Infection of murine keratinocytes with herpes simplex virus type 1 induces the expression of interleukin-10, but not interleukin-1α or tumour necrosis factor-α

Abstract

Herpes simplex virus (HSV) is known to possess several mechanisms whereby it can evade the normal host immune defences. In this study the expression of the immunosuppressive cytokine, interleukin (IL)-10, was monitored following infection of a murine keratinocyte cell line (PAM-212) and compared with the expression of two proinflammatory cytokines: IL-1α and tumour necrosis factor (TNF)-α. The PAM-212 cells were infected at a multiplicity of 0·5 with a clinical isolate of HSV type 1, and the mRNA of the three cytokines was assessed by semiquantitative reverse transcription–polymerase chain reaction (RT–PCR) over the following 24 hr. By 12 hr postinfection the amount of IL-10 mRNA had increased significantly to five-fold greater than that found in uninfected cells (P < 0·01), and this elevated level was maintained until at least 24 hr postinfection. In contrast, IL-1α and TNF-α mRNAs were not significantly up-regulated by the HSV infection. Immunostaining with an IL-10 monoclonal antibody (mAb) revealed that cytoplasmic IL-10 protein had increased by 6–12 hr postinfection. This quantity was further increased at 24 hr postinfection, when the viral cytopathic effect was apparent. Viral replication was necessary, but not sufficient on its own, for IL-10 induction. Experiments with HSV mutants lacking functional transactivating factors suggested that the viral transactivating proteins ICP-0 and VP-16 may be necessary for HSV-induced IL-10 expression. Thus, the up-regulation in the expression of IL-10 mRNA and protein induced by HSV early in the infection of keratinocytes represents a specific response and may be part of the viral strategy to avoid local immune defence mechanisms in the skin.

Introduction

In natural human infection with herpes simplex virus (HSV), the primary site of replication is at mucosal or epithelial surfaces. During the acute stage, the virus ascends the local sensory nerves and establishes latency in the neural bodies for an indefinite period of time. Subsequently it can reactivate and cause recrudescent lesions in the periphery, often in response to stress or immunomodulation. The immune response of the host is complex and depends on both innate mechanisms, especially macrophages and natural killer (NK) cells, and acquired mechanisms. However, the major role is thought to be played by T cells (reviewed in ref. 1).

In the skin, HSV is confined almost exclusively to the epidermis. There is focal necrosis of keratinocytes and intraepithelial vesicle formation. The inflammatory infiltrate is found mainly in the dermis and consists of macrophages and T cells with smaller numbers of NK cells and B cells.² CD4⁺ T cells predominate early, followed by an increase in the number of CD8⁺ T cells. Basal cells, keratinocytes and Langerhans' cells express major histocompatibility complex (MHC) class II antigens and adhesion molecules within 2 days of infection, as do endothelial cells and most infiltrating mononuclear cells. There is evidence from mouse models that the T helper 1 (Th1) cytokines, such as interferon (IFN)-γ and interleukin (IL)-2, provide the principal means of recovery. For example, treatment of the animals with antibodies to IFN-γ or IL-2 resulted in exacerbation of periocular skin lesions after corneal inoculation³ and in a diminished ability to clear the virus from the skin after footpad inoculation.⁴ Kanangat et al.⁵ have provided evidence that IL-12 is induced locally and in the draining lymph nodes in response to HSV infection of the eye which may promote, in turn, a Th1 type of cytokine response. Cytokines, including IL-10 and β (C–C) chemokines, produced in the vesicular fluid of human recurrent HSV lesions have been monitored, the source of which was presumed to be mainly from early infiltrating mononuclear cells.⁶

The induction of cytokines in keratinocytes infected with HSV has not been fully investigated, although this cell type is capable of producing a range of mediators in response to viral infection. One such study revealed the production of IL-1α, TNF-α and IL-6 in infected murine epidermal cells;⁷ another reported the up-regulation of IL-1α and IL-6, while IL-10 remained at a basal level.⁸ In a third study, Mikloska et al.⁶ examined the time course for the production of various cytokine and chemokine proteins in primary human keratinocyte cultures infected in vitro with HSV. The chemokines were stimulated first, followed by IL-12, IL-10, IL-1α, IL-1β and IL-6.

