|Home | About | Journals | Submit | Contact Us | Français|
Mutations in the genes that encode Fas or Fas ligand (FasL) can result in poor restraints on lymphocyte activation and in increased susceptibility to autoimmune disorders. Because these mutations portend a continuously activated immune state, we hypothesized that they might in some cases confer resistance to infection. To examine this possibility, the immune response to, morbidity caused by, and clearance of vaccinia virus (VACV) Western Reserve was examined in 5- to 7-week-old Fas mutant (lpr) mice, before an overt lymphoproliferative disorder was observable. On day 6 after VACV infection, C57BL/6-lpr (B6-lpr) mice had decreased morbidity, decreased viral titers, and an increased percentage and number of CD4+ and CD8+ T cells. As early as day 2 after infection, B6-lpr mice had decreased liver and spleen viral titers and increased numbers of and increased gamma interferon (IFN-γ) production by several different effector cell populations. Depletion of individual effector cell subsets did not inhibit the resistance of B6-lpr mice. Uninfected B6-lpr mice also had increased numbers of NK cells, γδ+ T cells, and CD44+ CD4+ and CD44+ CD8+ T cells compared to uninfected B6 mice. Antibody to IFN-γ resulted in increased virus load in both B6 and B6-lpr mice and eliminated the differences in viral titers between them. These results suggest that IFN-γ produced by multiple activated leukocyte populations in Fas-deficient hosts enhances resistance to some viral infections.
Autoimmune lymphoproliferative syndrome (ALPS) is a genetic disorder characterized by a chronic nonmalignant lymphadenopathy and/or splenomegaly, increased relative percentages of CD3+ TCRαβ+ CD4− CD8− (double-negative [DN]) T cells, and defective lymphocyte apoptosis. It is commonly associated with genetic mutations in FAS, FASL, or CASP10 (42). Mutations in the genes that encode either Fas or FasL have also been associated with non-ALPS autoimmune disorders such as systemic lupus erythematosus (SLE) (16, 26, 68, 69). Two commonly used mouse models of SLE and ALPS are lpr (lymphoproliferation) and gld (generalized lymphoproliferative disorder) mice. lpr and gld mice possess mutations in FAS and FASL, respectively.
The lpr mutation is associated with increased total anti-double-stranded DNA and antinuclear antibodies in sera, decreased life spans, and a severe lymph node hyperplasia noted at the end of life (1). lpr mice express very little FAS mRNA or cell surface Fas protein, and the decreased apoptosis due to low levels of Fas protein in lpr mice was found to be the cause of their lymphoproliferative disorder (66). The gld mutation is an inactivating point mutation in FASL that affects FasL activity (58), and gld mice present with a phenotype similar to that of lpr mice (10, 45).
Mutations at FAS or FASL not only are associated with ALPS and SLE but can also play a role in disease progression and outcome during pathogen infections. When cells expressing FasL interact with Fas-expressing cells, the Fas-expressing cells are caused to undergo apoptosis (56), and this is one mechanism by which T cells (23, 46) and NK cells (2) can eliminate infected cells. Previous work has demonstrated that mice that have inactivating mutations in FAS (lpr) or FASL (gld) have trouble controlling infections with West Nile virus (54), influenza virus (60), herpes simplex virus-1 (22), herpes simplex virus-2 (HSV-2) (11, 21), and mouse hepatitis virus (43). These mutations also result in worse disease and clinical score after infection with Toxoplasma gondii (17), increased parasitemia and mortality due to Tyrpanosoma cruzi subcutaneous infections (4, 33), and increased parasite load by and susceptibility to Leishmania major (18). In examining human cohorts, it was demonstrated that patients with the FAS-670 A/A polymorphism experience difficulties with human T lymphotropic virus type I (HTLV-1), where the A/A polymorphism is overrepresented in patients that eventually develop adult T-cell leukemia (13).
