PrP^c Expression Influences the Establishment of Herpes Simplex Virus Type 1 Latency

Abstract

PrP^c is a glycophosphatidylinositol-linked cell-surface protein expressed principally by neural tissue. The normal function of this protein is unestablished, although a role in either transmembrane signaling, cell-cell adhesion, or copper metabolism has been proposed. In this study we have investigated the effect of the neurotropic virus herpes simplex virus type 1 (HSV-1) in strains of mice which express different levels of PrP^c. Viral gene expression under the control of the HSV-1 early promoter IE110, detected either by in situ hybridization for RNA transcripts or by β-galactosidase (β-Gal) activity from an inserted lacZ gene, showed that the magnitude of HSV replication was retarded in PrP^−/− mice. This was reflected in the lower level of acute viral titers in tissues from these virus-inoculated mice. However, HSV-inoculated PrP^−/− mice contained higher levels of latent virus in both peripheral and central nervous tissue than those seen in mice which express PrP^c. Our observations show that lack of PrP^c expression favors the establishment of HSV latency whereas HSV replication proceeds more efficiently in neuronal tissue that expresses this protein. The data further suggest that PrP^c may be involved in a metabolic pathway that culminates in apoptosis of neurons that have been infected by neurotropic viruses.

PrP^c is a glycophosphatidylinositol-linked cell surface protein expressed by several cell types including those of the nervous system, both peripheral (PNS) and central (CNS) (5, 14). Despite a clear conservation of DNA and protein sequence homology among a wide variety of mammalian species, and presumably therefore function, the normal role of this protein remains unknown. It is established that PrP^c is involved in prion disease, as PrP^−/− mice are resistant to infection with these agents (8). Several studies have addressed the normal function of PrP^c through the use of PrP^−/− mice. NPU-1 (34) and Zürich I (9) strains of PrP^−/− mice have been reported to show altered neuronal electrophysiological properties. These include late after-hyperpolarization current (I_AHP) (12), changes in long-term potentiation and changes in the kinetics of inhibitory postsynaptic currents of hippocampal neurons (13, 27). PrP^−/− mice also show alterations in circadian rhythm (56, 57) and increased sensitivity to seizures (59).

The causes for the nonpathogenic changes in neuronal function of PrP^−/− mice are not yet known. They may arise through a decreased copper concentration in synaptic membranes of PrP^−/− mice (22). In vitro PrP^c has been shown to bind this metal ion at the N-terminal octapeptide repeat regions (6) and within the C-terminal portion of the molecule (21). Alternatively, the absence of PrP^c may perturb the regulation of neuronal oxidative stress in which PrP has been implicated through a proposed superoxide dismutase activity (7). Recently, a role for PrP^c in signal transduction was proposed through its linkage to the tyrosine kinase Fyn (37). An association of PrP^c with a metabolic pathway, in particular one involved with neuronal degeneration, is supported by the recent discovery of the PrP-related protein Doppel. Overexpression of Doppel in Ngsk (48), RCM0 (36), and Zürich II (47) PrP^−/− mice leads to Purkinje cell loss with the subsequent development of ataxia. Mice may be rescued from this phenotype by the reintroduction of a PrP transgene (36, 42). These observations support the hypothesis that PrP^c may interact with a cellular receptor involved with the regulation of neuronal survival (52).

As neurons express high levels of PrP^c the function of this molecule is presumed to play an important role in these cells under normal conditions. We have reasoned that changes in the physiology of neurons, which arise through changes in PrP^c expression, may modulate the fate of neurotropic viral pathogens in these cells. Accordingly, we have investigated the pathogenicity of herpes simplex virus type 1 (HSV-1) in mice that express different levels of PrP^c. This neurotropic virus initially invades epidermal or epithelial cells of mucosal surfaces. During this process HSV-1 enters the termini of local sensory neurons of the PNS. Retrograde axonal transport carries the virus to the neuronal cell bodies, where it establishes productive or latent infection. Intermittent reactivation may result in the production of infectious HSV from latently infected neurons (19, 46). Our initial studies with HSV-1 (SC16) wild-type virus showed that increased susceptibility to infection with this virus correlated with increased PrP^c expression. Here, we have extended these observations and have used the recombinant virus HSV-1 SC16 110lacZ to correlate virus pathogenicity, distribution, and latency in mice that express different amounts of PrP^c protein. Our observations show that lack of PrP^c expression favors the establishment of HSV latency whereas HSV replication proceeds more efficiently in neuronal tissue that expresses this protein.

MATERIALS AND METHODS

Mice.

PrP^−/− (Zürich I) (9) and tga20 (18) mice obtained from Charles Weissmann were bred in isolators and transferred to conventional animal facilities for inoculation with HSV-1. Wild-type mice were 129Sv × C57BL/6 bred from parental strains purchased from Harlan, Loughborough, United Kingdom. All regulated procedures involving experimental animals were carried out under Project and Personal licence authority issued in accordance with The Animals (Scientific Procedures) Act 1986.

Virus strain.

The virus used was either HSV-1 SC16 (wild type) (23) or SC16 110lacZ, a recombinant strain containing a 968-bp promoter fragment extending from position −818 to position +150 with respect to the immediate-early 110 (IE110)promoter transcription start site (43), linked to lacZ and inserted into the nonessential US5 locus of HSV-1 strain SC16 (2). Virus working stocks were prepared at a low multiplicity of infection in BHK-21 cells and stored at −70°C until needed. The recombinant virus was a gift from S. Efstathiou, University of Cambridge, Cambridge, United Kingdom.

Virus inoculation of mice.

Female mice were inoculated with virus by the intradermal route at 4.5 weeks of age. Virus was inoculated into the skin of the left ear pinna (41) with 10 μl of virus suspension containing 10⁶ PFU of wild-type SC16 or recombinant SC16 110lacZ in a single, discrete inoculation site. Forty-two mice were inoculated with virus and six were mock infected per strain. Mice were observed twice daily for clinical signs including increase in ear pinna thickness, weight gain or loss, and mortality from day 0 to 20 postinoculation (p.i.) as previously described (41).