Many viruses, particularly the DNA-containing viruses, have evolved mechanisms to evade the host immune responses. Some of these mechanisms include the production of cytokines or cytokine receptors (reviewed in ref. 9). HSV is known to employ several strategies of immunomodulation, including down-regulation of the expression of MHC class I molecules, avoidance of complement activation, the induction of Fc receptors and a block in apoptosis. The following study aimed to investigate the effect of HSV infection of the murine keratinocyte cell line, PAM-212, on the expression of IL-10. PAM-212 cells have been reported previously to contain a baseline level of IL-10 mRNA.¹⁰ IL-10 was originally identified as a cytokine synthesis inhibitory factor because of its ability to inhibit the production of IFN-γ by Th1 cell clones.¹¹ It is now known to be an important suppressor of both T-lymphocyte and antigen-presenting cell effector function. It affects the ability of Langerhans' cells to present antigen, allowing them to stimulate T helper 2 (Th2) cells but inducing anergy in Th1 cells.¹² It also acts on macrophages to down-regulate MHC class II expression, resulting in the inhibition of cytokine synthesis by activated T cells and NK cells.

The expression of IL-10 in the PAM-212 cells was assessed at the level of mRNA and protein, and compared with two other cytokines, namely tumour necrosis factor-α (TNF-α) and IL-1α. Our preliminary results have been published regarding IL-10 mRNA in infected keratinocytes and in mice.¹³

As HSV is known to encode several transcriptional regulatory proteins (trans- activating factors; reviewed in ref. 14) which can, in some cases, bind directly to specific host cell DNA sequences, motifs and direct host gene expression, for example infected cell protein (ICP)-4 (reviewed in ref. 14). Alternatively, they can influence transcription by interacting with other proteins. One example of this latter category is VP16, a tegument protein which activates immediate-early gene transcription through interaction with the cellular proteins, Oct-1 and host cell factors. VP16 carries an acidic domain that binds to multiple components of the preinitiation complexes, thus stimulating transcription of not only viral, but also of particular cellular, genes (reviewed in ref. 14). We hypothesized that IL-10 induction in HSV-infected murine keratinocytes could be directed by viral transactivating proteins. Evidence to support that conclusion and the time course of IL-10 expression are presented in this report.

Materials and methods

Cell culture

PAM-212 cells comprise a line of spontaneously transformed murine keratinocytes¹⁵ and were cultured in Dulbecco's modified Eagle's minimal essential medium (DMEM) supplemented with 100 IU/ml penicillin, 200 µg/ml streptomycin and 5% heat-inactivated fetal calf serum (all Gibco Life Sciences Ltd, Paisley, UK). Cultures were maintained at 37° in a humidified atmosphere containing 5% CO₂. Cells were routinely cultured in 75 cm² flasks (Sarstedt, Leicester, UK) and were transferred to six-well dishes (Sarstedt) for all subsequent experiments.

In some assays, the PAM cells, grown on coverslips (see below), were incubated with agents (sodium selenite and sodium bisulphite menadione¹⁶; Sigma-Aldrich, Poole, UK) that generate oxidative stress. The menadione was added in complete DMEM at a final concentration of 50 µm and this was incubated with the cells at 37° for 30 min before the monolayers were washed three times and the original medium replaced. Sodium selenite was added to the cells in complete DMEM at 1 µm, and incubation was continued for a further 24 hr before immunostaining for IL-10 (see below).

Virus and cell infection

For the initial experiments, the virus was a plaque-purified isolate of HSV type 1 from a clinical case.¹⁷ It was passaged in Vero cells at a multiplicity of infection (MOI) of 0·2, on three occasions, from the original isolate. When the cytopathic effect was advanced, the cells were harvested, sonicated and stored at −70° in small aliquots. The number of plaque-forming units (PFU) per ml was calculated following a plaque assay in Vero cells and was 4·0 × 10⁷. PAM-212 cells were dispensed (at least in triplicate), at a cell density of 1 × 10⁵ cells per well, into six-well dishes and cultured until 70% confluent before infection. The virus, at an MOI of 0·5, was added to the cell monolayer in 0·5 ml phosphate-buffered saline (PBS), after removal of the DMEM. After 30 min at 37°, the original DMEM was added back and the cells cultured for a further 48 hr. Mock-infected cells were treated in the same manner, but with PBS only. For immunostaining, the PAM-212 cells were dispensed on to sterile glass coverslips placed in the six-well dishes and infected as above. When desired, the coverslips were removed, washed with PBS, and placed in freshly made 4% (wt/vol) paraformaldehyde in PBS for 15 min, followed by storage in absolute ethanol at 4° until required.

