Nuclear Translocation of NF-κB Precedes Apoptotic Poly(ADP-ribose) Polymerase Cleavage during Productive HSV-1 Replication in Corneal Epithelial Cells

Abstract

purpose. Herpes simplex virus (HSV)-1 infections of the human cornea range in severity from uncomplicated episodes that readily resolve to severe, recurring disease that invades the stroma, having a devastating permanent effect on vision. Recent published data implicate an apoptotic component to stromal HSV-1 infection. In a prior study, it was found that wild type (wt) HSV-1 infection induces, then blocks, apoptosis in epithelial cells derived from skin and that this block requires infected cell proteins (ICPs) synthesized between 3 and 6 hours post infection (hpi). This inhibition of apoptosis is in part dependent on the activation of inducible nuclear transcription factor κB (NF-κB).

methods. HSV-1-dependent apoptosis in rabbit corneal epithelial (SIRC) cells was compared with that in infected human epithelial (HEp-2) cells.

results. SIRC cells were sensitive to apoptotic cell death induced by environmental treatment with tumor necrosis factor (TNF)-α plus cycloheximide (CHX). HSV-1 stimulated the degradation of regulatory IκBα protein, resulting in nuclear translocation of NF-κB. This phenomenon was dependent on ICP synthesis. Neither wt nor apoptotic HSV-1 infection resulted in apoptosis in these cells. However, wt HSV-1-infected cells produced detectable levels of cleaved poly(ADP-ribose) (PARP). Inhibition of SIRC cell protein synthesis with CHX during wt HSV-1 infection led to a reduction in the amount of PARP cleavage. Whereas PARP cleavage defined cell death in most other cell types, its processing in SIRC cells was a reproducible characteristic of wt HSV-1 infection.

conclusions. This is the first report of such an effect, and it suggests that in corneal epithelial cells, activation of apoptotic pathways may be necessary for productive viral replication. Thus, efficient replication of HSV-1 in the corneal milieu proceeds via a different mechanism than it does in skin. However, it appears that NF-κB participates in inhibiting apoptosis during HSV-1 infection in both systems.

Herpes simplex virus (HSV)-1 infection is one of the major causes of ocular morbidity and blindness in the developed world. Scarring after severe or repeated infections may necessitate corneal transplantation to restore useful vision to the patient (see review). ¹ Current research efforts have been focused on defining the mechanisms of herpetic eye disease, including ways in which viral infection of ocular structures differs from infection of the skin and mucous membranes. One new concept under increased scrutiny is the role that apoptosis plays in herpetic eye disease. Traditionally, HSV-1 infection was assumed to stem from successful viral replication leading to lysis of infected cells. However, it now appears that there is probably an apoptotic component as well. Human corneal epithelial cells obtained from patients with herpetic eye disease are positive for certain markers of apoptosis. ² Consistent with these observations, ocular HSV-1 infection also causes apoptosis in mouse ^{3 4} and rabbit ⁵ model eye systems. These findings suggest that modulation of apoptosis by HSV-1 plays a role in the development of herpetic eye disease (see review). ⁶  

Most of what is known about apoptosis during HSV-1 infection is derived from studies using the prototype human epithelial HEp-2 cell system (see review). ^{7 8} Briefly, stimulation of apoptosis by wild-type (wt) HSV-1 is independent of de novo viral protein synthesis, ^{9 10} requiring only transcription of the viral immediate early ICP0 gene. ^{9 11 12} If left unchecked, this apoptosis would cause accelerated destruction of the infected cell, leading to decreased viral replication efficiency. However, infected cell proteins (both viral and cellular) are produced during a cell death “prevention window,” between 3 and 6 hours after wt HSV-1 infection, and they block the process from killing the cells. ⁹ Thus, HSV-1 viruses that do not produce these prevention factors, either as a result of pharmacologically blocking their synthesis or by possessing specific genetic mutations, are apoptotic. A cellular factor that participates in the inhibition of HSV-1-dependent apoptosis is the inducible nuclear transcription factor-κB (NF-κB), ¹³ a key regulator of immunity, growth, differentiation, and fate of mammalian cells (see review). ^{14 15 16 17} Thus, wt, but not apoptotic, HSV-1 infection induces the nuclear translocation and DNA binding activities of NF-κB. 

Our goal was to determine whether HSV-1 modulates apoptosis in corneal cells as it does in the skin and whether NF-κB plays a role in its prevention in the corneal milieu. Although we expected that infection would proceed as we had previously found with human HEp-2 cells, we discovered that SIRC cornea cells are resistant to HSV-1-induced apoptosis. Although SIRC cells undergo apoptosis in response to environmental stimulants, infection with wt HSV-1 in the presence of a protein synthesis inhibitor or infection with an apoptotic HSV-1 did not lead to programmed cell death. Unexpectedly, we discovered that PARP, generally a marker of apoptotic processing, was cleaved during wt HSV-1 infection. Coincident with this, NF-κB translocated to the nuclei, and both effects were dependent on infected cell protein synthesis. These findings indicate that, although HSV-1 modulates NF-κB in both the skin and corneal cells, its replication in the cornea may proceed via a different mechanism than it does in skin. The implication of our results is that the activation of apoptotic pathways may be necessary for efficient viral replication in corneal epithelial cells. 

