Lymphocyte maturation is not completely blocked in some strains of scid mice 414243. Thus, B and T lymphocyte function were evaluated in scid mice. Assessment of B cell function indicated that BALB/c mice had serum IgG levels of 6.4 ± 1.3 mg/ml, whereas serum IgG was undetectable in scid mice (Fig. 1A). Flow cytometric analysis indicated that BALB/c mice contained an average 110 million WBCs per spleen, of which 21% were CD4⁺ T cells, 10% were CD8⁺ T cells, and 2.5% were CD3^- CD49b⁺ NK cells (Fig. 1B). In contrast, scid mice contained an average 8 million WBCs per spleen, of which <0.1% were CD4⁺ or CD8⁺ T cells and 45% were CD3^- CD49b⁺ NK cells (Fig. 1B).

Adoptive transfer was performed to verify that adaptive resistance to HSV-1 could be restored to scid mice. Following ocular inoculation with HSV-1 strain KOS, scid mice shed high titers of virus between 1 and 7 days p.i. (Fig. 1C). On day 7 p.i., scid mice were i.v. administered either a. vehicle, b. total WBCs, c. purified B cells, or d. purified T cells from naïve BALB/c donors (Fig. 1C). Vehicle-treated scid mice continued to shed high levels of virus (Fig. 1C) and succumbed to the infection within 17 ± 2 days p.i. (Fig. 1D). Scid mice reconstituted with total WBCs shed 30-fold less virus than vehicle-treated controls on day 14 p.i. (Fig. 1C) and 8 of 8 survived the infection (Fig. 1D). Scid mice reconstituted with purified B cells eventually died, but the mean time of survival was increased to 22 ± 3 days (Fig. 1D). Reconstitution with purified T cells controlled HSV-1 infection in 8 of 8 scid mice, but viral shedding continued for ~3 days longer than scid mice reconstituted with total WBCs (Fig. 1C). Thus, all measures indicated that scid mice are effectively devoid of B and T lymphocyte function.

To determine if innate resistance to HSV-1 is dependent on NK cell cytotoxic function, infection with HSV-1 strain KOS was compared in BALB/c scid mice versus non-obese diabetic (NOD) scid mice. Consistent with previous reports 4445, WBCs isolated from the spleens of NOD scid mice were functionally deficient in NK cell cytotoxic activity relative to BALB/c mice and BALB/c scid mice (Fig. 2A). Following ocular inoculation with 2 × 10⁵ pfu/eye, HSV-1 strain KOS replicated to high and equivalent titers in BALB/c scid mice and NOD scid mice between 1 and 14 days p.i. (not shown). No differences were observed in the progression of viral pathogenesis or the duration of survival of BALB/c scid mice versus NOD scid mice (Fig. 2B). Flow cytometry demonstrated that approximately one-third of the peripheral WBCs of NOD scid mice possessed the CD3^- CD49b⁺ phenotype of NK cells (Fig. 2C) 4647. Thus, despite the lack of in vitro cytotoxic activity (Fig. 2A), NOD scid mice still possessed significant numbers of NK cells that could control HSV-1 infection via other mechanisms (e.g., IFN-γ secretion).

Preliminary experiments indicated that two treatments with 0.32 or 1.0 mg rabbit anti-asialo GM1 reduced the number of NK cells in BALB/c scid mouse spleens by >10- and >50-fold, respectively, whereas control rabbit IgG produced no such effect (Fig. 3). Thus, anti-asialo GM1 antibody was used to determine if NK cells are necessary for innate resistance to HSV-1.

BALB/c mice and BALB/c scid mice were treated with PBS, control IgG, or anti-asialo GM1 and were inoculated with 2 × 10⁵ pfu/eye of HSV-1 strain KOS. In BALB/c mice, KOS replicated to similar viral titers in mice treated with PBS, control rabbit IgG, or anti-asialo GM1, with one notable exception (Fig. 4A). On days 5 and 7 p.i., BALB/c mice treated with rabbit anti-asialo GM1 shed ~5-fold more virus than PBS-treated controls (Fig. 4A; p < 0.05, denoted by asterisks). Between days 9 and 14 p.i., viral shedding was detected in none of the BALB/c mice (Fig. 4A). Likewise, viral pathogenesis was limited, and 100% of BALB/c mice survived infection with HSV-1 strain KOS (Fig. 4B).