HSV mutants

The HSV mutants were a kind gift from Dr Chris Preston, Virology Institute, University of Glasgow, Glasgow, UK. They lack specific functional transactivating factors and were used as described for the clinical isolate above except that they were innoculated at an MOI of 5·0 to ensure a productive infection, which was verified by observing the cytopathic effect (CPE) at 48 hr postinfection. The viruses were as follows: 1814R: basically HSV-1 strain 17 wild-type virus,¹⁸ titre 3 × 10⁹ PFU/ml; in 1814: VP16 mutant,¹⁸ titre 2·5 × 10⁸ PFU/ml; tsK: ICP4 mutant,¹⁹ titre 1·5 × 10⁹ PFU/ml; in 1312: VP16 mutant, ICP0 mutant, ICP4 mutant,²⁰ titre 2·5 × 10⁹ PFU/ml; and dl 1403: ICP0 mutant,²¹ titre 1 × 10¹⁰ PFU/ml.

UVB radiation

In some experiments, virus suspensions were irradiated with UVB from a bank of TL20W/12 lamps (Philips, Croydon, UK) which were a kind gift from Dr Neil Gibbs (Photobiology Unit, University of Dundee, Dundee, UK). The TL20W/12 lamps have an emission spectrum of between 270 and 350 nm, with an E_max of 308 nm. The cumulative dose of UVB received by the cells was measured using an IL1400A radiometer (International Light Inc., Newburyport, MA). For some experiments, inactivated virus was prepared by UVB irradiation of the virus inoculum in PBS (UVB-HSV) with a dose of 2930 J/m². The virus was then used immediately at an erstwhile MOI of 5·0. Wild-type virus was also used at a MOI of 5·0 for this part of the study.

RNA isolation

Total cellular RNA was prepared for reverse transcription–polymerase chain reaction (RT–PCR) by the acidic phenol procedure of Chomczynski & Sacchi.²² When required, the medium was removed from the dishes and the monolayers washed twice with PBS to dryness. Then, 650 µl of solution D (4 m guanidinium isothiocyanate, 25 mm sodium citrate, pH 7·0, 10% wt/vol lauryl sarcosine, 8 mg/ml 2-mercaptoethanol) was added to each well.²² After 1 min, the cell lysates were collected and stored at −70°. The RNA was quantified by spectrophotometry at 260 nm.

PCR primers and RT–PCR amplification

PCR primer sequences for murine IL-1α, IL-10, TNF-α and β-actin were as described previously²³ and synthesized by Oswell Ltd (Edinburgh, UK). The RT–PCR conditions were a modification of those used previously.²⁴ Briefly, synthesis of cDNA employed 1 µg of each RNA sample, incubated at 37° for 1 hr with 300 pmol of random primers (Pharmacia LKB, London, UK) and 200 U of murine-Moloney leukaemia virus reverse transcriptase (Gibco Life Technologies). The reaction mixture (20 µl) also contained 50 mm Tris–HCl, pH 8·3, 70 mm KCl, 10 mm dithiothreitol, 3 mm MgCl₂ and 1·0 mm of each dNTP (Boehringer Mannheim, Lewes, UK).