Materials and Methods

Cell Lines and Viruses

HEp-2 and SIRC cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA). HEp-2 cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) containing 5% fetal bovine serum (5% FBS). SIRC cells were maintained in SIRC medium, composed of Eagle’s minimum essential medium with 2 mM l-glutamine and balanced saline (BSS; Alcon, Fort Worth, TX) adjusted to 1.5 g/L sodium bicarbonate, 0.1 mM nonessential amino acids, 1.0 mM sodium pyruvate, and 10% fetal bovine serum. The wt viruses used in these experiments were the HSV-1(F) virus, which was obtained from Bernard Roizman (University of Chicago) and the HSV-1(KOS1.1) virus which was obtained from Saul Silverstein (Columbia University). An ICP27-null mutant HSV-1 virus, vBSΔ27, was also obtained from Saul Silverstein. This virus was derived from HSV-1(KOS1.1) and contains a deletion of the α27 gene, replaced by the Escherichia coli lacZ gene ¹⁸ and was propagated in Vero 2.2 cells (Saul Silverstein). The HEp-2 and SIRC cell monolayers were infected with a multiplicity of infection (MOI) of 10 PFU/cell. This multiplicity of 10 insures that all cells are infected within the single step of virus growth. However, we have observed essentially identical results at MOIs of 5 and 1 (data not shown), indicating that our findings are not a consequence of viral multiplicity. HSV-1 inoculations were performed in medium 199 containing 1% bovine serum albumin (BSA; 199V) at 37°C for 1 hour. After adsorption, the infected cells were maintained in the appropriate medium for the remainder of the experiment, as indicated in the text. All tissue culture reagents were purchased from Invitrogen-Life Technologies (Gaithersburg, MD). 

Pharmacological Inhibition of Protein Synthesis in Infected HEp-2 and SIRC Cells

At the specified times, cycloheximide (10 μg/mL; CHX) in the appropriate medium was added to the infected cells and maintained until the end of the experiment. As a positive control for the induction of apoptosis, TNFα (10 ng/mL) and CHX were added to mock infected HEp-2 or SIRC cells. Lyophilized TNFα (Sigma-Aldrich, St. Louis, MO) was dissolved in sterile PBS containing 1% BSA. Aliquots of TNFα stocks at a concentration of 10 μg/mL were stored at −80°C and thawed once before using. 

Preparation of Infected Cell Extracts

The infected HEp-2 and SIRC cells were harvested by scraping directly into the medium. The cells were centrifuged at low speed (8000g), after which the cell pellets were washed in phosphate-buffered saline (PBS) containing 10 mM each of the protease inhibitors N-tosyl-l-phenyl-alanine-chloromethylketone (TPCK), phenylmethylsulfonyl fluoride (PMSF), and tosyl-l-lysine-chloromethylketone (TLCK). Cell pellets were resuspended in a lysis solution of 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 5 mM EDTA containing TPCK, TLCK, and PMSF. The infected cells were then disrupted by sonication on ice, three times for 10 seconds at an output level of 2 with a sonifier (Branson; Danbury, CT). The protein concentration of each sample was determined by a modified Bradford protein assay, as recommended by the manufacturer (Bio-Rad Laboratories, Hercules, CA). All biochemical reagents were obtained from Sigma-Aldrich, unless stated otherwise. 

Preparation of Cytoplasmic and Nuclear Fractions

The subfractionation of infected cells into nuclear and cytoplasmic extracts was described previously. ^{13 19} Cytoplasmic and nuclear extracts were prepared at the times after infection indicated in the text. All manipulations were performed on ice. Infected cells were scraped directly into the medium and pelleted by low-speed (8000g) centrifugation for 3 minutes. The cell pellets were washed once with PBS, transferred to a 1.5-mL tube, and pelleted again in a microcentrifuge at very low speed (1000g) for 3 minutes. The cells were lysed by gentle resuspension in 150 μL PBS containing 0.4% NP40 followed by centrifuging at very low speed for 3 minutes. The supernatants (cytoplasmic extract) were removed and placed in fresh tubes. The nuclear pellets were washed once with 100 μL PBS containing 0.1% NP40, pelleted at very low speed for 3 minutes, and resuspended in 100 μL of PBS containing 0.4% NP40. Nucleic fractions were frozen at −80°C for at least 30 minutes and thawed on ice to increase extraction of nuclear proteins from the insoluble material. The nuclear samples were then sonicated on ice three times for 5 seconds each to create the final nuclear fractions. In experiments in which both whole-cell extracts and subcellular (cytoplasmic and nuclear) fractions were prepared from the same cells, infected cells were resuspended in a total of 250 μL PBS. Of this, 50 μL was processed for whole extracts and 200 μL were processed for subcellular fractionation. 