Isolation of tissue samples.

Tissue samples (left ear pinna, left and right trigeminal ganglia [TG], left and right cervical dorsal root ganglia CII/CIII/CIV, brain stem, brain, and cervical and axillary lymph nodes) were obtained on days 4, 6, and 8 p.i. and were used for a variety of tests as described below.

Detection of ICP0-sense RNA by in situ hybridization.

In situ hybridization was performed with various tissues from three mice per group and two mock-infected controls to detect ICP0-sense RNA during the acute infection. Mice were sacrificed on days 4, 6, and 8 p.i., and left and right TG, left and right CIII, brain, and brain stem were fixed in periodate-lysine-paraformaldehyde fixative. Plasmid pSLAT 5 was used to generate the probe using a digoxigenin (DIG) detection system as described in detail by Arthur et al. (1). S. Efstathiou, University of Cambridge, Cambridge, United Kingdom, kindly provided the plasmid. After transcription, the reaction mixtures were ethanol precipitated and the product was resuspended in 100 μl of 10 mM Tris (pH 8)-1 mM dithiothreitol with RNase inhibitor. Tissues were fixed in periodate-lysine-paraformaldehyde at 4°C for 16 h, transferred to 50% ethanol, and then paraffin embedded. Sections (5 μm thick) were collected onto glutaraldehyde-activated, 3-aminopropyl-triethoxysilane-coated slides and were dewaxed in xylene before use. Sections were digested with 100 μg of proteinase K per ml at 37°C for 8 min. Overnight hybridization was carried out at 71°C. One to 3 μg of DIG-labeled probe was used in each 100 μl of hybridization solution. One stringent wash in 0.1× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate)-30% formamide-10 mM Tris-HCl (pH 7.5) was carried out at 70°C for 30 min. Bound probe was detected with alkaline phosphatase-conjugated anti-DIG Fab fragments according to the manufacturer's instructions (Roche). RNase- and DNase-treated tissue sections were included in each hybridization test as internal controls. Alternate sections from all tissues were counted and used to calculate the mean number of positive neurons per section per group. Cells containing a level of black or brown staining clearly above the background level were scored as positive.

Detection of β-Gal-positive neurons.

Three mice per group and one mock-infected control were sacrificed on days 4, 6, and 8 p.i., and β-galactosidase (β-Gal)-positive neurons were detected by histochemical staining of whole ganglia and sliced brain and brain stem with 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal). Tissues were reacted with X-Gal for 6 h prior to clarification in glycerol and enumeration of positive neurons using previously described methods (30). Following a preliminary enumeration of blue cells, the ganglia and brain tissue were wax embedded and sectioned (5 or 10 μm). The positive neurons were then counted microscopically in each section, and the results were recorded.

Detection of β-Gal-positive cells in draining lymph nodes.

Three mice per group and one mock-infected control were sacrificed on days 4, 6, and 8 p.i., and β-Gal-positive cells were detected by histochemical staining of cervical and axillary lymph nodes. Squashed, intact whole mounts were examined microscopically, and representative samples are shown in Fig. 2b.

FIG. 2.

Detection of acute viral RNA and protein. Various mice were inoculated with 10⁶ PFU of SC16 110lacZ by intradermal inoculation of the left ear pinna. (a) In situ hybridization for ICP0. Representative brain stem sections are shown for day 6 p.i. The staining pattern was predominantly nuclear. Magnification, ×200. (b) Detection of β-Gal-positive cells in draining lymph nodes. Representative cervical lymph nodes are shown as whole mounts for the different mouse strains sampled on day 6 p.i. Magnification, ×40.

Acute virus titers in various tissues following intradermal inoculation with HSV.

Three mice per time point per group plus one mock-infected control were sampled and placed in 1 ml of Eagle's minimal essential medium. Each tissue sample was minced with scissors, homogenized, and sonicated on ice for 2 min (16). The samples were centrifuged at 1,500 × g to remove debris. Either 200 μl of undiluted, 10-fold diluted, or 100-fold diluted tissue sample supernatant was inoculated onto triplicate BHK-21 cell monolayers in 24-well plates and allowed to adsorb for 1 h. Each well was subsequently overlaid with 1 ml of Eagle's minimal essential medium containing 2% heat-inactivated fetal calf serum plus 2% carboxymethylcellulose. Plates were stained and fixed with crystal violet following incubation at 37°C, 5% CO₂ for 48 h, and viral plaques were counted. All samples were coded prior to the start of the experiment, and this code was not revealed until all the results were recorded as mean virus titers (log₁₀ PFU) per tissue per group ± standard deviations.

TUNEL assay to detect apoptotic neurons.

Tissue samples taken for acute in situ hybridization were also used for the terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling (TUNEL) assay in order to detect apoptotic cells. The in situ cell death detection kit POD (Roche) was used according to the manufacturer's instructions. Briefly, tissue sections (8 μm) were dewaxed in xylene and rehydrated through ethanols. Endogenous peroxidase was blocked with hydrogen peroxide, and the tissues were permeabilized on ice for 2 min. The tissue sections were incubated with TUNEL reaction mixture for 60 min at 37°C in a humidified chamber; this was followed by incubation with an antifluorescein antibody, Fab fragment from sheep, conjugated with horseradish peroxidase (POD). Apoptotic cells were visible by light microscopy after treatment with 3,3"-diaminobenzidine (DAB; Vector Labs) and counterstaining with hematoxylin. Positive TUNEL staining was indicated by a brown precipitate. Positive and negative control slides were included in every TUNEL assay. Positive controls were pretreated with DNase.

Detection of latently infected neurons by means of in situ hybridization.