Enzymatic amplification of the specific cDNA sequences was performed according to a modified procedure of Kondo et al.²⁴ Each condition was assayed in triplicate wells and each sample was amplified in duplicate. To 10-µl (final volume) of amplification reaction mix, made up in 1× Buffer IV (Advanced Biotechnologies, Epsom, UK), the following components were added: 0·2 mm of each dNTP (Boehringer Mannheim), 10⁻⁶ m tetramethylammonium chloride (Fisher Scientific, Toronto, Ont., Canada), 2·0 µl of cDNA and 0·25 U of ‘Red Hot’ Taq polymerase (Advanced Biotechnologies). The mixture was overlaid with 15 µl of paraffin oil (Sigma, Poole, UK) and heated to 85° in a thermal cycler (Techne, Cambridge, UK) before adding the primers. To increase the sensitivity of detection of amplified products, both primers were end-labelled with [γ-³²P]ATP, specific activity > 7000 Ci/mmol²⁴ (ICN Biomedicals, Basingstoke, UK), as follows: for each reaction, 0·5 µl of 20 µm 5′ primer was incubated with 0·6 U of T4 polynucleotide kinase (PNK) (New England Bio-Labs, Hitchin, UK) in 0·1 µl of 10×PNK buffer (700 mm Tris–HCl, pH 7·6, 100 mm MgCl₂, 50 mm dithiothreitol and 0·025 µl of [γ-³²P]dATP – specific activity 7000 Ci/mmol) for 1 hr at 37°. The mixture was then boiled for 10 min to inactivate the enzyme. Labelled primers from each set were mixed in equal volumes (6·7 µm each) and a 1·5-µl aliquot was carefully added to each PCR reaction mixture. The Mg²⁺ concentration for each of the primer pairs was optimized, and was as follows: IL-1α, 2·1 mm; IL-10, 1·1 mm; TNF-α, 1·1 mm; β-actin, 2·1 mm. The corresponding annealing temperatures used were: 60°, 65°, 65° and 60°, respectively. A ‘hot-start’ was used to improve the specificity of primer binding. The samples were then cycled at 95° for 1 min, followed by a 1-min cycle at the chosen annealing temperature. The optimum number of cycles was empirically derived for each primer set by plotting the band intensity (see below) obtained when samples consisting of a mixture of RT reactions from all the experimental treatments were amplified at increasing cycle number. When the band intensity was plotted against cycle number, a point was chosen within the upper portion of the linear region of the curve preceding the plateau seen at a higher cycle number. This point was extrapolated to the x-axis to give the optimum cycle number for future amplifications.

Controls

Positive controls for IL-10 were created by purifying RNA from the EL-4 mouse T-cell line that had been incubated with 50 ng/ml phorbol ester for 24 hr previously. The negative control was HSV-infected PAM-cell RNA, which was amplified without reverse transcription (‘+’ and ‘−’ in Fig. 1).

Figure 1.

Interleukin (IL)-10 and β-actin mRNA expression at 0–24 hr after infection of PAM-212 cells with herpes simplex virus (HSV) at a multiplicity of infection (MOI) of 0·5. Mock-infected cells were used as a control (MO24) for the 24-hr time-points. The RNA was extracted, reverse transcription–polymerase chain reaction (RT–PCR) amplification performed and the PCR products resolved by sodium dodecyl sulphate (SDS) gel electrophoresis, as described in the Materials and Methods. The sizes of the PCR products are given in base pairs (bp). Autoradiographs are shown of samples from one experiment of five. The positive (+) and (−) PCR controls are illustrated and are described in the Materials and Methods. L=molecular weight ladder. MO=mock-infected, 0 hr.

Electrophoresis and autoradiography

The PCR reaction was stopped by addition of 2 µl of 10× gel loading buffer III tracking dye.^25,26 Six microlitres of the total amplification mixture was size fractionated by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE), as described previously,²⁷ but with the following modifications: no stacking gel was used and samples were resolved on a 12% acrylamide gel run at 200 V for 1 hr. Product sizes were verified using radioactive 100-bp markers (Pharmacia), which had been labelled with ³²P by the exchange reaction of PNK using exchange reaction buffer (Gibco-BRL).²⁶ The gel was exposed to X-ray film (XAR-5, Kodak Corp.; Sigma Chemicals) for 10–30 min and the resulting images quantified by image analysis (SeeScan, Cambridge, UK). The gels for each individual primer set in a given experiment were exposed together for the same period of time. The expression was normalized for total RNA content by expressing the ratio of cytokine signal to that of the corresponding β-actin signal. To control for contaminating DNA, samples of cellular RNA not incubated with reverse transcriptase were also included: no product was detected.