Denaturing Gel Electrophoresis and Immunoblot Techniques

Approximately 50 μg of infected cell proteins from whole-cell extracts or subcellular fractions were electrophoretically separated in 15% SDS-polyacrylamide gels cross-linked with N,N′ diallyltartardiamide and electrically transferred to nitrocellulose in a tank apparatus (Bio-Rad Laboratories), then probed overnight with one or more primary antibodies. Secondary goat anti-rabbit and anti-mouse antibodies conjugated to alkaline phosphatase were obtained from Southern Biotechnology (Birmingham, AL). Anti-rabbit or-mouse antibodies conjugated to horseradish peroxidase was used for chemiluminescence detection, as recommended by the vendor (GE Healthcare, Piscataway, NJ). Specific proteins were detected after development with chemiluminescence reagents (GE Healthcare) and autoradiography at 25°C on autoradiograph film (X-OMAT Eastman-Kodak, Rochester, NY). Prestained molecular weight markers (Invitrogen-Gibco, Grand Island, NY) were included in all acrylamide gels (data not shown). Equal protein loading and integrity of subcellular fractions were determined by ponceau S staining ¹³ before immunoblot analysis (data not shown). 

Immunologic Reagents

Antibodies used to detect viral and cellular proteins were as follows. ICP4 was detected with the specific mouse monoclonal antibody 1114 (Rumbaugh-Goodwin Institute for Cancer Research, Plantation, FL). Mouse monoclonal antibodies specific for the p65 subunit of NF-κB, IκBα, caspase 3, and poly(ADP-ribose) polymerase (PARP) were obtained from Santa Cruz Biotechnology (Santa, Cruz, CA), Imgenex (San Diego, CA), BD-Transduction Laboratories (Lexington, KY), and BD-Pharmingen (San Diego, CA), respectively. Our PARP antibody reacts with both the full-length 116,000-molecular-weight form and the 85,000-molecular-weight cleavage product. ⁹ Our caspase 3 antibody is capable of recognizing both the uncleaved, inactive form (procaspase 3, 32 kDa) and one of the processed forms (20 kDa); however, under our electrophoresis conditions, only the uncleaved form may be resolved. Thus, we used loss of caspase 3 reactivity to demonstrate apoptosis induction. ^{9 20}  

Immunofluorescence, Microscopy, and Computer Analysis

Hoechst 33258 dye (0.05 μg/mL; Sigma-Aldrich) was added to cells 1 hour before harvesting, to visualize chromatin condensation. ^{21 22} Fluorescence was observed with an inverted fluorescence microscope (IX70/IX-FLA; Olympus, Melville, NY), and images were acquired with a digital camera (DKC-5000; Sony, Tokyo, Japan). The images were processed with image-analysis software (Photoshop; Adobe, San Jose, CA) Immunoblots were digitized at 600 dots per inch with a scanner (Arcus II; AGFA, Ridgefield Park, NJ). Raw digital images were saved as tagged image files (TIF; Photoshop; Adobe) and were organized into figures (Illustrator; Adobe). 

Results

SIRC Epithelial Cells Undergo Apoptosis Induced by Extrinsic Stimulation

We have reported that human HEp-2 cells undergo apoptosis induced by the administration of TNFα+CHX. ¹³ TNFα is a potent apoptotic stimulator, which in the presence of CHX, induces most human cells to die via caspase-dependent apoptosis due to the absence of NF-κB-dependent prevention factors. ^{16 23} Our first goal was to determine whether rabbit corneal SIRC epithelial cells were susceptible to this pharmacological instigation of apoptosis. SIRC cells were treated with TNFα+CHX for 18 hours at which time whole-cell extracts were prepared, separated in a denaturing gel, electrically transferred to nitrocellulose, and probed with anti-PARP antibodies, as described in the Materials and Methods section Cleavage of the intact 116,000-molecular-weight PARP protein to its 85,000-molecular-weight processed form is regarded as one of the hallmarks of apoptosis ²⁴ in human HEp-2 cells. ⁹ We observed that treatment of SIRC cells with TNFα+CHX led to cleavage of PARP, generating the 85,000-molecular-weight degradation product (Fig. 1 , lanes 2 and 4). As expected, untreated cells did not show significant PARP cleavage. These findings indicated that SIRC cells are capable of succumbing to apoptosis induction by environmental stimulants in the manner previously described for human HEp-2 cells. Thus, the cells used in this study were not senescent and did not have a general resistance to apoptosis induction. 

NF-κB Translocates to the Nuclei of Infected SIRC Cells

Because SIRC epithelial cells respond to extrinsic apoptotic cell death stimulation (Fig. 1) , we next investigated in more detail the process of signaling through the TNF receptor. The extrinsic activator, TNFα, is a cytokine that both induces apoptosis and activates NF-κB. In human epithelial HEp-2 cells, the apoptotic effects of TNFα manifest only during cotreatment with CHX. ^{13 25} In addition, CHX prevents the synthesis of NF-κB’s inhibitor, IκBα. ^{16 26} Therefore, NF-κB cannot be cycled back out into the cytoplasm and remains in the nucleus. ¹³ We set out to determine the response of SIRC cells to various TNFα treatments and to viral infection. Two sets of experiments were performed. In the first set, SIRC cells were untreated; treated with TNFα, CHX, or TNFα+CHX; or infected with HSV-1(F). The cells were maintained for 18 hours, at which time the cultures were treated with Hoechst DNA dye and visualized with phase-contrast and fluorescence microscopy, as described in the Materials and Methods section. 