In situ hybridization was performed with tissues from five mice per group and two mock-infected controls to detect latency-associated transcripts (LATs) >30 days postinfection (54). The probes used to detect LATs were generated by T7 polymerase transcription of HindIII-linearized pSLAT 2 (kindly provided by S. Efstathiou) with a DIG detection system as previously described for ICP0 (1). Overnight hybridization was carried out at 72°C. One to 3 μg of DIG-labeled probe was used in each 100 μl of hybridization solution. One stringent wash in 0.1× SSC-30% formamide-10 mM Tris-HCl (pH 7.5) was carried out at 75°C for 30 min. Bound probe was detected with alkaline phosphatase-conjugated anti-DIG Fab fragments according to the manufacturer's instructions (Roche). RNase- and DNase-treated tissue sections were included in each hybridization test as internal controls. Alternate (5-μm) sections from all tissues were counted and used to calculate the mean number of positive neurons per section. Cells whose nuclei contained a level of black or brown staining clearly above the background level were scored as positive.

Histology (H&E).

Alternate brain stem sections were taken from latent HSV-1-infected PrP^−/−, wild-type, and tga20 mice and stained with hematoxylin and eosin (H&E). Representative sections are shown in Fig. 5c.

FIG. 5.

Detection of latent HSV by in situ hybridization. Various mice were inoculated with 10⁶ PFU of SC16 110lacZ by intradermal inoculation of the left ear pinna. In situ hybridization for major LAT was performed on tissues from five mice per group and two mock-infected controls >30 days p.i. Sections were analyzed by means of a DIG-labeled major LAT riboprobe as described in Materials and Methods. Representative sections are shown for each HSV-inoculated group of mice. The punctate brown-black signal is characteristic of this probe and is confined to the nucleus. Alternate 5-μm sections were used for each tissue. Similar sections were used to enumerate the LAT-positive cells for the data shown in Table 5. (a) Whole brain (magnification, ×200). (b) Whole brain (magnification, ×400). (c) Alternate brain stem sections were taken from the HSV-1-infected mice and stained with H&E. Representative sections are shown. Magnification, ×100.

Statistical analysis.

Statistical analysis of the data was performed by one-way analysis of variance together with the Duncan multiple-range test (P < 0.01) and the Student-Newman-Keuls test (P < 0.05) for post hoc analysis.

RESULTS

Decreased susceptibility to HSV-1 infection of PrP^−/− mice.

PrP^−/−, wild-type, and tga20 mice express zero, normal, and 10 times normal levels of PrP^c protein, respectively. These three strains of mice were inoculated with HSV-1 to determine the effect of PrP^c expression on viral neuropathogenesis. An initial experiment with wild-type HSV-1 (SC16) showed that while PrP^−/− mice did not succumb to virus-induced mortality, wild-type and tga20 mice showed 30 and 90% mortality, respectively (Fig. 1a). We have carried out the infection of these three strains of mice with wild-type HSV-1 three times, and on each occasion the same trends were seen with respect to mortality for wild-type mice and tga20 mice, i.e., wild-type mice showed 20 to 30% mortality and tga20 mice showed 70 to 90% mortality. To investigate further whether PrP^c expression affected viral pathogenicity, we used the less pathogenic recombinant strain of HSV SC16 110lacZ, which contains the lacZ reporter gene inserted under the control of the IE110 promoter (31).

FIG. 1.

Clinical signs following intradermal inoculation with HSV. Various mice (n = 10) were inoculated with 10⁶ PFU of SC16 wild-type HSV-1 by intradermal inoculation of the left ear pinna. The percent survival is shown in the upper panels (a). Other mice (n = 10) were inoculated with 10⁶ PFU of SC16 110lacZ by the same route (closed symbols). Clinical signs were recorded daily for days 0 to 20 p.i. Mock-infected mice were inoculated with culture medium (open symbols). (b) Weight gain or loss. (c) Percent survival. Error bars are shown. The data presented are from one of two experiments, which displayed the same trends.

Administration of the virus by the ear pinna route induced classical early clinical signs of virus infection in the form of acute inflammation and redness at the site of inoculation followed by ear paralysis. Initial clinical signs were first seen in tga20 mice usually on day 1 or 2 p.i., subsequently in wild-type mice, and lastly in PrP^−/− mice, typically several days later. All tga20 mice had macroscopic ear lesions by day 3 p.i., while PrP^−/− and wild-type mice showed ear lesions by day 5 p.i. The three strains of mice displayed significant differences with regards to weight gain or loss and mortality as a consequence of viral infection (Fig. 1). HSV-infected PrP^−/− and wild-type mice showed weight gain that was similar to, albeit slightly less than, that observed with the mock-infected control animals (Fig. 1b). In contrast, HSV-1-infected tga20 rapidly lost weight during the experiment and never managed to regain their start weights. The weights of tga20 mice were significantly less than those of mock-infected mice on days 4 and 5 p.i. (P < 0.05). While HSV SC16 110lacZ is regarded as less pathogenic than wild-type HSV-1, this recombinant form of virus had a significant effect on the tga20 mice as there was 90% mortality in this group of infected animals. All of the tga20 mice that succumbed to terminal disease died on day 6, 7, or 8 p.i. (Fig. 1c). A significant feature of the mortality of the tga20 mice was their rapid decline in health in response to the virus. Mice progressed from a grossly normal appearance to terminal disease within a matter of hours. This is in comparison to other strains of mice that are susceptible to HSV-1 infection, such as BALB/c, which take 2 to 3 days to progress through a well-defined series of clinical signs leading to terminal disease (17). In comparison, all of the PrP^−/− and wild-type mice survived the viral challenge, although they all displayed mild clinical signs associated with HSV infection. PrP^−/− and wild-type mice showed no signs of HSV-induced hind limb paralysis during the experiment. In contrast, the tga20 mice displayed complete loss of use of their hind limbs just hours before reaching terminal disease. It appears that increased expression of PrP^c correlates with increased severity of clinical signs following HSV inoculation.