Immunostaining for IL-10

The coverslips, prepared as described above, were incubated with 1% H₂O₂ in PBS to block endogenous peroxidase activity. This was followed by incubation in PBS (containing a 20% vol/vol dilution of rabbit serum) for 10 min at room temperature to block non-specific sites. A rat anti-mouse monoclonal antibody (mAb) (MAS629, clone 2AS, IgG1; Harlan Sera Labs, Loughborough, UK) was applied to the drained coverslips, at a dilution of 1:100 in PBS, for 18 hr at 4°. Biotinylated rabbit anti-rat IgG (Dako Corporation, Cambridge, UK) was used at a dilution of 1:400 for 30 min at 20°, followed by ABC horseradish peroxidase reagent (Dako), using diaminobenzidene (Dako) as substrate. The coverslips were mounted on glass slides in DPX mount (BDH Ltd, Poole, UK) and photographed at an original magnification of ×40 using PAN F50 film (Ilford Imaging UK Ltd, Mabberley, UK). Ten fields (× 40 magnification) in each of two or three slides were examined for each time-point. The staining intensity was assessed by a naı¨ve observer, who graded the overall staining intensity per field for each experiment. Staining was graded as 0 or 1–4, as described in the legend of Tables 1 and 2. Ten fields were assessed: the mean staining score per field was calculated for the pooled result from two or three experiments and expressed as the staining intensity±the standard error of the mean, which was then used for the statistical analysis.

Table 1.

Immunostaining for interleukin-10 (IL-10) protein in PAM-212 cells at various time-points after infection with herpes simplex virus (HSV), after UVB irradiation and after treatment with sodium selenite (Se) and sodium bisulphite menadione

Time after infection	IL-10 scoring
0 hr after HSV	2·08 ± 0·33
6 hr after HSV	3·19 ± 0·36*
12 hr after HSV	4·00 ± 0·44*
24 hr after HSV	4·00 ± 0·36*
24 hr mock-infected	1·22
24 hr after Se	2·41
24 hr after menadione	1·15
24 hr after UVB-inactivated HSV	2·22 ± 0·13

0, no staining; 1, faint staining (just visible); 2, readily visible staining; 3, stronger staining; 4, very strong dark-brown staining.Immunostaining of IL-10 protein expression in PAM murine keratinocytes infected with HSV at a multiplicity of infection (MOI) of 0·5, at 0–24 hr after infection. Some cultures were ‘infected’ with UVB-inactivated virus or treated with the oxidizing agents ‘Se’ (24 hr of treatment with 1 µm sodium selenite as described in the Materials and methods) or ‘menadione’ (24 hr after a 30-min exposure to 50 µm sodium bisulphite menadione). Scores represent the pooled means±SEM of three experiments. Where SEM values are not shown, only two experiments were carried out.

^*Significantly different from 0-hr HSV, P < 0·05.

Table 2.

Effects of herpes simplex virus (HSV) mutants on interleukin (IL)-10 protein induction in PAM cells

Mutation	Virus	Score at 48 hr	Score at 72 hr
None	Mock-infected	0·80 ± 0·2	1·50 ± 0·28
Wild type	1814R	3·40 ± 0·25*	3·75 ± 0·25*
VP16	in 1814	1·25 ± 0·25§	2·75 ± 0·25‡
ICP-4	tsK	4· 00± 0*¶	3·50 ± 0·29*¶
ICP-0	dl 1403	1·75 ± 0·25†	1·50 ± 0·3§
ICP-0,4 and VP16	in 1312	1·50 ± 0·29§	1·50 ± 0·3§

0, no staining; 1, faint staining (just visible); 2, readily visible staining; 3, stronger staining; 4, very strong dark-brown staining.Immunostaining of IL-10 protein expression in PAM murine keratinocytes infected with mutant HSV lacking the functional transactivating factors indicated in the left hand column. Cells were infected at an MOI of 0·5 (wild type-1814R) and an MOI of 5 for the mutants, to ensure a productive infection. Cells were harvested for immunostaining at 48 or 72 hr after infection. Scores were the pooled means±SEM of three experiments.

^*Significantly different from mock-infected (P < 0·001).

^†Significantly different from mock-infected (P < 0·01).

^‡Significantly different from mock-infected (P < 0·02).

^§Not significantly different from mock-infected.

^¶Not significantly different from the wild-type virus.

Statistics

The significance of differences in the cytokine: β-actin ratios and IL-10 immunostaining was assessed by the unpaired Student's t-test. Values of P < 0·05 were considered significant.