Figure 2Ashows the morphologies of treated or infected cells in this experiment. Mock-treated cells appeared flat and elongated, with large, round nuclei. HSV-1(F)-infected cells had smooth, rounded edges in phase-contrast microscopy, characteristics of wt virus cytopathic effect, and their nuclei were large and round (Hoechst). Control cells treated with TNFα or CHX alone had morphologies similar to that of mock-treated cells. However, TNFα+CHX-treated SIRC cells had a shrunken appearance with membrane blebbing, characteristic of apoptotic death. Their nuclei were fragmented with condensed chromatin, imparting the bright staining observed (Hoechst). Combined with the findings in Figure 1 , these results indicate that SIRC epithelial cells are sensitive to apoptotic cell death triggered through the TNFα receptor. In addition, infection with wt HSV-1 does not seem to result in any significant apoptotic SIRC cell morphologies. 

In the second series of experiments, we determined whether NF-κB translocation to nuclei was activated after the treatments shown in Figure 2A . SIRC cells were mock HSV-1(F) infected or treated with TNFα and CHX, alone and in combination. At 18 hours, nuclear and cytoplasmic fractions were prepared from the cells and immunoblotted for NF-κB, as described in the Materials and Methods section. As expected, NF-κB localized to the cytoplasm in most of the untreated, mock-infected cells (Fig. 2B ; lane 1). Control CHX-treated cells (lane 5) were similar to these untreated cells. We did not observe NF-κB in the nuclei of TNFα-treated cells (compare lanes 3 and 4). Although treatment with TNFα is known to stimulate NF-κB translocation in many cell types, this effect is transient due to the recycling of IκB, ^{16 26} and so this finding was expected. Treatment with TNFα and CHX showed two things. First, very little NF-κB staining was detected, probably because of the treated cells’ dying at this time point, such that intracellular membranes breakdown and proteins, such as NF-κB, were degraded. Thus, TNFα stimulated apoptosis and CHX prevented the synthesis of the survival proteins needed to prevent the cell death process. The second finding was that the detected NF-κB was cytoplasmic. Although this may also be a consequence of cell death, we cannot exclude the possibility that NF-κB nuclear translocation does not occur in these cells in response to TNFα and CHX treatment. In contrast, we observed that the bulk of NF-κB in the HSV-1(F)–infected cells partitioned in the nuclear fraction (Fig. 2B ; compare lanes 9 and 10). Based on these results, we conclude that wt HSV-1 infection stimulates a sustained translocation of NF-κB to nuclei. This is consistent with earlier observations in other cell types. ^{13 27}  

Resistance of SIRC Cells to HSV-1-Dependent Apoptosis

The results above (Figs. 1 and 2)indicate that NF-κB is activated during wt HSV-1 infection and suggest that these infected cells are nonapoptotic. We next set out to determine the mechanism through which the virus induces NF-κB in these cells. Since we showed that SIRC cells are sensitive to TNFα and CHX treatment, we also tested their response to infection with an apoptotic HSV-1 strain. SIRC cells were mock infected or were infected with wt HSV-1(F), HSV-1(KOS), or the apoptotic mutant strain HSV-1(vBSΔ27). vBSΔ27 has been shown to trigger but not prevent apoptosis in human epithelial cells. ²⁸ At 17 hours post infection (hpi), the cells were stained with Hoechst DNA dye, visualized by light and fluorescence microscopy. At 18 hpi, whole-cell extracts were prepared and immunoblotted for viral ICP4 (infection marker) and cellular PARP and IκBα. The results (Fig. 3)were as follows. 

As expected, untreated mock-infected SIRC cells showed no membrane blebbing, cell shrinkage, nuclear fragmentation, or chromatin condensation (Fig. 3A) . Neither of the wt strains, HSV-1(F) or HSV-1(KOS), showed morphologic features of apoptosis. Surprisingly, the vBSΔ27-infected SIRC cells also did not exhibit apoptotic morphologies in these experiments. Several previous studies have shown that various cell cultures infected with the vBSΔ27 mutant virus undergo apoptosis in the absence of any additional stimulant because this virus is unable to block the cell death program. ^{20 21 28 29} Thus, this result was novel and particularly unexpected, since we know that SIRC cells die by apoptosis induced by TNFα and CHX. 

Figure 3Bshows the corresponding biochemical analysis of these infected SIRC cells. The presence of ICP4 in the F- and KOS-infected cells (Fig. 3 , lanes 2 and 3, respectively) confirmed productive infection of these cells. The ICP4 level was barely detectable with vBSΔ27 (Fig. 3 ; lane 4), as expected. ²⁸ Mock-infected cells had abundant levels of IκBα and unprocessed PARP (Fig. 3 ; lane 1). Detection of IκBα with vBSΔ27 (lane 4) is consistent with this virus’ being unable to activate NF-κB. ¹³ Complete loss of IκBα with F and KOS complements the findings in Figure 2and confirms the mechanism of NF-κB activation by wt HSV-1 in SIRC cells. 