BALB/c scid mice shed infectious KOS continuously during the 14-day sampling period (Fig. 4A). Treatment with rabbit anti-asialo GM1 had no effect on the titers of infectious KOS recovered from the eyes of BALB/c scid mice between 1 and 14 days p.i. (Fig. 4A). Scid mice treated with PBS survived for 16 ± 1 days p.i. (Fig. 4B). Treatment with rabbit anti-asialo GM1 did not shorten the duration of survival of KOS-infected scid mice (Fig. 4B). Paradoxically, treatment with rabbit anti-asialo GM1 or control rabbit IgG increased the duration of survival of KOS-infected BALB/c scid mice to 24 ± 1 and 35 ± 3 days p.i., respectively (Fig. 4B; p <0.001). Multiple experiments confirmed this unexpected effect that rabbit immunoglobulin (with no reactivity against HSV-1) prolonged the survival of KOS-infected scid mice. The interpretation of these data was complicated by this caveat. However, it was clear that NK cell depletion did not fundamentally alter the progression of HSV-1 infection in scid mice during the first week p.i.

Cyclophosphamide (CyP) is an alkylating agent that is rapidly converted in vivo into metabolites that cause lethal DNA damage in dividing cells 4849, and transiently reduce peripheral WBC counts by ~90% in mice 31. To determine if WBCs are necessary for innate resistance to HSV-1, BALB/c mice and scid mice were treated with PBS or CyP and were inoculated with 2 × 10⁵ pfu/eye of HSV-1 strain KOS. On day 4 p.i., peripheral WBC counts (WBCs per ml × 10⁶) in each group were, as follows: BALB/c + PBS = 6.6 ± 0.5; scid + PBS = 2.3 ± 0.2; BALB/c + CyP = 0.8 ± 0.1; and scid + CyP = 0.4 ± 0.1. Similar viral titers were recovered from the eyes of all mice at 24 hours p.i. (Fig. 5A). However, BALB/c mice treated with CyP shed 30- to 1000-fold more virus than PBS-treated BALB/c mice between 5 and 9 days p.i. (Fig. 5A; p < 0.05, denoted by asterisks). Likewise, CyP-treated scid mice shed 2- to 7-fold more virus than PBS-treated scid mice between 5 and 9 days p.i. (Fig. 5A). Viral titers were not determined in CyP-treated mice on 11 and 14 days p.i. because the extent of ocular pathogenesis precluded a reliable measurement.

BALB/c mice treated with PBS uniformly survived ocular HSV-1 infection (Fig. 5B). In contrast, 0% of CyP-treated BALB/c mice survived HSV-1 infection (Fig. 5B). The death of these mice was not a direct consequence of CyP's toxicity, because 100% of uninfected BALB/c controls survived the same course of CyP treatment (Fig. 5B). PBS-treated scid mice survived ocular inoculation with HSV-1 strain KOS for 18 ± 1 days (Fig. 5C). In contrast, CyP-treated scid mice succumbed to HSV-1 infection within 12 ± 1 days (Fig. 5C; p < 0.001). This reduced duration of survival was not a direct consequence of CyP's toxicity, because 7 of 8 uninfected scid mice survived CyP treatment (Fig. 5C). Therefore, depletion of total WBCs in scid mice was correlated with decreased innate resistance to HSV-1 infection.