Results

Following infection of PAM-212 cells with the clinical isolate of HSV-1, the first signs of a CPE were visible after 12 hr and were clearly apparent after 24 hr, with many cells rounded and ≈50% detached from the plate. By 48 hr postinfection, almost all the cells had detached and this time-point was therefore not included in the mRNA studies of infections with the clinical isolate.

Signals representing amplified DNA were detected at 33 cycles for IL-10, at 30 cycles for β-actin and at 28 cycles for both IL-1α and TNF-α: the cycle numbers chosen from the linear region of the curve, when band intensity was plotted against cycle number, were 36, 32, 34 and 34 cycles, respectively. The effect of HSV infection of PAM-212 cells on the expression of IL-10 mRNA is demonstrated in Fig. 1 and Fig. 2: Fig. 1 shows the result of a single experiment, while Fig. 2 represents the mean result of five different experiments. After 6 hr, the level of IL-10 mRNA showed a twofold increase (P < 0·001), and a fivefold increase in mRNA compared with uninfected cells was found at 12 hr and 24 hr postinfection (P < 0·01). In contrast, the amounts of IL-1α and TNF-α mRNA had not changed significantly during the same period of infection (Figs 3a, b).

Figure 2.

Interleukin (IL)-10 mRNA expression at 0, 6, 12 and 24 hr after infection of PAM-212 cells with herpes simplex virus (HSV) at a multiplicity of infection (MOI) of 0·5. Mock-infected cells were used as a control (MO). The density of the bands for IL-10 was expressed as a ratio of the density of the bands for β-actin for the same sample and represents the mean value±standard error of the means of five different experiments. Difference from controls: *P < 0·01; **P < 0·001.

Figure 3.

Lack of effect of herpes simplex virus (HSV) infection on mRNA for (a) interleukin (IL)-1α and (b) tumour necrosis factor-α (TNF-α) cytokine mRNA expression at 6, 12 and 24 hr after infection of PAM-212 cells with HSV at a multiplicity of infection (MOI) of 0·5 (HSV series). Mock-infected cells were used as a control (MO) and samples were taken at the time-points indicated after infection. The density of the bands for cytokines was expressed as a ratio of the density of the bands for β-actin for the same sample and represents the mean value±standard error of the mean of three experiments. There were no significant differences at any time-point between the expression of the cytokine messages in the infected cells and the corresponding mock-infected cell cultures.

Immunostaining with the anti-IL-10 antibody revealed constitutive IL-10 in the cytoplasm of uninfected PAM-212 cells (Fig. 4a). An increase in the level of this protein was evident by 6 hr postinfection (Table 1) (P < 0·05). At 12 hr postinfection, the maximum number of cells was stained (Fig. 4b) and Table 1. The pattern of expression was striking; under higher power it was confirmed to be concentrated in perinuclear rings (as indicated by the arrows in Fig. 4b), but sparing the nucleus and the plasma membrane, possibly indicating synthesis of IL-10 in the Golgi complex. By 24 hr postinfection (Fig. 4c), some cells had detached but those remaining showed the greatest amount of staining per cell (Fig. 4c). At 48 hr, almost all the cells had detached owing to the CPE of the virus. The mock-infected cells at 24 hr are shown, at the same magnification, in Fig. 4(d). Table 1 presents the results of the immunostaining and demonstrates that there was a significant increase in the intensity of IL-10 protein immunostaining from 6 to 24 hr postinfection with HSV. This up-regulation was greater than that induced by UVB-inactivated HSV at 24 hr (Table 1). These results suggest that viral replication is necessary for IL-10 up-regulation.

Figure 4.

Immunostaining of interleukin-10 (IL-10) protein in PAM-212 cells at various time-points after infection with a clinical isolate of herpes simplex virus (HSV) at a multiplicity of infection (MOI) of 0·5. Original magnification: ×40. (a) Mock-infected, 0 hr; (b) 12 hr after HSV infection, perinuclear staining and sparing of the cell membrane is indicated by the arrows; (c) 24 hr after HSV infection; (d) mock-infected cells after 24 hr.