No PARP processing was observed in cells infected with vBSΔ27 (Fig. 3 ; lane 4). This finding, along with the data in Figure 3Aobtained with this virus indicates that SIRC cells are resistant to HSV-1-dependent apoptosis. However, we observed that F and KOS infection led to levels of PARP cleavage that were above the mock background (Fig. 3B ; compare lanes 2 and 3 with lane 1). The levels of full-length (unprocessed) PARP with mock and vBSΔ27 were much greater than with F and KOS. This cleavage was consistently found with repeated experiments (data not shown). In general, the level of detected ICP4 correlated with that of processed PARP: more ICP4, more PARP cleavage. Taken together, these data suggest that while productive wt HSV-1 infection of SIRC cells does not induce apoptosis, it does stimulate the processing of at least one apoptotic mediator. 

Effect on Apoptosis of Inhibition of Protein Synthesis in WT-Infected SIRC Cells

We and others have reported that the addition of protein translation inhibitors during the course of wt HSV-1 infection leads to detectable apoptosis in the infected cells. ^{9 10} Moreover, infected cell apoptotic prevention factors produced between 3 and 6 hpi block the process from killing the cells. ⁹ We next set out to determine whether such treatments might enhance the amount of PARP cleavage that we observed (Fig. 3) . 

SIRC cells were HSV-1(F) infected, and CHX was added at either 3 or 6 hpi and maintained until the end of the experiment. In addition, CHX was added 1 hour before infection (−1 hpi) as an additional control. At 23 hpi, the cells were stained with Hoechst DNA dye and visualized by fluorescence and phase-contrast microscopy. At 24 hpi, whole-cell extracts were prepared and immunoblotted for ICP4 and PARP. Control experiments included untreated mock- and HSV-1(F)-infected cells and mock cells treated with either CHX or TNFα+CHX. In mock-infected cells, only those treated with TNFα+CHX showed apoptotic morphologic features (Fig. 4A) . None of the infected cells showed nuclear fragmentation and condensed chromatin, characteristics of apoptotic cells (Fig. 4B) . These results suggest that the addition of CHX does not facilitate HSV-1-dependent apoptosis in SIRC cells. 

Figure 4Cshows a biochemical analysis of the infected or treated cells. Untreated and CHX-treated mock-infected cells had abundant unprocessed PARP bands while the TNFα+CHX-treated cells did not (Fig. 4C ; compare lane 1, 2 with 3). Consistent with the results in Figure 3 , untreated HSV-1(F)-infected cells had a large amount of PARP processing (Fig. 4C ; lane 5). Pretreating with the protein synthesis inhibitor caused an expected decrease in virus replication, as evidenced by the absence of ICP4 production (Fig. 4C ; lane 4). However, this CHX treatment did not yield detectable PARP cleavage (Fig. 4C ; lane 4). No processed PARP was observed when CHX was added at either 3 or 6 hpi (Fig. 4C ; lanes 6, 7). These results suggest that infected cell protein synthesis is necessary for this processing of PARP to occur. 

Taken together, these data lead to the following conclusions: (1) Although wt HSV-1 infection does not ultimately lead to apoptosis, PARP processing occurs during productive viral replication in SIRC cells. (2) Infected cell protein synthesis is essential for this PARP processing, and these factors are produced after 6 hpi. 

Occurrence of NF-κB Translocation before PARP Processing in HSV-1-Infected SIRC Cells

The results presented suggest that the response of SIRC cells to HSV-1 differs in many ways from that of human HEp-2 cells, which are the prototypical cells for characterizing HSV-1-dependent apoptosis. In the final portion of this study, we directly compared the responses of SIRC and HEp-2 cells to HSV-1. HEp-2 and SIRC cells were simultaneously mock- and HSV-1(F)-infected, and CHX was added at either 3 or 6 hpi and maintained for 18 hours. At 17 hpi, the cells were stained with Hoechst DNA dye and visualized by fluorescence and phase-contrast microscopy. At 18 hpi, whole cell, nuclear, and cytoplasmic extracts were prepared and immunoblotted for PARP, NF-κB, caspase 3, and ICP4. Caspase 3 was included as an alternative marker for apoptosis in the HEp-2 cells but was omitted from the experiments with the SIRC cells, as the antibody does not react with rabbit cell protein. Controls included untreated mock- and HSV-1(F)-infected cells and mock cells treated with either CHX or TNFα+CHX. The results (Figs. 5 and 6)were as follows. 