The innate resistance of scid mice to HSV-1 infection was not adversely affected by NK cell depletion, but was impaired by CyP-induced depletion of total WBCs (Table I). To assure that inter-experimental variance was not the source of these differences, the effect of NK cell versus total WBC depletion was directly compared in scid mice infected with KOS-GFP, a GFP-expressing recombinant virus 50. Scid mice were treated with PBS, rabbit IgG, anti-asialo GM1, or CyP and were inoculated with 2 × 10⁵ pfu/eye HSV-1 strain KOS-GFP. GFP expression provided a measure of the extent of KOS-GFP spread in scid mice (Fig. 6A). Anti-asialo GM1 antibody treatment caused a >20-fold reduction in NK cell abundance, as determined in n = 2 KOS-GFP-infected scid mice sacrificed on day 5 p.i Despite effective depletion of NK cells, neither treatment with control rabbit IgG nor anti-asialo GM1 had a measurable effect on KOS-GFP spread in the eyes or periocular skin of scid mice during the first 6 days p.i. (Fig. 6A). In contrast, CyP treatment enhanced the spread of KOS-GFP into the periocular skin of scid mice on day 6 p.i. relative to the other treatment groups (Fig. 6A).

PBS-treated scid mice infected with HSV-1 strain KOS-GFP survived for 22 ± 1 days p.i. (Fig. 6B). NK cell depletion with anti-asialo GM1 did not decrease the duration of survival of HSV-1 infected scid mice (Fig. 6B). Rather, treatment with control IgG or anti-asialo GM1 increased the duration of survival of KOS-GFP infected scid mice (Fig. 6B; 38 ± 3 and 35 ± 2 days p.i, respectively). In contrast, treatment with CyP significantly reduced the duration of survival of HSV-1-infected scid mice (Fig. 6B; p < 0.001; 14 ± 0.2 days p.i.). The reduced duration of survival was not due to CyP's toxicity, because 7 of 7 uninfected scid mice survived CyP treatment (Fig. 6B). Thus, while depletion of total WBCs was correlated with decreased innate resistance to HSV-1 infection, depletion of NK cells had no such effect.

Stat 1 is an IFN-activated transcription factor that is essential for the intracellular response of cells to the cytokines IFN-α/β and IFN-γ 3940. Lymphocyte-deficient rag2^-/- mice, which were genetically stat1^+/+ versus stat1^-/-, were inoculated with 2 × 10⁵ pfu/eye HSV-1 strain KOS-GFP. As controls, wild-type strain 129 and stat1^-/- mice were also inoculated with KOS-GFP. At 24 hours p.i., GFP expression (Fig. 7A) and infectious KOS-GFP (Fig. 7B) were detected in the eyes of all mice. Between 48 and 96 hours p.i., GFP-expression steadily decreased in the eyes of strain 129 mice and rag2^-/- mice infected with KOS-GFP (Fig. 7A). In contrast, GFP expression continued to spread in the eyes of stat1^-/- mice and rag2^-/-stat1^-/- mice such that 25 to 50% of the ocular surface was GFP-positive by 72 hours p.i. (Fig. 7A). Likewise, stat1^-/- and rag2^-/-stat1^-/- mice shed ~300-fold more virus on day 3 p.i. than wild-type or rag2^-/- mice (Fig. 7B). This rapid response at the site of inoculation was not lymphocyte-dependent, because wild-type mice and rag2^-/- mice shed equivalent, low titers of KOS-GFP on day 3 p.i. (Fig. 7B). Viral titers were not determined in stat1^-/- mice or rag2^-/-stat1^-/- mice on day 7 p.i. because the extent of ocular pathogenesis precluded a reliable measurement.

Strain 129 (wild-type) mice uniformly survived KOS-GFP infection (Fig. 7C). In contrast, 0% of rag2^-/- mice survived and their duration of survival was 25 ± 2 days p.i. Thus, the duration of survival of KOS-GFP-infected rag2^-/- mice was similar to KOS-GFP-infected scid mice (i.e., 21 ± 3 days; Table I). Rag2^-/- stat1^-/- mice succumbed to HSV-1 infection much more rapidly than rag2^-/- mice, and survived for only 7.8 ± 0.4 days after inoculation with KOS-GFP (Fig. 7C). Likewise, stat1^-/- mice also succumbed to HSV-1 by 7.8 ± 0.8 days p.i., presumably because the viral infection overwhelmed these mice before an adaptive immune response could be mounted. Collectively, the results indicate that innate resistance to HSV-1 infection is intimately dependent on Stat 1-induced gene expression.