Slower accumulation of HSV gene products in PrP^−/− mice.

Acute viral gene expression was investigated on days 4, 6, and 8 p.i. by detection of ICP0 transcripts using in situ hybridization on various tissues of the PNS and CNS as shown by representative sections in Fig. 2a and enumerated in Table 1. PrP^−/− and wild-type mice showed a gradual increase in the number of ICP0-positive neurons with time, whereas tga20 mice showed a gradual decrease. In tissues from PrP^−/− mice, the number of ICP0 in situ-positive neurons was minimal on day 4 p.i. and subsequently increased on days 6 and 8 p.i. A similar trend was seen in the majority of tissues from wild-type HSV-inoculated mice but with a greater number of positive cells than those observed in PrP^−/− samples. In contrast, in tga20 HSV-infected mice, the number of ICP0-positive neurons was maximal on day 4, when the left and right TG and the LCIII dorsal root ganglia, brain stem, and brain had positive neurons, and was subsequently decreased on days 6 and 8 p.i. The highest number of positive cells was seen in the brains of tga20 mice, which represented the tissue with the largest surface area that was examined. As shown by the data in Table 1, many of the values for tga20 ICP0 samples were significantly greater (at either P < 0.01 or P < 0.05) than the values for equivalent wild-type or PrP^−/− samples. In turn, many wild-type samples showed significantly greater numbers (at P values of either <0.01 or <0.05) than those of equivalent PrP^−/− samples. It was noted that there was an absence of any detectable ICP0 in the contralateral ganglia of the PrP^−/− and wild-type mice although wild-type mice did have detectable levels of acute virus. This may reflect differences in the sensitivities of the tests employed or the fact that the tissues were tested at only three defined time points as opposed to every day, in which case more positive results might have been seen. These data show that HSV infection clearly had less of an effect on PrP^−/− mice than on those that expressed PrP^c.

TABLE 1.

Assessment of acute viral RNA and protein

Substance detected	Tissuea	Mean no. ± SD of neurons testing positive per section per group at indicated day p.i.
		PrP−/−			Wild type			tga20
		4	6	8	4	6	8	4	6	8
ICP0 in situ	Left TG	4 ± 2	22 ± 15	41 ± 6	30 ± 5b	81 ± 15b	96 ± 11b	90 ± 9b,d	22 ± 8d	12 ± 2b,d
	Right TG	0	0	0	0	0	0	28 ± 4b,d	0	0
	LCIII	0	4 ± 2	15 ± 5	2 ± 1	6 ± 2	26 ± 4c	52 ± 6b,d	45 ± 7b,d	20 ± 3
	RCIII	0	0	0	0	0	0	0	0	0
	Brain stem	42 ± 11	77 ± 11	90 ± 14	24 ± 6	53 ± 20	68 ± 7	87 ± 12b,d	59 ± 13	34 ± 13b,e
	Brain	20 ± 7	51 ± 10	200 ± 40	50 ± 16	200 ± 48b	80 ± 16b	250 ± 36b,d	150 ± 30c	44 ± 18c
β-Gal	Left TG	0	0	22 ± 6	0	6 ± 2	24 ± 6	37 ± 10b,d	8 ± 2b,d	2 ± 2b,d
	Right TG	0	0	0	0	0	0	0	0	0
	LCIII	0	0	20 ± 12	0	14 ± 4c	34 ± 11	36 ± 12b,d	30 ± 9b	12 ± 3
	RCIII	0	0	0	0	0	0	0	0	0
	Brain stem	0	0	15 ± 5	0	25 ± 16c	47 ± 15c	50 ± 13b,d	35 ± 14c	28 ± 10
	Brain	0	0	45 ± 7	0	0	63 ± 21	68 ± 12b,d	41 ± 18b,d	21 ± 11e

^aLCIII, left third cervical dorsal root ganglia; RCIII, right third cervical dorsal root ganglia.

^bP < 0.01 (comparison with PrP^−/−).

^cP < 0.05 (comparison with PrP^−/−).

^dP < 0.01 (comparison with wild type).

^eP < 0.05 (comparison with wild type).

A similar trend was seen when β-Gal staining on acute tissues from the different HSV-inoculated mice was carried out to detect IE110 expression. Tissues were stained and initially recorded as whole mounts. All tissues were then sectioned and recounted microscopically to give the mean number ± standard deviation of β-Gal-positive neurons per section per group (Table 1). All mice had detectable viral antigen in peripheral lymphoid tissue. Figure 2b shows X-Gal-stained cervical lymph node tissue from SC16 110lacZ-infected mice. As the cervical lymph nodes drain the CNS and neck region these results indicated that HSV was effectively delivered to the immune system of all three mouse strains. In PrP^−/− mice β-Gal activity was seen only on day 8 p.i. in the tissues listed in Table 1. Tissues from wild-type mice showed the presence of β-Gal activity on day 6 and increased amounts on day 8 p.i. In the case of tga20 mice, most β-Gal activity was seen on day 4 p.i. with a decrease on days 6 and 8 p.i. Many of the tga20 β-Gal samples were significantly greater (at either P < 0.01 or P < 0.05) than equivalent wild-type or PrP^−/− samples, while wild type samples, in particular brain stems, were significantly greater (at P values of either <0.01 or <0.05) than equivalent PrP^−/− samples as shown in Table 1. No β-Gal activity was seen in the contralateral ganglia of any mice on the days tested.

Reduced virus titers in PrP^−/− mice.