As expected, both the untreated and CHX-treated HEp-2 (Fig. 5A)and SIRC (Fig. 5B)cells were flat and elongated, with large, round, intact nuclei. Treatment of both cell types with TNFα+CHX resulted in the cells’ displaying extensive apoptotic features: cell shrinkage, membrane blebbing, condensed chromatin, and nuclear fragmentation. Thus, both HEp-2 and SIRC cells undergo apoptosis in a similar manner when stimulated by exogenous environmental factors. Only infected HEp-2 cells treated with CHX at 3 hpi exhibited these characteristic morphologies. Infected HEp-2 cells treated with CHX at 6 hpi showed almost no apoptosis compared with either the 3 hpi-addition sample or the mock-infected HEp-2 cells treated with TNFα+CHX (Fig. 5A) . These results are consistent with our previous data, ⁹ confirming that delaying protein synthesis inhibition until 6 hpi reduces apoptosis due to production of antiapoptotic proteins during the prevention window. In contrast, none of the infected SIRC cells displayed apoptotic morphologies. Of note, the 3 hpi-addition sample had no evidence of chromatin condensation or nuclear fragmentation (Fig. 5B) . These findings are consistent with SIRC cells being unable to undergo HSV-1-dependent apoptosis. Thus, HEp-2 and SIRC cells differed in their response to infection in the presence of CHX. 

Figure 6Ashows the corresponding immunoblot analysis of the infected HEp-2 cells shown in Figure 5A . In mock-infected cells, NF-κB localized to the cytoplasmic fraction and neither PARP nor caspase 3 was processed, as expected (Fig. 6A , lanes 1, 2). CHX treatment of the mock-infected cells yielded similar effects (Fig. 6A ; lanes 3, 4). Control TNFα+CHX-treated cells had completely processed PARP and caspase 3 (Fig. 6A ; lanes 5, 6). Most of the NF-κB partitioned in the nuclear fraction (Fig. 6A ; lanes 5, 6). In this case, TNFα treatment stimulated NF-κB translocation to the nucleus but, due the presence of CHX, no new IκBα was synthesized to return NF-κB to the cytoplasm, and so it remained in the nuclear fraction of those cells. 

The HSV-1-infected HEp-2 cells without CHX did not display PARP cleavage or caspase activation (Fig. 6A ; lanes 7, 8). As expected, NF-κB predominated in the nuclear fraction of these cells due to its sustained stimulation during infection. ^{13 27} In the infected cells treated with CHX at 3 hpi, both PARP (complete loss of the full-length band) and caspase 3 were completely degraded (Fig. 6A ; lanes 9, 10). Combined with the related panels in Figure 5A , these results indicate that these cells were undergoing apoptosis. The bulk of NF-κB was cytoplasmic in these cells (lanes 9, 10). With the 6-hpi CHX addition, no PARP or caspase cleavage was observed, and the lanes resembled those with no CHX addition (Fig. 6A ; compare lanes 11, 12 with 7, 8). There was also more nuclear relative to cytoplasmic NF-κB at the 6-hpi versus the 3-hpi addition (Fig. 6A ; compare lanes 11, 12 with 9, 10). These findings are consistent with previously published experiments indicating that NF-κB translocation to the nuclei of HSV-1-infected HEp-2 cells correlates with apoptosis prevention. ^{13 30}  

Figure 6Bshows the biochemical analysis of the SIRC cells shown in Figure 5A . The goal was to determine whether the features described in the experiments with HEp-2 cells were replicated in SIRC cells. As expected, there was no NF-κB in the nuclear fraction and no PARP cleavage in the mock-infected, untreated SIRC cells (Fig. 6C , lanes 1, 2) or the control CHX-treated cells (Fig. 6C ; lanes 3, 4). In contrast to the results with HEp-2 cells (Fig. 6A) , there was almost complete disappearance of both PARP and NF-κB from the TNFα+CHX-treated SIRC cells (Fig. 6C ; lanes 5, 6). Any detected NF-κB was cytoplasmic (Fig. 6C ; lane 5). This observation seems to suggest that the SIRC cells are actually more sensitive to this extrinsic proapoptotic stimulation than are the HEp-2 cells. 

The untreated HSV-1-infected SIRC cells accumulated the same amount of viral ICP4 protein as the cells tested with CHX at 6 hpi (Fig. 6B ; compare lane 4 with 6). This indicates that the infectious programs were identical up to 6 hpi in both cases. As with infected HEp-2 cells, the nearly all of the NF-κB in the untreated SIRC cells was partitioned in the nuclear fraction (Fig. 6B , lanes 7, 8). NF-κB was cytoplasmic in the 3-hpi addition (Fig. 6C ; lanes 9, 10) and had more nuclear NF-κB in the 6-hpi addition (Fig. 6C ; lanes 11, 12). These results indicate that NF-κB activation requires de novo protein synthesis in HSV-1-infected SIRC cells and that its nuclear translocation initiates between 3 and 6 hpi. 

In contrast to the results with HEp-2 cells, the only detectable PARP cleavage occurred in the untreated, infected cells (Fig. 6C ; lane 8). The greatest PARP cleavage is found in wt-infected, untreated cells. The absence of detectable PARP cleavage in the 6-hpi-addition sample is particularly striking, as it implies that factors synthesized later in infection (>6 hpi) are necessary for this processing. Overall, these data indicate that NF-κB translocates to the nuclei of HSV-1-infected rabbit cornea cells as it does in human epithelial cells. Although NF-κB translocation in HEp-2 cells correlates with apoptosis prevention (inhibition of PARP processing), this event appears to correlate with promotion of PARP processing in SIRC cells, even though SIRC cells did not show any other apoptotic features. 