Despite a large differential in NK cell cytotoxicity, NOD scid mice possessed an innate resistance to HSV-1 infection that was equivalent to BALB/c scid mice. One interpretation of the results is that NK cell function is not essential for innate resistance to HSV-1. However, the fact that peripheral WBCs of NOD scid mice lack detectable NK cell cytotoxic activity in vitro does not prove that NOD scid mice are devoid of NK cell function in vivo. Indeed, one-third of peripheral WBCs in NOD scid mice were found to be CD3^- CD49b⁺ NK cells. Thus, we were hesitant to use the behavior of HSV-1 in NOD scid mice as the basis for concluding that NK cells play no role in innate resistance to HSV-1. Likewise, we question the validity of comparisons of HSV-1 infection in mice that exhibit "high" or "low" natural cytotoxicity in vitro, because there is no evidence that these mice lack NK cell function in vivo 2151.

T cells that become activated in response to viral infections express the "NK cell" markers asialo GM1, NK1.1, and CD49b (i.e., antigen recognized by DX5 monoclonal antibody; Ref. 47). Thus, depletion of asialo GM1⁺ T cells or NK1.1⁺ T cells may account for the capacity of anti-asialo GM1 or anti-NK1.1 to impair the resistance of BALB/c and C57BL/6 mice to HSV-1 infection 122627. Consistent with this hypothesis, anti-asialo GM1 increased ocular viral titers in BALB/c mice on days 5 and 7 p.i., but produced no such effect in scid mice (Fig. 4). In BALB/c scid mice, NK cell depletion had no effect on (a) ocular viral titers or (b) the rate of KOS-GFP spread to the periocular skin. An important caveat of the NK cell depletion experiments was that control IgG or anti-asialo GM1 had the unanticipated effect of prolonging the survival of HSV-1 infected scid mice. Neither IgG preparation possessed neutralizing activity or reactivity with HSV-1 antigens by ELISA. This effect raises questions about the homeostatic mechanisms and Fc-γ-receptor dependent processes that are influenced when IgG is introduced into a scid mouse for the first time in its life 52535455. However, the relevant point for this study is that a >95% reduction in NK cell abundance does not impair the capacity of a scid mouse to limit HSV-1 spread from the site of inoculation.

The lack of effect of NK cell depletion on innate resistance to HSV-1 is only relevant if scid mice possess a measurable innate resistance to HSV-1. CyP-induced depletion of total WBCs in scid mice was associated with 1. increased HSV-1 spread from the site of inoculation (evident by day 6 p.i.) and a 2. reduced duration of survival. CyP impairs the immune response of normal mice to HSV-1 31. This study provides the first evidence that CyP can also be used to blunt the innate immune response to HSV-1. CyP has a narrow therapeutic window; there is only a 4-fold difference between the minimum effective dose and a fatal dose. Thus, we suspect that antibody-based depletion of the relevant WBC effector(s) would cause a more profound decrease in innate resistance to HSV-1. Based on recent evidence, professional APCs such as dendritic cells may be the relevant cellular targets whose depletion accounts for CyP's capacity to impair innate resistance to HSV-1 5657. Further study is needed to test this hypothesis.