Here we questioned whether an activated immune system associated with a FAS mutation could, in contrast to the above studies demonstrating increased susceptibility to infection, at times enhance resistance to infection. We will show that this is the case with VACV and will discuss this along with other recent studies supporting this point (8, 24, 30, 31, 40, 41).
B6.MRL-Faslpr/J mice were used between 5 and 7 weeks of age and were bred in the vivarium of the University of Massachusetts Medical School (UMMS) Department of Animal Medicine. B6Smn.C3-Faslgld/J (B6-gld) and C57BL/6J (B6) mice were used between 5 and 7 weeks of age and were purchased from The Jackson Laboratory (Bar Harbor, ME). B6.SJL (Ly5.1+ Thy1.2+) host mice were used between 5 and 7 weeks of age and were either purchased from Taconic Farms (Germantown, NY) or bred at UMMS. B6.Cg-IgHaThy-1aGPi-1a/J (Ly5.2+ Thy1.1+) donor mice were used between 5 and 6 weeks of age and bred at UMMS. All experiments were done in compliance with the Institutional Animal Care and Use Committee of UMMS. VACV, strain WR, was propagated on L929 cells as previously described (71). Intranasal (i.n.) delivery of VACV first required anesthetization of mice by methoxyflurane (Anafane; Ivesco, Iowa Falls, IA). Anesthetized mice were i.n. infected with 104 PFU in 50 μl of Hanks' balanced salt solution (HBSS; 14025-076; Invitrogen, Carlsbad, CA). Intraperitoneal (i.p.) inoculation of mice was performed with 4 × 106 PFU diluted to 200 μl in HBSS. Plaque assays of VACV were done on Vero cell monolayers, as previously described (52).
Splenocytes were labeled with CFSE (carboxyfluorescein succinimidyl ester) or DDAO-SE [9H-(1,3-dichloro-9,9-dimethylacridin-2-one-7-yl) phosphate, diammonium salt succinimidyl ester] (C34553; Molecular Probes, Carlsbad, CA), as previously described (34, 67). Briefly, a single-cell suspension of splenocytes was prepared, red blood cells were lysed in an 0.84% NH4Cl solution, and spleen leukocytes were washed in cold HBBS and resuspended in HBSS for counting. Spleen leukocytes were next resuspended in HBSS at 2 × 107 cells per ml and labeled in a solution of 2 μM CFSE for 15 min in a 37°C water bath, with mixing every 5 min. After CFSE labeling, cells were again washed twice with cold HBSS and counted immediately before transfer. An aliquot of splenocytes was used for a surface stain, and the rest of the splenocytes were diluted in HBSS for adoptive transfer. DDAO-SE labeling of splenocytes was the same, except that the DDAO-SE final working concentration was 0.5 μM.
Single-cell suspensions of splenocytes were prepared, red blood cells were lysed in 0.84% NH4Cl, and the leukocytes were washed in RPMI 1640 medium (11875-093; Sigma-Aldrich, St. Louis, MO). Cells were then counted and resuspended in fluorescence-activated cell sorter (FACS) buffer (HBSS containing 2% fetal calf serum and 0.02% sodium azide) for staining. Fc receptors were blocked with antibody to CD16/CD32 (Fcγ III/II receptor; 553142; BD Biosciences, San Diego, CA), and cells were then stained in 96-well plates. After the surface stain with the antibodies indicated below, cells were fixed using Cytofix (554655; BD) and resuspended in FACS buffer for analysis. For intracellular stains, cells were permeabilized and fixed using Cytofix/Cytoperm (554722; BD) and stained intracellularly with the indicated antibodies as per the manufacturer's instructions. Samples were run on an LSRII (BD) flow cytometer and analyzed using FlowJo (Treestar, Ashland, OR).