Discussion

Corneal infection by HSV is a complex process in which proinflammatory cytokine release ³¹ plays a significant role in disease progression (see review). ³² Initially, we endeavored to establish the baseline of apoptotic activity in the rabbit SIRC cell system in the presence of a known proapoptotic cytokine stimulant. However, very little is known about the molecular mechanisms through which corneal epithelial cells respond to HSV infection. Accordingly, we set out to determine whether HSV modulates apoptosis in prototype SIRC cells and if NF-κB participates in the process. Our significant findings may be summarized as follows. 

First, the prototypical corneal epithelial SIRC cells are sensitive to extrinsic proapoptotic stimuli (Table 1) . Thus, these cells are fully capable of dying by programmed cell death. However, in contrast to other standard cell systems, ^{13 16 23} SIRC cells do not appear capable of retaining nuclear NF-κB on TNFα+CHX treatment. Thus, these cells may possess a fundamental difference in their ability to regulate NF-κB. A consequence of such an effect would be enhanced apoptotic sensitivity to proinflammatory, extrinsic cytokines, such as TNFα and Fas. Additional studies are needed to determine whether this is a unique feature of SIRC cells or whether this represents a more fundamental aspect of the corneal epithelium. 

Second, SIRC cells are resistant to HSV-1-dependent apoptosis. As noted earlier, both the SIRC and HEp-2 cells exhibited apoptosis in response to environmental stimulants. Therefore, it was surprising that infection with the vBSΔ27 virus, a replication defective mutant known to stimulate apoptosis, or with the wt virus plus CHX did not yield any apoptotic features (Table 1) . This response resembles that of African green monkey kidney fibroblast Vero cells, which were initially described as being resistant to HSV-1-dependent apoptosis. ²⁸ However, a detailed analysis of these Vero cells revealed that they undergo apoptosis, albeit inefficiently. ²² The SIRC cells seem to respond to HSV-1 infection in a manner most similar to that previously described in primary human and mouse fibroblast cells. ²⁹ In this case, it was suggested that the p53 status of the cells may define their apoptotic response to infection and it is assumed that the SIRC p53 is functionally wt. 

Third, HSV-1 activates the nuclear translocation of NF-κB during infection in SIRC cells (Table 1) . The mechanism through which this occurs involves the loss of IκBα, a protein that serves to maintain NF-κB in the cell’s cytoplasm. When IκBα is degraded, NF-κB then translocates to the cell’s nucleus where it binds DNA and stimulates gene transcription, many of which are involved in antiapoptotic survival (see review). ^{14 26} The ability of SIRC cells to “survive” infection may be due, at least in part, to the activation of antiapoptotic factors such as NF-κB. 

Fourth, the apoptotic death factor PARP is cleaved during wt HSV-1 infection of SIRC cells. This processing occurs in the absence of any other characteristic features of apoptosis (Table 1) . Such degradation is considered one of the hallmarks of apoptosis in other standard cell types. ³³ A possible explanation for this is that productive viral infection of corneal epithelial cells may actually require a certain level of PARP processing to proceed successfully. Therefore, these findings represent the first description of such a phenomenon of which we are aware. Our current data imply that the viral infection must proceed to the later stages of infection for PARP processing to occur. Since NF-κB is activated by the virus at that time, it is conceivable that it may play a role blocking other obvious proapoptotic events. A potential conclusion that may be drawn from all of the results is that in SIRC cells, productive HSV-1 infection actually requires PARP processing for efficient replication. This processing may require the translocation/activation of NF-κB to occur in the infected cells. It is important to characterize additional cornea-derived epithelial cells to determine the prevalence of this effect. These future studies will help assess whether cells configured for tissue culture reflect those in their natural habitat. Targeted inhibition of PARP may represent a possible novel therapeutic target for the treatment of herpetic eye disease. 

The most intriguing remaining question is whether this has an impact on herpetic eye disease. In one animal study, Wilson et al. ⁵ found that infection of the rabbit corneal epithelium stimulates apoptosis of the underlying anterior stromal keratocytes. Apoptosis was observed using a TUNEL assay at 18 hpi, compared with 2 hours after a corneal scrape control. ⁵ They hypothesized that this apoptosis was induced by release of soluble mediators from infected epithelial cells as a means of creating a barrier to stop the spread of the viral infection. Questions raised by that study include whether the underlying cells were actually infected with HSV-1 and whether the virus’ ability to inhibit apoptosis was the cause of the delay from 2 hours after mechanical injury to 18 hours after viral infection. In other studies, apoptotic cells were detected in mouse eyes infected with HSV-1. ³ These results suggest that modulation of apoptosis by HSV is essential in the development of herpetic eye diseases. 

These SIRC cell-specific responses have not been described before. They suggest that HSV-1 infection of ocular structures may proceed via a different mechanism than that which facilitates infection of the skin and mucous membranes. Additional studies ongoing in our laboratory should elucidate whether early inhibition of NF-κB activation or PARP processing can block productive infection. Further investigations are needed to determine what impact this information may have on rational drug design and prevention of the severe ocular sequelae of herpetic disease. 