The biological functions of IFN-α/β and IFN-γ are dependent on the phosphorylation of Stat 1, which results in Stat 1 dimerization, nuclear translocation, and transcriptional activation of IFN-stimulated genes 3958. HSV-1 infection was compared in rag2^-/- mice versus rag2^-/-stat1^-/- mice to determine if innate resistance to HSV-1 is dependent on Stat 1-induced gene expression. Profound differences in viral titers and viral spread were evident in stat1^-/- versus stat1^+/+ mice by 3 days p.i. The rapidity with which HSV-1 infection spread in the absence of Stat 1 (i.e., a relevant effector) underscored the remarkable lack of effect that NK cell depletion had on innate resistance to HSV-1. The defect in Stat 1 rendered rag2^-/- stat1^-/- mice incapable of limiting HSV-1 spread, and these mice succumbed to the viral infection just 7.8 ± 1 days p.i. In contrast, the weakly virulent KOS-GFP strain 50 caused a slowly progressing infection in rag2^-/- mice that was not lethal until 25 ± 2 days p.i. Thus, Stat 1 plays an integral role in innate resistance to HSV-1 infection. Likewise, a similar phenotype of uncontrolled HSV-1 spread has been observed in IFN-α/βR^-/- IFN-γR^-/- double knockout mice 2230. Thus, the available evidence suggests that IFN-α/β and IFN-γ are important activators of Stat 1-induced resistance to HSV-1 in vivo. However, further studies are necessary to determine if chemokines and pro-inflammatory cytokines are also important contributors to Stat 1-dependent innate resistance to HSV-1.

The wild-type HSV-1 strains KOS 65 and KOS-GFP, a KOS strain engineered to express GFP 50, were generously provided by Dr. Priscilla Schaffer and Dr. John Balliet (Harvard University Medical School, Boston, MA). The viruses were propagated in Vero cells (American Type Culture Collection, Manassas, VA) and stored as viral stocks at -70°C until used. Vero cells were propagated in Dulbecco's modified Eagle medium (DMEM) containing 10% fetal bovine serum (FBS), hereafter referred to as "complete DMEM." YAC-1 cells (American Type Culture Collection) were propagated in RPMI-1640 containing 10% FBS.

Female BALB/c, BALB/c scid, and NOD scid mice were obtained from the Jackson Laboratory (Bar Harbor, ME). Female strain 129 mice, rag 2^-/- mice, stat1^-/- mice, and rag2^-/- stat1^-/- mice were purchased from Taconic Farms (Germantown, NY). These studies were reviewed and approved by an IACUC, and animals were handled in accordance with the NIH Guide for the Care and Use of Laboratory Animals. Prior to ocular inoculation, mice were anaesthetized by intraperitoneal (i.p.) administration of xylazine (6.6 mg/kg) and ketamine (100 mg/kg). Mice were inoculated by scarifying the cornea with the blunted tip of a 25-gauge needle, blotting tear film from the eyes, and by placing 4 μl of complete DMEM containing 2 × 10⁵ pfu/eye of virus on each eye. Viral titers were determined by swabbing the ocular surface of both eyes at times after inoculation with a cotton-tipped applicator. The tip of the applicator was removed, incubated in 0.4 ml complete DMEM for 1 hour on ice, and viral titers were determined in the supernatant by a microtiter plate plaque assay. Following anaesthetization of mice, fluorescent images of mouse eyes infected with HSV-1 strain KOS-GFP were taken at 2× or 4× magnification on a Nikon TE300 fluorescent microscope (Nikon Instruments, Lewisville, TX).

Randomly chosen BALB/c, BALB/c scid, and NOD scid mice were tested to confirm that they were deficient for T- and B-cell function. Serum from mice was tested for immunoglobulin G (IgG) levels using an ELISA kit specific for the Fc fragment of mouse IgG (Bethyl Laboratories, Montgomery, TX). Flow cytometric analysis was used to measure the abundance of CD4⁺ T cells and CD8⁺ T cells in the spleens of selected mice, as described below.