We attempted on several occasions to measure IL-10 in the culture supernatants from the infected cells using an ultrasensitive enzyme-linked immunosorbent assay (ELISA) kit (range 0·62–80 pg/ml; Biosource International, Camarillo, CA), but none was detected, although the control standards supplied with the kit gave the expected results. In addition, culture supernatants from PAM-212 cells collected 24 or 48 hr following UVB irradiation showed no detectable IL-10 protein. We attempted Western blotting in a effort to detect IL-10 protein, but the IL-10 antibody MAS629 did not work under these conditions.

We were concerned that the expression of IL-10 protein was simply a non-specific stress response elicited from dying cells. The role of IL-10 as a stress cytokine has been reviewed.²⁸ To rule out this possibility, cells were incubated with agents known to cause oxidative stress and death. A pulse treatment with 50 µm menadione for 30 min followed by replacement of the original medium for 24 hr had no effect on IL-10 protein expression (Table 1). Sodium selenite, which can act as a pro-oxidant at high concentrations, also had no significant effect (Table 1).

These immunostaining experiments were repeated with the mutant viruses (Table 2) using the wild-type 1814R as a positive control instead of the clinical isolate, because the mutant in 1814 was derived from this virus.¹⁹ The 1814R gave a lower IL-10 staining score than our clinical isolate used in Table 1. Differences in the staining intensity of mock-infected cells in these experiments (Tables 1 and 2) could be the result of culture conditions or because the mutant experiments were performed by different operators.

A significant increase in the IL-10 staining with 1814R was not seen until 48 hr postinfection. These infected cultures were scored at 48 hr and 72 hr after infection to ensure that there was productive infection and obvious CPE in the cultures infected with mutants (C. Preston, personal communication). At 48 hr after infection, the in 1814 (VP16) mutant was unable to produce a significantly greater IL-10 score than the mock-infected culture, although a significant difference from the mock-infected culture was seen at 72 hr after infection. However, the score was diminished compared to cultures infected with the wild-type virus. Staining in the cultures infected with the tsK (ICP-4) mutant was not significantly different from that seen in cultures infected with the wild-type control at either 48 hr or 72 hr. However, the dl 1403 (ICP-0) mutant had the greatest effect of any single mutation in diminishing IL-10 protein expression, compared to wild-type-infected cultures, at 72 hr. The cultures infected with the in 1312 mutant, bearing mutations in all three transactivating factors, did not show significantly greater IL-10 staining than the mock-infected cultures at either the 24 or 72-hr time-points. This was despite the presence of similar degrees of CPE achieved in all virus-infected cultures.

Discussion

In this study we have shown that IL-10 mRNA and protein are induced as a result of HSV infection of murine keratinocytes. The increase in IL-10 expression may be specific, as the levels of the two other cytokines (IL-1α and TNF-α) also monitored were not changed by the infection. Viral replication also seems to be necessary, as UV-inactivated virus was incapable of inducing IL-10 (Table 1). However, replication alone may not be sufficient, as indicated by the results obtained with the dl 1403 and in 1312 mutants shown in Table 2. The IL-10 response is not simply a consequence of stress, for several reasons. First, under the same conditions in which IL-10 was induced, TNF-α mRNA and the message for the prototypic stress cytokine IL-1α were not induced. Second, treating the cells with 1 µm sodium selenite or 50 µm menadione, which can act as pro-oxidants at these concentrations,¹⁶ had no effect on the expression of IL-10 protein (Table 1). Finally, there was no induction of IL-10 protein in cultures infected with the in 1312 mutant, despite a visible CPE (Table 2).

HSV inhibits host cell mRNA and protein synthesis at an early stage of the viral replication cycle in permissive cells. However, this effect is not absolute, as the expression of β-actin mRNA was not affected. The sparing and augmentation of IL-10 expression may therefore be useful in the viral life cycle. The lack of effect on the proinflammatory cytokines IL-1α and TNF-α is also interesting, as stress and infection normally augment the expression of these cytokines. The poxviruses have evolved various strategies to minimize the biological action of IL-1 during infection.⁹ Perhaps a similar strategy is also used by HSV; this aspect warrants further inspection.