 Supported in part by National Institute of Allergy and Infectious Diseases AI38873 and AI48582 (JAB).

 Submitted for publication April 23, 2007; revised June 14, 2007; accepted August 21, 2007.

 Disclosure: M.L. Goodkin, None; S. Epstein, None; P.A. Asbell, None; J.A. Blaho, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Immune reactivities of treated rabbit SIRC cells. SIRC cells were untreated (−) or treated with CHX+TNFα and whole-cell extracts (50 and 80 μg) were separated in a denaturing gel, transferred to nitrocellulose, and probed for immune reactivity with anti-PARP antibody. Molecular weight markers are indicated in the margins; 116 and 85 refer to uncleaved and cleaved PARP, respectively.

Figure 2.

 

Live cell microscopy (A) and immune reactivities (B) of infected SIRC cells. SIRC cells were untreated (−), treated with CHX, TNFα, or TNFα+CHX, mock-infected or infected with HSV-1(F) (MOI = 10) and stained with Hoechst DNA dye before visualization by phase-contrast and fluorescence (Hoechst) microscopy at 18 hpi. Cytoplasmic (C) and nuclear (N) extracts were separated in a denaturing gel, transferred to nitrocellulose and probed for immune reactivity with anti-NF-κB antibody. Left: molecular weight marker.

Figure 3.

 

Live cell microscopy (A) and immune reactivities (B) of infected rabbit SIRC cells. SIRC cells were mock-, HSV-1(F)-, HSV-1(KOS)-, or HSV-1(vBSΔ27)-infected (MOI = 10), stained with Hoechst DNA dye before visualization by phase-contrast and fluorescence (Hoechst) microscopy at 18 hpi and whole-cell extracts were separated in a denaturing gel, transferred to nitrocellulose, and probed for immune reactivity with anti-ICP4, -PARP, and -IκBα antibodies. Molecular weight markers are indicated in the margins; 116 and 85 refer to uncleaved and cleaved PARP, respectively. At least three slight bands were observed in mock-infected cells that are most likely nonspecific degradation products.

Figure 4.

 

Live-cell microscopy (A) and immune reactivities (B) of infected rabbit SIRC cells. SIRC cells infected with HSV-1(F) (MOI = 10) were either untreated or CHX was added at 1 hour before (−1 hpi) or at either 3 or 6 hpi. Mock-infected cells were untreated or had CHX or CHX+TNFα added at 0 hpi. Cells stained with Hoechst DNA dye were visualized by phase-contrast and fluorescence (Hoechst) microscopy at 24 hpi, and whole-cell extracts were separated in a denaturing gel, transferred to nitrocellulose, and probed for immune reactivity with anti-ICP4, -PARP, and -IκBα antibodies. Molecular weight markers are indicated in the margins; 116 and 85 refer to uncleaved and cleaved PARP, respectively.

Figure 5.

 

Live-cell microscopy of infected human HEp-2 (A) and rabbit SIRC cells (B). HEp-2 and SIRC cells infected with HSV-1(F) (MOI = 10) were either untreated (−) or CHX was added at either 3 or 6 hpi. Mock-infected cells were untreated (−) or had CHX or CHX+TNFα added at 0 hpi. Cells were stained with Hoechst DNA dye before visualization by phase-contrast and fluorescence (Hoechst) microscopy at 18 hpi.

Figure 6.

 

Immune reactivities of fractionated HEp-2 (A), whole SIRC (B), and fractionated SIRC (C) infected cell extracts. HEp-2 and SIRC cells infected with HSV-1(F) (MOI = 10), and cells were untreated (−) or CHX was added at either 3 or 6 hpi. Mock-infected cells were untreated (−) or had CHX or CHX+TNFα added at 0 hpi. Whole-cell extracts (B) and nuclear (N) and cytoplasmic (C) fractions were prepared at 18 hpi and separated in a denaturing gel, transferred to nitrocellulose, and probed for immune reactivity with anti-PARP, -NF-κB, -caspase 3, and -ICP4 antibodies. Molecular weight markers are indicated in the margins; 116 and 85 refer to uncleaved and cleaved PARP, respectively.

Table 1.

 

Different Cell Susceptibilities to Environmental and HSV-Induced Apoptosis

Cell Name	Origin	Source	Apoptosis			ICP Dependence of Activation*
						Environmental, †	HSV Dependent, ‡	NF-κB, §	PARP Cleavage, ∥
HEp-2	Human	Skin	+	+	+	−
SIRC	Rabbit	Cornea	+	−	+	+

*  Dependence on infected cell protein (ICP) synthesis was determined using CHX.

†  Sensitivity to environmental stimulators of apoptosis was assayed as described in the text.

‡  Susceptibility to HSV-dependent apoptosis was determined after infection with either wild-type HSV + CHX or apoptotic vBSΔ27. Neither cell type displays apoptotic features following wild-type HSV infection alone.

§  NF-κB activation was determined by nuclear translocation.

∥  PARP cleavage was assessed by immunoblot analysis.

The authors thank Elise Morton for expert technical assistance.