Mice that lacked PrP^c did not appear to be as permissive for HSV-1 replication as those mice that expressed normal or elevated levels of PrP^c. This was confirmed when acute viral titers were assessed. Figure 3 shows the acute viral titers in left and right TG for all three strains of mice (Fig. 3a and b) and cervical and axillary lymph nodes (Fig. 3c and d). Significantly higher virus titers were seen on day 6 p.i. with wild-type left and right TG than with equivalent PrP^−/− tissues (P < 0.01) and on days 4 and 6 p.i. with tga20 mice than with wild-type and PrP^−/− mice (P < 0.01). In contrast, significantly lower virus titers were seen on day 8 p.i. with tga20 mice than with wild-type and PrP^−/− mice (P < 0.01). Significantly higher virus titers were seen for cervical lymph nodes on days 4 and 6 p.i. with tga20 than with equivalent wild-type and PrP^−/− tissues (P < 0.01) and significantly lower titers were evident on day 8 p.i. with tga20 mice than with wild-type and PrP^−/− mice (P < 0.01). Significantly higher virus titers were seen for axillary lymph nodes on days 6 and 8 with wild-type mice than with PrP^−/− tissue (P < 0.01) and on day 8 than with tga20 mice. Significantly higher virus titers were seen for axillary lymph nodes on day 6 with tga20 mice than with wild-type and PrP^−/− tissue (P < 0.01). Table 2 shows the acute viral titers for the left and right third cervical dorsal root ganglia and brain stem and brain samples with all three mice strains. All mice had infectious virus present by day 4 p.i., and this subsequently increased by day 6. At both of these time points, the level of infectious virus in the PrP^−/− mice was reduced between 1 and 2 logs compared to the level seen in those strains that expressed PrP^c. On day 8 p.i., the level of infectious virus in PrP^−/− and wild-type mice continued to rise. However, for tga20 mice, virus titers detected in the brain and brain stem samples (Table 2) and all the tissues shown in Fig. 3 were significantly decreased compared to those of earlier time points. These trends were seen in both the ipsilateral and contralateral TG with higher viral titers seen in the ipsilateral side (those from the same side as the ear pinna inoculation). These data indicated that virus replication appeared to peak at an earlier time point for tga20 mice. Similarly, infectious virus titers were higher for the tga20 mice on day 6 p.i. than for the other mouse strains but were reduced on day 8 p.i. in the draining lymph nodes (Fig. 3c and d) and the brain and brain stem (Table 2). All samples from tga20 mice were significantly different from all equivalent samples from wild-type and PrP^−/− mice at a P value of <0.01, with the exception of LCIII samples on day 8 p.i., which were significantly different from the wild type at a P value of <0.05. All wild-type samples were significantly different from PrP^−/− at a P value of <0.01, with the exception of LCIII samples on day 8, which were significantly different at a P value of <0.05, while brain stem and brain samples on days 6 and 8 were not significantly different. The greater severity of clinical signs exhibited by HSV-1-infected tga20 mice may reflect a failure of these mice to clear the virus, with a resultant increase in viral pathogenesis. We have established that tga20 mice have high numbers of γδ T cells (A. M. Thackray and R. Bujdoso, unpublished data), which are known to contribute to anti-HSV viral immunity (51). When infectious virus was measured in peripheral lymphoid tissue, it was clear that while tga20 mice responded quicker to the viral challenge, levels of infectious virus at these sites in the tga20 mice were never significantly greater than those seen in mice with lower PrP^c expression. This implied that tga20 mice were as effective as the other mouse strains at neutralizing HSV when it was presented to the immune system. Collectively, these data suggested that expression of PrP^c by neurons might be directly responsible for the observed neuropathology when these cells are invaded by a lytic neurotropic pathogen such as HSV.

FIG. 3.

Acute virus titers in TG and peripheral lymph nodes. Various mice were inoculated with 10⁶ PFU of SC16 110lacZ by intradermal inoculation of the left ear pinna. Acute viral titers were assessed for tissues isolated from mice on days 4, 6, and 8 p.i. Three mice per group and one mock-infected control were tested, and the results are shown as the mean virus titer (log₁₀ PFU) per tissue per group ± standard deviation. The limit of sensitivity for this test is 0.5 log₁₀ PFU per tissue. (a) Left TG. (b) Right TG. (c) Cervical lymph nodes. (d) Axillary lymph nodes.

TABLE 2.

Acute viral titers in various tissues following intradermal inoculation with HSV^a

Tissueb	Mean virus titer ± SD (log10 PFU) at indicated day p.i.
	PrP−/−			Wild type			tga20
	4	6	8	4	6	8	4	6	8
LCIII	0.4 ± 0	2.4 ± 0.2	3.9 ± 0.2	2.7 ± 0.2	3.2 ± 0.1	4.3 ± 0.2	3.5 ± 0.1	4.2 ± 0.1	4.8 ± 0.1
RCIII	0.4 ± 0	0.4 ± 0	0.4 ± 0	2.1 ± 0.1	2.8 ± 0.3	2.8 ± 0.1	3.3 ± 0.1	4.3 ± 0.1	3.6 ± 0.3
Brain stem	3.5 ± 0.1	4.4 ± 0.1	4.7 ± 0.1	4.3 ± 0.1	4.6 ± 0.2	4.7 ± 0.1	5.6 ± 0	5.7 ± 0	3.5 ± 0.5
Brain	3.7 ± 0.1	3.9 ± 0.2	5.2 ± 0.2	4.2 ± 0.2	4.3 ± 0.1	5.4 ± 0.4	5.5 ± 0.1	5.6 ± 0.1	2.8 ± 0.2

^aMice were inoculated with 10⁶ PFU of SC16 110lacZ; three mice and one mock-infected control were tested per group. The limit of sensitivity was 0.5 log₁₀ PFU per tissue.

^bLCIII, left third cervical dorsal root ganglia; RCIII, right third cervical dorsal root ganglia.

Numerous TUNEL-positive neurons in tga20 HSV-infected mice.