Three viruses, two of them herpesviruses, are known to encode IL-10. The first viral IL-10 to be described is produced by Epstein–Barr virus which shares some, but not all, of the properties of cellular IL-10.²⁹ The second is coded by orf virus belonging to the parapox family.³⁰ Very recently a third viral version of IL-10 was discovered, coded by cytomegalovirus, which has minimal homology with the cellular IL-10.³¹ No gene coding for IL-10 has been described in the HSV genome thus far, and it seems probable that the cellular IL-10 is up-regulated, rather than a viral gene. The human group of rhinoviruses have recently been shown to induce IL-10 protein release from human monocytes during infection.³²

As HSV has been sequenced and does not contain an open reading frame coding for IL-10, we speculated that IL-10 induction could occur through the action of a viral regulatory protein or transactivating factor on host cytokine genes. Our experiments with mutant viruses containing certain non-functional transactivating factors suggest that ICP-0 and VP-16, but not ICP-4, may be necessary for IL-10 protein induction. It is also possible that the effect of HSV may be indirect, i.e. a factor could be released from infected cells which has the ability to induce IL-10 expression in adjacent uninfected cells. This seems unlikely, given the rapid increase in IL-10 mRNA and protein which follow HSV infection.

Although the level of IL-10 mRNA was increased five-fold in HSV-infected keratinocytes compared with that found in uninfected cells, and IL-10 protein also increased intracellularly, no IL-10 protein was detected extracellularly. The pattern of immunostaining suggests that the cytokine may be expressed preferentially in the Golgi apparatus and accumulates in that site. As a similar result was obtained following UVB irradiation of PAM-212 cultures (data not shown), the transport of IL-10 may be defective in this cell line. In support of this hypothesis, it is known that TNF-α protein also fails to be transported extracellularly from PAM-212 cells.^16,33 Furthermore, another study was also unable to detect extracellular IL-10 from irradiated PAM-212 cells that were obtained from two different sources (P. McLoone, University of Dundee, personal communication). However, it is possible that IL-10 may indeed be secreted from infected keratinocytes in vivo during a natural HSV infection, or released by cell lysis as a result of the virus or by CD8⁺ cytotoxic cells. Interleukin-10 was detected by ELISA in HSV-infected primary human keratinocytes from three different donors.⁶ As yet, studies of IL-10 protein expression have not been carried out after epidermal inoculation of mice with HSV. However our preliminary results indicate an up-regulation in IL-10 mRNA from ≈3 days postinfection when the vesicular HSV lesions first become apparent.¹³ The production of IL-10 early in the replication cycle of HSV in the epidermis may be a significant immune evasion mechanism which operates locally in the skin. For example, Beissert et al.³⁴ showed that UVB-induced IL-10 from PAM 212 cells was able to inhibit the presentation of tumour-associated antigens by Langerhans' cells. Furthermore, Manickan et al. have shown that intramuscular injection of a cDNA directing expression of IL-10 impairs cutaneous delayed-type hypersensitivity reactions to HSV challenge in mice.³⁵

Posavad et al.³⁶ suggested recently that, owing to the down-regulation of MHC class I antigens on infected keratinocytes by the action of the product of the immediate-early gene ICP47, CD4⁺ T cells and NK cells are the first immune-effector cells to enter the site of infection. The presentation of HSV antigens by Langerhans' cells and monocytes, which are more resistant to the viral inhibition of MHC expression³⁷ is likely to lead to the stimulation of the CD4⁺ T cells. IFN-γ produced by them and by activated NK cells could then up-regulate MHC class I peptide expression on the surface of infected keratinocytes. This would enable the CD8⁺ T cells, which infiltrate into the lesion a day or two after the CD4⁺ T cells and NK cells, to be activated and lyse the infected cells, thereby curtailing the spread of the virus. Indeed, there is evidence to indicate that local cytotoxic T-cell activity correlates with viral clearance from cutaneous HSV lesions.³⁸ Thus, the presence of IL-10 may interfere with the expansion of Th1 cells and their production of IFN-γ in the lesion and inhibit the antiviral cytokine synthesis by NK cells.

Abbreviations

CPE

cytopathic effect

DMEM

Dulbecco's modified Eagle's minimal essential medium

HSV

herpes simplex virus

ICP

infected cell protein

IFN

interferon

IL

interleukin

MOI

multiplicity of infection

NK

natural killer

PBS

phosphate-buffered saline

PFU

plaque-forming units

PNK

polynucleotide kinase

RT–PCR

reverse transcription–polymerase chain reaction

TNF-α

tumour necrosis factor-α

UVB

ultraviolet emission at 290–315 nm