applications demonstrated here, have reduced the characteristic side effects of these compounds. However, it remains important to consider the potential issues that arise from the off-target effects of these compounds or of LSD1 inhibition in long-term treatment. Specific LSD1 inhibitors are in development for the treatment of various cancers. At this time, these inhibitors are less well characterized than the MAOIs but could represent antivirals that would have advantages over the nonselective MAOI compounds.

Overall, the described studies represent a comprehensive analysis of epigenetic inhibition of viral infection and reactivation from latent pools. This is in contrast to approaches used to attempt to “purge” latent HIV reservoirs by epigenetically inducing viral reactivation (5658). Thus, epigenetic suppression can represent an approach to antivirals with extensive potential derived from studies of viral epigenetics and the emergence of the field of epi-pharmaceuticals.

MATERIALS AND METHODS

Study design

Three animal model systems of herpetic disease were used to study the potential for epigenetic suppression of HSV infection, subclinical shedding, and spontaneous recurrence using an inhibitor of the histone demethylase LSD1. Experiments were designed to use the most appropriate animal model for each aspect of HSV infection (primary infection, mouse model; spontaneous reactivation, rabbit eye model; clinical recurrence, guinea pig vaginal model).

Sample sizes for all experiments were based on standards ascribed to the National Institute of Allergy and Infectious Diseases (NIAID) Anti-Viral Program and the specific expertise of each contributing research group. For rabbit and guinea pig experiments, infected animals were randomized to study groups on the basis of initial acute disease. After infection in the rabbit eye model, slit-lamp examination was used to determine the extent of corneal damage, and those eyes with extensive damage were discarded from the study. Animal groups were not blinded.

Animals and infections

Mice. Ocular infections: Six- to 7-week-old Balb/c mice were infected with 1 × 105 PFU of HSV-1(F) per eye after corneal scarification and treated topically or via intraperitoneal injection as described in the appropriate figure legend and table S1.

Intranasal infections: Four-week-old Balb/c mice were infected intranasally with 1.1 × 104 PFU (one LD50) or 1.1 × 105 PFU (one LD90) of HSV-2(MS) and treated orally or via insertion of time-release drug pellets as described in the appropriate figure legend and table S1.

Rabbits. Latency: New Zealand white (NZW) rabbits (McNeil Rabbitry; 2.0 to 2.5 kg) were infected with 2.5 × 105 to 3 × 105 PFU of HSV-1 (17 syn+ or McKrae) per eye after corneal scarification. Infection was verified by slit-lamp examination at 3 dpi. Latently infected animals (>17 dpi) were assessed, and those with normal corneas and ocular tissues were used. Rabbits were treated orally, via intraperitoneal injection, via insertion of time-release drug pellets, or via topical drug patches as described in the appropriate figure legend and table S1.

Acute infection: NZW rabbits were infected as detailed above and treated as described in the appropriate figure legend. Ocular disease was assessed according to clinical parameters detailed in fig. S2.

Eye swabs: Rabbit eye swabs were analyzed for infectious HSV-1 or viral DNA as described (59).

Guinea pigs. TCP efficacy: Female Hartley guinea pigs (15 per group, 250 to 350 g, Charles River) were inoculated intravaginally with 1 × 106 PFU of HSV-2(MS) by rupturing the vaginal closure membrane with a moistened calcium alginate tipped swab and instilling 0.1 ml of a virus suspension into the vaginal vault (60). Vaginal swabs were obtained at 2 dpi to assess infection. After recovery from primary infection (15 dpi), animals were randomized on the basis of acute disease. Animals were treated orally (15 to 35 dpi) with vehicle, TCP (15 mg/kg per day, 7.5 mg/kg twice daily), or ACV (5 mg/ml ad libitum in drinking water) and monitored for recurrent herpetic lesions (15 to 56 dpi) and recurrent viral shedding (21 to 35 dpi). Experimental details and scoring criteria are defined in table S1 and fig. S3A, respectively. Dorsal root ganglia and spinal cords were collected from each animal at 57 dpi and processed for viral DNA loads. In a second study, animals were treated orally (15 to 35 dpi) with vehicle, TCP (30 mg/kg per day), or ACV (5 mg/ml in drinking water). At 35 dpi, dorsal root ganglia were collected for ChIP analyses.

Viral shedding: Vaginal viral swabs were collected three times per week (21 to 35 dpi). HSV-2 DNA levels were determined by quantitative polymerase chain reaction (qPCR) as described (61).

TCP-ACV combination: Female Hartley guinea pigs (Charles River) were inoculated intravaginally with 2 × 105 PFU of HSV-2(MS). At 22 dpi, animals were randomized on the basis of acute disease. Animals were treated intraperitoneally (22 to 43 dpi) with TCP (5 mg/kg per day), ACV (6 mg/kg per day), or TCP and ACV (5 and 6 mg/kg per day, respectively). Recurrent lesions, defined as new lesions on the external genitalia, were scored (22 to 43 dpi). Experimental details and scoring criteria are defined in table S1 and fig. S3B, respectively. Dorsal root ganglia were collected from each animal immediately after sacrifice and processed for viral DNA loads and ChIP assays.

Animal care

Mice and guinea pigs were housed in American Association for Accreditation of Laboratory Animal Care–approved facilities. Rabbit experimental procedures were performed in accordance with the Association for Research in Vision and Ophthalmology (ARVO) resolution for the use of animals in ophthalmic and vision research. All animal care and procedures were done according to Institutional Animal Care and Use Committee–approved animal protocols and in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. All animals were monitored by experienced animal care staff and by the responsible research scientists. No indication of MAOI-mediated toxicity (that is, weight loss, persistent hyperexcitability or behavioral changes, consumption of food and water, morbidity, or mortality) was evident in these experiments.