Cells were harvested from the spleens of mice and red blood cells were removed by hypotonic lysis in 0.84% NH₄Cl. WBCs were labeled with fluorescent-labeled monoclonal antibodies obtained from BD Biosciences (San Jose, CA) according to the manufacturer's directions. For each mouse analyzed, 1 × 10⁶ WBCs were labeled with either 1. nothing, 2. fluorescein-isothiocyanate (FITC)-labeled anti-CD3 (clone 17A2) + phycoerythrin (PE)-labeled anti-CD4 (clone GK1.5), 3. FITC-labeled anti-CD3 + PE-labeled anti-CD8 (clone Ly-2), or 4. FITC-labeled anti-CD3 + PE-labeled anti-CD49b (clone DX5). Additionally, controls were included for gating and compensation, and these included WBCs labeled with 5. nothing, 6. FITC-labeled IgG_2b isotype control (clone G27-35), 7. PE-labeled IgG_2b isotype control (clone G27-35), 8. PE-labeled IgG_2a isotype control (clone G155-178), or 9. PE-labeled IgM isotype control (G155-258). Flow cytometry was performed immediately after antibody labeling on a FACSCalibur using CellQuest Pro software (BD Biosciences). A minimum of 10,000 events was recorded per sample. When NK cell depletion was monitored in mice, a minimum of 25,000 events were recorded per sample. The threshold between fluorescence-positive and -negative was set such that >99.5% of WBCs incubated with control antibodies were considered negative.

Unfractionated and purified WBCs were obtained from naïve BALB/c donors for adoptive transfer to BALB/c scid mice, as follows. Spleens and cervical lymph nodes were removed from ten naïve BALB/c mice and dissociated to yield a single cell suspension. WBCs were purified using Lympholyte M according to the manufacturer's directions (CedarLane Laboratories Ltd., Hornby, Ontario, Canada). Purified lymphocytes were obtained by passing BALB/c spleen WBCs through immunoaffinity B cell and T cell columns according to the manufacturer's directions (Cytovax Biotechnologies Inc., Edmonton, Alberta, Canada). Adoptive transfer of WBCs was achieved by intravenous (i.v.) tailvein injection of BALB/c scid mice with 0.5 ml complete RPMI-1640 containing nothing (vehicle), 5 × 10⁶ unfractionated WBCs (total WBCs), 5 × 10⁶ purified B cells, or 5 × 10⁶ purified T cells.

YAC-1 cells (1 × 10⁴ cells) were labeled with ⁵¹Cr and incubated with BALB/c, BALB/c scid, and NOD scid spleen WBCs in round bottom 96-well plates for 6 hours at 37°C at effector : target ratios of 100:1, 50:1, 25:1, 12.5:1, and 6.25:1 (n = 3 per group). One hundred microliters of supernatant were collected from each culture for determination of ⁵¹Cr release from target cells. Controls for the assay included 1 × 10⁴ cells target cells incubated alone in culture medium (spontaneous release) and target cells incubated with 0.5% Triton X-100 (maximal release). The percent cytotoxicity in each group of three replicate cultures was calculated, as follows:

Between days -1 and 10 p.i., BALB/c mice and BALB/c scid mice were given four to six i.p. injections of normal rabbit IgG (Rockland Immunochemicals, Gilbertsville, PA) or rabbit IgG containing anti-asialo GM1 antibody (Wako Chemicals USA, Richmond, VA). The efficacy of NK cell depletion was validated by flow cytometric comparison of NK cell (CD3^- CD49b⁺) frequency. Cyclophosphamide (Mead Johnson Oncology Products, Princeton, NJ) was diluted with phosphate-buffered saline (PBS) such that a volume of 0.25 ml delivered i.p. would achieve a dose of 125 mg/kg (e.g., 11 mg/ml for 22 g mice). Vehicle-treated controls received 0.25 ml PBS. Intraperitoneal injections of PBS or cyclophosphamide were administered on days -1, +1, and +3 after viral inoculation.

Analysis of numerical data and statistical analyses were performed with the software packages Microsoft Excel (Redmond, WA) and Modstat (Modern Microcomputers, Mechanicsville, VA). Unless otherwise indicated, all data are presented as means ± standard errors of the means (SEM). Viral titers were transformed by adding a value of 1 to the number of pfu per eye such that negative results (i.e. no plaques detected) could also be analyzed on a logarithmic scale. The significance of differences in viral titers between three or more groups was statistically evaluated by one way analysis of variance (ANOVA) followed by Tukey's post hoc t-test. The significance of differences in duration of survival between each treatment group and PBS-treated controls was evaluated by a two-way t-test.