HSV infection leads to virally induced apoptosis in neural tissue despite the fact that this viral genome encodes viral analogues that act as cellular antiapoptotic proteins. We therefore examined the extent of apoptosis in the different mouse strains by TUNEL to see if PrP^c expression affected neuronal survival. The staining patterns observed for the CNS differed greatly among mouse strains. Figure 4 shows representative pictures from the various tissues examined for TUNEL stain, and Tables 3 and 4 show the quantified data. When sections of the brain stem were examined (Fig. 4a), numerous TUNEL-positive neurons could be seen for all three mice strains. PrP^−/− samples contained weakly stained neurons at the base of the brain stem with a narrow band of positive cells around the outside of the tissue. Wild-type mice had more stain in individual neurons, with a thicker band of positive peripheral cells. In contrast, apoptotic neurons were seen all over the brain stem in tga20 samples. Tga20 appeared to have approximately half the number of neurons per section that PrP^−/− or wild-type mice had (Table 4). While all three strains of mouse had comparable numbers of TUNEL-positive neurons, which gave rise to similar percentages of positive cells, the appearance of apoptotic cells was different. Many TUNEL-positive tga20 nuclei were tightly packed with dense TUNEL precipitate, resulting in cells which appeared to be two or three times larger than those in the other two mouse strains. Figure 4b shows that the dentate gyrus from both PrP^−/− and wild-type mice was negative for TUNEL, as was the cortex (Fig. 4c). However, the tga20 samples showed approximately 25% apoptotic neurons in the brain. Numerous positive cells were seen in the hippocampal ribbon of tga20 mice. In contrast, TG from all three strains of mice were negative for TUNEL (Fig. 4d), even though all the ganglia contained infectious virus (Tables 1 and 2). Lack of HSV-induced apoptosis in TG is consistent with other previously published data (55).

FIG. 4.

Apoptosis in HSV-infected neural tissue. Various mice were inoculated with 10⁶ PFU of SC16 110lacZ by intradermal inoculation of the left ear pinna. Representative sections are shown for each mouse strain demonstrating apoptosis as detected by TUNEL on day 6 p.i. (a) Brain stem. (b) Dentate gyrus. (c) Cerebral cortex. (d) Left TG. Magnification, ×200.

TABLE 3.

TUNEL assay results for brain and brain stem

Tissue	No. of TUNEL-positivea/total no. of neuronsa (% apoptotic)
	PrP−/−	Wild type	tga20
Brain	0/5,214 ± 576 (0)	0/5,308 ± 277 (0)	654 ± 46/2,550 ± 441 (25.60)
Brain stem	119 ± 80/1,018 ± 395 (11.70)	157 ± 55/1,003 ± 219 (15.70)	112 ± 91/977 ± 333 (11.50)

^aMean number per section ± standard deviation.

TABLE 4.

TUNEL assay results for different brain sections^a

Section	% Apoptotic neurons
	Brain			Brain stem
	PrP−/−	Wild type	tga20	PrP−/−	Wild type	tga20
A	0	0	27.7	10.5	12.7	6.1
B	0	0	30.2	16.3	12.4	5.6
C	0	0	19.6	7.6	17.4	21.8
D	0	0	27.3	10.5	19.3	9.3

^aBrain and brain stem samples were divided by Frontal cuts into quarters labeled A, B, C, and D, with A being the front section of each tissue. Twenty nonconsecutive sections were tested for each of the quarters.

Increased number of latent HSV-infected neurons in PrP^−/− mice.

A feature of HSV infection is the establishment of latency in sensory neurons such as those of the TG. Latent HSV may be detected by in situ hybridization for major LATs, which are believed to mediate the establishment of latency (45, 55). Surviving mice were sampled 1 month p.i., and alternate tissue sections were tested with a DIG-labeled riboprobe for major LAT to obtain the data shown in Table 5. LAT-positive neurons were detected in the ipsilateral and contralateral trigeminal and CIII ganglia and the brain stem and brain for all three mouse strains. PrP^−/− mice contained higher numbers of LAT-positive neurons than wild-type mice in all the tissues tested. The tga20 sections showed fewer positive neurons than did either wild-type or PrP^−/− mice. All samples from PrP^−/− mice gave significantly greater numbers than those of equivalent samples from wild-type mice (P < 0.01). All numbers from wild-type samples, with the exception of right TG, were significantly greater than those of equivalent samples from tga20 (P < 0.01). The intensity of LAT-positive signal, which is believed to be proportional to the number of genome copies of HSV present within neurons, varied considerably among mouse strains. While tga20 mice showed few LAT-positive neurons, approximately 80% of the positive neurons were jet black in color. In contrast, PrP^−/− mice that showed the greatest number of positive neurons did so with the majority displaying weak brown staining. The wild-type samples had a mixture of weakly and strongly stained LAT-positive neurons. These data suggest that a lack of PrP^c expression favors the establishment of HSV latency while PrP^c expression appears to be permissive for viral replication, with the extent of replication dependent upon the extent of PrP^c expression. Sections of ganglia, brain stem, and brain from mock-infected mice processed identically were completely negative for major LAT and only showed some weak background staining. RNase-treated, latently infected sections were negative (data not shown). A sense riboprobe also gave no positive neurons when tested on latently infected tissues. DNase-treated, latently infected sections gave positive staining (data not shown).

TABLE 5.

Increased latent HSV in PrP^−/− mice^a

Tissueb	Mean no. ± SD of LAT-positive neurons
	PrP−/−	Wild type	tga20
Left TG	44 ± 5	31 ± 2	6 ± 1
Right TG	12 ± 3	2 ± 1	2 ± 1
LCIII	40 ± 10	23 ± 8	4 ± 3
RCIII	23 ± 6	15 ± 7	2 ± 2
Brain stem	128 ± 34	82 ± 16	13 ± 4
Brain	185 ± 23	96 ± 14	27 ± 12

^aMice were inoculated with 10⁶ PFU of SC16 110lacZ; latent virus was detected by in situ hybridization in alternate 5-μm sections with a DIG-labeled major LAT riboprobe (see Materials and Methods).