For surface stains and intracellular cytokine assays, monoclonal antibodies (MAb) to CD3ε fluorescein isothiocyanate (FITC) (555274; BD), CD3ε phycoerythrin (PE)-Cy7 (552774; BD), γδ T-cell receptor (TCR) PE (553178; BD), CD4 FITC (553047; BD), CD4 PE (553653; BD), CD4 allophycocyanin (APC) (553051; BD), CD4 AF700 (557956; BD), CD4 PB (558107; BD), CD8α PE (55303; BD), CD8α APC (553035; BD), CD8α V500 (556077; BD), CD8α Alexa Fluor 700 (557959; BD), CD8α PE (12-0083083; eBioscience), CD44 PE-Cy7 (25-0441-82; eBioscience), CD44 peridinin chlorophyll protein (PerCP) Cy5.5 (45-0441-82; eBioscience), DX5 PE-Cy7 (25-5771-82; eBioscience), DX5 APC (17-5971-81; eBioscience), Thy1.1PE-Cy7, Thy1.1PE (554898; BD), Ly5.2 PerCP-Cy5.5 (552950; BD), Thy1.2 AF700 (105319; BioLegend, San Diego, CA), Thy1.2 PE (553006; BD), Thy1.2 APC (553007; BD), CD4 AF700 (557956; BD), and NK1.1 PerCP-Cy5.5 (551114; BD) were used, and MAbs to gamma interferon (IFN-γ) APC (554413; BD) and IFN-γ eFluor450 (48-7311-80; eBioscience) were used for the intracellular stains.
MAbs for depletions were diluted in HBSS and given i.p. 1 day before VACV infection. All antibodies, including isotype controls, were purchased from BioXCell (West Lebanon, NH). MAbs used for depletions were anti-CD4 clone GK1.5 (BE0003-1) and the control isotype rat IgG2b clone LTF-2 (BE0090), anti-CD8α clone 2.43 (BE00061), anti-γδ clone UC7-13D5 (BE0070), and anti-NK1.1 antibody clone PK136 (BE0036). MAb against IFN-γ was diluted in HBSS and given i.p. 2 days before and again on the day of the virus infection. Anti-IFN-γ clone XMG1.2 (BE0055) and control isotype rat IgG1 clone HRPN (BE0088) were used.
Stimulations for intracellular cytokine assays were performed as previously described (63). Briefly, single-cell suspensions of lymphocytes were cultured for 4.5 to 5 h with human recombinant interleukin-2 (IL-2; 10 U/ml) and GolgiPlug (555029; BD) and purified MAb to CD3ε (1 μg/ml) (553058; BD) was added for a polyclonal T-cell stimulation or B8R peptide (TSYKFESV, 1 μM; 21st Century Biochemicals, Marlboro, MA) as a VACV-specific stimulation.
Mice infected with VACV i.n. develop severe disease associated with immune suppression and low lymphocyte counts. We initially tested the hypothesis that activation-induced cell death (AICD) of T cells by Fas/FasL interactions limited the T-cell response to VACV, as has been shown with influenza (31). Age- and weight-matched wild-type B6 and B6-lpr mice were infected with 1 × 104 VACV i.n. (approximately 1 50% lethal dose [LD50]), and virus titers in the livers and lungs and immune responses in the mediastinal lymph nodes (MLNs) were examined. Five- to 7-week-old mice were used to avoid the lymphoproliferative disorder that occurs with the lpr mutation at about 3 months of age in mice of the C57BL/6 background (37, 38). B6-lpr mice had nearly 2 log10 less virus at day 6 in the livers (Fig. 1A) compared to B6 wild-type mice, although there was no difference in lung virus titers (Fig. 1B). There was also an increased percentage and number of CD4+ and CD8+ T cells in the MLNs of B6-lpr mice (Fig. 1C to F), with an increased number of IFN-γ+ CD8+ T cells after stimulation with the immunodominant B8R peptide epitope (Fig. 1G). B6-lpr mice also had slightly decreased morbidity after i.n. VACV infection, as in one experiment B6-lpr mice had lost 6.5% ± 2.8% of their weight by day 3, while B6 mice had lost 11.8% ± 1.6% of their weight (5 mice per group; P = 0.006); this slight decrease in morbidity continued throughout the experiment, as on day 5 B6-lpr mice had lost 19.1% ± 2.2% of their weight, compared to B6 mice, which had lost 23.1% ± 1.1% of their weight (P = 0.0007).