Compounds

The following compounds were used in these studies: TCP, trans-2-phenylcyclopropylamine hydrochloride; SEL, selegiline, R-(–)-deprenyl hydrochloride; ACV, acyclovir, acycloguanosine (Sigma-Aldrich); VCV-Valtrex, valacyclovir (GlaxoSmithKline Pharmaceuticals); TCP and placebo time-release pellets (Innovative Research of America); EMSAM, selegiline transdermal patches (Somerset Pharmaceuticals Inc.).

Viral loads

Total genomic DNAs from mouse, rabbit, and guinea pig tissues were isolated and analyzed according to standard procedures as described (32, 6163). Levels of viral DNA were determined by qPCR and are represented per microgram of total DNA or as the ratio to the levels of the cellular GAPDH gene. The sequences of the relevant primer sets are listed in table S3.

ChIP assays

ChIP assays were done from latently infected mouse, rabbit, and guinea pig sensory ganglia essentially as described (39, 42), with minor modifications detailed in Supplementary Materials and Methods. Antibodies and primer sets are listed in table S3.

Statistical analyses

Data are expressed as means ± SEM. Statistical comparisons were made using Prism (V6.0) and included unpaired two-tailed t tests with a statistical significance of <0.05, ANOVA (analysis of variance) with post hoc Dunnett’s multiple comparison test with a statistical significance of <0.025, Holm-Sidak multiple t tests, and linear regression slope analyses. Additional details are given in the appropriate figure legends and in table S2.

SUPPLEMENTARY MATERIALS

www.sciencetranslationalmedicine.org/cgi/content/full/6/265/265ra169/DC1

Materials and Methods

Fig. S1. Viral loads in select organs of mice treated with ACV or TCP at 10 dpi.

Fig. S2. Scoring criteria for ocular disease in the rabbit eye model.

Fig. S3. Scoring criteria for HSV disease in the guinea pig model.

Table S1. Experimental conditions.

Table S2. Statistical analyses.

Table S3. Antibodies and primers.

Supplementary file. Numerical values of experimental data.

REFERENCES AND NOTES

  1. Acknowledgments: We thank J. Arbuckle and T. Pierson for editorial comments; J. Skinner (Laboratory of Immunogenetics, NIAID) for statistical analyses; NIAID Bld33 Animal Facility staff (NIH), L. L. Birke and M. Johnson [Louisiana State University Health Science Center (LSUHSC) Animal Care], D. Pullum and H. Shen (University of Cincinnati), R. Crosby II and M. Simpson (LSUHSC); D. J. Collins and T. L. Rice (University of Alabama Birmingham) for technical assistance; and J. Arbuckle (NIAID, NIH) for assistance to T.M.K. with mouse studies. Portions of this work, including murine studies, were presented at the 2012 and 2013 International Herpesvirus Workshops and the 2012 and 2013 International Conference on Antiviral Research. Funding: These studies were supported by the Laboratory of Viral Diseases, Division of Intramural Research, NIAID, NIH (T.M.K.); contract HHSN272201000027I, Task Order HHSN2720001 (D.C.Q.); National Eye Institute R01-EY006311 (J.M.H.), Research to Prevent Blindness (RPB) Senior Investigator Award (J.M.H.), RPB unrestricted grant to the LSU Eye Center (J.M.H.), LSUHSC unrestricted grant to J.M.H.; U.S. Food and Drug Administration Intramural Funding (P.R.K.); contract HHSN27220100008I (D.I.B. and R.D.C.); and NIH P01-AI098681 (D.M.K.). Author contributions: Mouse studies: D.C.Q., J.L.V., and T.M.K. designed the studies and analyzed the experimental results; D.C.Q. conducted the intranasal infection therapeutic studies; T.M.K. and J.L.V. performed the mouse ocular infection studies; J.L.V. performed viral loads; D.M.K. and T.M.K. designed the mouse ChIP experiments and analyzed the experimental results; P.R. and J.S.L. performed mouse ganglia ChIP assays; T.M.K. performed mouse infections, drug implants, and ganglia preparation. Rabbit studies: J.M.H., C.C., T.P.F., and T.M.K. designed the studies and analyzed the experimental results. J.M.H., C.C., and T.P.F. conducted the studies. J.L.V. performed viral yields and rabbit ganglia ChIP assays. Guinea pig therapeutic studies: D.I.B., R.D.C., and T.M.K. designed and analyzed the therapeutic studies. D.I.B. and R.D.C. were responsible for study oversight and analyses. F.J.B. conducted the treatment and monitored the guinea pigs. J.L.V. performed guinea pig ChIP assays. Guinea pig combination treatment study: M.B.-M., P.R.K., and T.M.K. designed the study and analyzed the experimental results. M.B.-M. and P.R.K. conducted the study. J.L.V. performed guinea pig ChIP assays. T.M.K. assembled all data and wrote the manuscript. All authors commented on the manuscript prior to publication. Competing interests: The NIH has the following patent applications: (i) Methods of preventing or treating viral infection or reactivation from latency in a host using inhibitors of the LSD1 protein (T.M.K., J.L.V., and Y.L.; U.S. patent application no. 61/083,304; International patent application no. PCT/US2009/051557); (ii) Method of preventing or treating viral infection via inhibition of the JMJD2 proteins (T.M.K., Y.L., and J.L.V.; U.S. patent application no. 61/366,563). All other authors declare no competing interests.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article