^bLCIII, left third cervical dorsal root ganglia; RCIII, right third cervical dorsal root ganglia.

Increased microvacuolation in latently infected tga20 mice.

There was a clear correlation between the lack of latent in situ hybridization and TUNEL staining in the HSV-infected tga20 mice. Many neurons appeared to have been destroyed during the acute phase of the infection, but those that survived were loaded with LATs. The inference was that HSV infection had resulted in virally induced apoptosis with a subsequent loss of neurons from infected mice. Sections of brain stem were stained with H&E and visualized microscopically (Fig. 5c). It was clear that sections from tga20 mice had an increased microvacuolation typical of damage as a result of apoptosis of neurons. This type and extent of vacuolation was routinely seen in the brain stems of HSV-infected tga20 mice and to a much lesser extent in the other two mice strains, where it was regarded as tissue damage.

DISCUSSION

We have analyzed the effect of HSV inoculation in mice that have different levels of PrP^c protein expression. When mice were inoculated with wild-type SC16, PrP^−/− mice did not succumb to virus-induced mortality, unlike those that expressed normal or elevated levels of PrP^c protein. When PrP^−/− and wild-type mice were inoculated with HSV-1 SC16 110lacZ, these animals showed similar mild clinical signs as a consequence of virus inoculation. SC16 110lacZ is regarded as less pathogenic than wild-type HSV-1 (53), and neither mouse strain succumbed to virus-induced mortality. However, analysis of the distribution of SC16 110lacZ RNA and protein expression clearly showed that virus replication appeared to proceed less efficiently in PrP^−/− mice than in wild-type mice. Following virus inoculation, neural tissues from wild-type mice were found to contain a greater number of cells with either ICP0 transcripts or β-Gal activity at earlier time points than PrP^−/− mice. These results indicated that PrP^−/− mice were not as permissive for HSV-1 viral replication as were wild-type mice. This view was supported by measurement of acute virus titers in neural tissues from HSV-1-infected mice. Both PNS and CNS tissues from wild-type mice contained higher levels of infectious virus at each of the time points analyzed than did samples from PrP^−/− mice. These data showed that the magnitude of the acute phase of HSV-1 infection in PrP^−/− mice was reduced compared to that seen in mice which expressed PrP^c.

The implication of this observation is that neural tissue with elevated levels of PrP^c expression may be expected to show increased pathology as a consequence of HSV-1 infection. This appeared to be the case when tga20 mice, which express approximately 10-fold more PrP^c in neural tissue than do wild-type mice, were analyzed for their response to HSV-1 inoculation. In HSV-1-inoculated tga20 mice, all aspects of viral pathogenicity were exacerbated compared to that seen in PrP^−/− and wild-type mice. Significantly, tga20 mice showed mortality after inoculation with SC16 110lacZ. The onset of clinical signs and progression to terminal neurological disease in these virus-inoculated mice were extremely rapid and occurred in a matter of hours. Normally, mice susceptible to HSV-1-induced neurological disease progress through a series of well-defined clinical signs and do so over a time course of 2 to 3 days (17). The rapid progression of clinical signs in tga20 mice was probably coupled with enhanced viral replication that also appeared to proceed very efficiently in these mice. Most significantly, tga20 mice that survived the HSV-1 infection had a reduced number of LAT-positive neurons compared to PrP^−/− and wild-type mice. The number of latently infected neurons in tga20 mice was approximately 10-fold less than that seen in PrP^−/− mice. In addition, LAT-positive neurons present in PrP^−/− mice were mainly light brown in color. This indicated that these neurons contained lower levels of latent HSV-1 nucleic acid than did latently infected neurons in wild-type or tga20 mice. Collectively, our results strongly suggest that expression of PrP^c by neural tissue correlates with an increased permissiveness for the acute phase of HSV-1 infection and that lack of PrP^c expression favors the establishment of HSV-1 latency.

Detection of LAT RNA by in situ hybridization is one of several techniques which are routinely used to assess latency. This assay provides an estimate of the number of cells which contain the viral genome, and it does not allow quantification of the number of viral-genome copies within individual latently infected cells. Several strategies may be employed to determine the number of viral genomes in DNA harvested from latently infected tissues. These include quantitative PCR that allows measurement of the total amount of viral DNA (24) or in situ PCR which has been used to detect latent viral genomes in sectioned mouse ganglia (33, 35). Using a PCR approach, Mehta et al. demonstrated that more neurons contained the viral genome than contained detectable levels of LATs measured by in situ hybridization (35). A quantitative analysis of viral nucleic acids within the individual cells constituting complex solid tissues is also possible (50). It will be important to characterize HSV-1 latently infected tissues from mice with various levels of PrP^c expression to correlate quantitatively the number of cells harboring the viral genome and the number of viral genomes contained within latently infected cells. This level of cellular quantification will be useful in clarifying the role of PrP^c in the establishment and maintenance of HSV-1 latency.

Clearly, a complex interaction must exist between HSV and the host to determine the fate of virus-infected cells. It is likely that the most important event for an infected neuron is to limit transcription of viral lytic phase genes, if it can, to ensure that the infection becomes latent. The combinations of viral and cellular factors that regulate this event have yet to be fully described. Viral components believed to participate in this event include the LAT locus (45, 55) and the US3 gene-encoded protein kinase (32). The LAT locus is proposed to regulate viral protein expression in neurons as the acute phase of infection proceeds (55). The LAT promoter contains neuron-specific expression elements, which are likely to function together with neuronal factors to regulate viral gene expression (3, 4, 61, 62). Candidate cellular factors involved in this role include c-fos, c-jun, Nf-kB, and p53, which can be detected in HSV-infected neurons (32, 60). This combination of cellular factors is not seen in neurons from TG, which are regarded as resistant to HSV-induced apoptosis. As latent infection at this site does lead to increased tumor necrosis factor alpha expression in neurons (10), it can be inferred that different levels of such factors may be responsible for the establishment of productive or latent HSV infection in cells. This balance would appear to be influenced by PrP^c, which may constitute part of a metabolic pathway within neurons, one which creates a cellular environment that is permissive for viral replication rather than latency. In this regard, it is relevant to note that one of the proposed binding proteins for PrP^c is the proapoptotic protein Bcl-2 (28, 29) and secondly that PrP^c has been reported to be functionally linked to the tyrosine kinase Fyn (37).

In our study we have utilized the TUNEL assay, which detects fragmented DNA, to measure apoptosis. It has been suggested that TUNEL identifies cells that have undergone not only apoptosis but also necrosis and autolysis and may therefore overestimate apoptosis (20). However, a comparison of TUNEL and laser scanning confocal microscopy has shown that TUNEL is capable of faithful identification and quantification of apoptosis (26, 44). Other markers of apoptosis may now be assessed with elucidation of the enzyme cascades activated by mediators of apoptosis such as tumor necrosis factor alpha and the Fas ligand (FasL)-Fas receptor system. The Fas signaling pathway is particularly well understood (40). Upon ligand engagement, Fas is trimerized, which facilitates the recruitment of FADD to its cytoplasmic domain (11), which in turn recruits and activates procaspase 8 via a death effector domain interaction. This scheme links the death domains of early signaling proteins to activation of the caspase cascade, including caspase 3 and caspase 7 (39). These enzymes are involved in cleavage events that result in activation of DNase, poly-ADP-ribose polymerase, and various structural proteins including actin, fodrin, and lamin (15, 25, 38, 49). A more complete assessment of apoptosis may be achieved through measurement of specific caspases as their enzyme activity leads to the generation of new protein epitopes that occur only during apoptosis. It will be of interest to examine the expression of caspase enzymes and their activity in HSV-1-mediated apoptosis of PrP-expressing tissue in order to identify the signal transduction pathway with which PrP^c is associated.

If PrP^c is involved in a signal transduction pathway and its expression by neurons does promote virus replication, its action may be expected to be deleterious to neuronal survival following infection by lytic viruses such as HSV-1. Following HSV inoculation, tga20 displayed enhanced virally induced apoptosis. Neural tissues from PrP^−/−, wild-type, and tga20 mice all showed apoptotic neurons in the brain stem, but only tga20-infected mice showed apoptotic neurons in brain tissue. An explanation therefore for the dramatic and accelerated onset of terminal neurological disease in HSV-infected tga20 mice may be a triggering of rapid and extensive apoptosis of essential neurons. While tga20 CNS neurons showed enhanced susceptibility to HSV-1-induced apoptosis, this effect was not a general property of all neurons in these mice. For example, TG neurons are resistant to HSV-1-induced apoptosis (55) and are a principal site of HSV-1 latency in natural and experimental HSV infections. However, TG from tga20 mice, like those from PrP^−/− and wild-type mice, did not contain any apoptotic neurons following HSV infection. Therefore, despite tga20 having enhanced expression of PrP^c, the site-specific regulation of HSV-1 pathogenicity, such as that seen between the PNS and CNS, is maintained.

While our data support a direct role for PrP^c in the regulation of HSV-1 pathogenicity, an alternative explanation is that the expression of PrP^c by the host immune system modulates its ability to clear the virus with accompanying differences in virus-induced immunopathology. In a separate study, we have compared immune cell subset composition in peripheral lymphoid tissue of mice that express different levels of PrP^c protein. PrP^−/− and wild-type mice have similar numbers of T cells, B cells, NK cells, and dendritic cells, while tga20 mice have a low number of αβ T cells and an increased number of γδ T cells. Despite the similarity in immune composition, PrP^−/− mice in our study clearly showed more resistance to HSV-1 infection than did wild-type mice. Furthermore, in studies with wild-type HSV-1 (SC16), mice with normal or elevated levels of PrP^c protein succumbed to virus-induced mortality whereas PrP^−/− mice did not. This was somewhat surprising with respect to tga20 mice, as others have shown that γδ T cells are capable of mediating antiviral immunity and can provide protection against HSV-1 pathogenicity even in the complete absence of αβ T cells (51). The fact that HSV-infected tga20 mice progressed to mortality much more rapidly than is usually seen with mice susceptible to this virus, such as BALB/c (17), further supports the view that events other than immune intervention induce the large amount of apoptosis seen for these mice. Increased PrP^c expression may inadvertently prime neurons to undergo apoptosis in the presence of a neurotropic pathogen.

Our report highlights a novel phenotypic difference between PrP^−/− mice and those that express PrP^c protein; namely, that lack of PrP^c expression favors the establishment of HSV latency whereas HSV replication proceeds more efficiently in neuronal tissue that expresses PrP^c. We have yet to establish the regulatory pathways within neurons that are influenced by PrP^c, but these initial experiments illustrate the utility of using HSV to address the function of this protein. It has been reported that PrP^c is involved in the regulation of neuronal oxidative stress. However, productive HSV-1 infection of CNS tissue and latency in TG is reported to cause oxidative stress at both of these sites (58). If PrP^c does have an important role in oxidative stress it may be expected that PrP^−/− mice would show a more severe pathology during acute HSV infection than do wild-type mice, but this was not seen in our study. The inference of our data is that PrP^c may indirectly modulate neuronal gene expression possibly through a signal transduction pathway and in doing so render these cells susceptible to HSV replication. In this scheme, it is clearly beneficial for HSV to invade cells that lack or have a low level of PrP^c expression, as this allows virus survival through promotion of latency. Conversely, events that increase PrP^c expression by neurons may render them more susceptible to viruses such as HSV. The implication of this is that during prion diseases, when levels of PrP protein expression are increased, individuals may be more susceptible to the cytopathic effects of lytic CNS viruses such as HSV, which in turn may enhance prion disease pathogenesis.