To examine if the increased protection in B6-lpr mice was specific to i.n. inoculations with VACV, we next examined mice infected by the i.p. route. Age- and weight-matched wild-type B6 and B6-lpr mice were infected with 4 × 106 PFU of VACV i.p., and morbidity was determined through weight loss. B6-lpr mice lost less weight after i.p. infection with VACV, indicating that morbidity was decreased in B6-lpr mice as compared to wild-type B6 mice (Fig. 2A to C). B6-lpr mice also had decreased liver virus titers on day 6 after infection compared to control B6 mice (Fig. 2D) and an increased number of splenocytes (Fig. 2E), splenic CD4+ T cells (Fig. 2F), and splenic CD8+ T cells (Fig. 2G) compared to B6 mice. The increase in CD4+ T-cell numbers was due to a significant increase in both the splenocyte number and CD4+ T-cell percentage (CD4 percentage in B6 = 10.6% ± 1.1%, in B6-lpr = 12.5% ± 0.4%; P = 0.009), while the increase in CD8+ T-cell numbers was due to the increased splenic leukocyte number (CD8 percentage in B6 = 34.3% ± 5.6%, in B6-lpr = 36.6% ± 5.1%; P = 0.52). Thus, the viral titers were low, morbidity was decreased, and the T-cell numbers were high in B6-lpr mice after infection with VACV.
It was unclear whether the high T-cell numbers suppressed virus growth or whether the high viral yields suppressed the T-cell response. We initially hypothesized that the increased T-cell numbers in lpr mice were due to resistance to AICD, which enabled protection in B6-lpr mice by keeping VACV-specific T cells alive. However, further experimentation revealed enhanced protection from VACV in FAS mutant B6 mice even at very early time points before a VACV-specific T-cell response would be expected to develop. We thus chose to focus on these earlier time points.
B6 and B6-lpr mice were inoculated with 4 × 106 VACV i.p., and virus titers were examined at day 2 or day 3 postinfection. B6-lpr mice had 0.75 log10 less virus in livers (Fig. 3A) and about 1 log10 less virus in spleens (Fig. 3B) at day 3 in comparison to B6 wild-type mice. At day 2, B6-lpr mice had 1 log10 less virus in the livers (Fig. 3C) and about 0.5 log10 less virus in the spleens compared to B6 wild-type mice (Fig. 3D). In complementary experiments with B6 and B6-gld mice, B6-gld mice had significantly less virus in the liver in comparison to B6 mice. Spleen virus titers in B6-gld mice similarly trended lower in individual experiments and reached significance when data from different experiments were combined. Analysis of the combined experiments indicated that the B6-gld mice had a statistically significant 0.5-log10 decrease in viral titers in both the liver (Fig. 3E) and the spleen (Fig. 3F) compared to wild-type mice.
Two days after i.p. infection, B6-lpr mice had increased numbers of splenocytes (Fig. 4A), CD8+ T cells (Fig. 4B), NK cells (Fig. 4C), and γδ+ T cells (Fig. 4D) compared to infected B6 wild-type mice. The increase in the number of CD8+ T cells was associated with an overall increase in total spleen leukocyte numbers, rather than a significant percentage increase in CD8+ cells (CD8 percentage in B6 = 6.34% ± 0.36%, in B6-lpr = 6.93% ± 0.46%; P = 0.3185). In contrast, increases in the number of NK cells and γδ+ T cells were due to both the increased splenocyte number and the increased percentages of these subsets within the splenocyte populations (NK percentage in B6 = 1.33% ± 0.23%, in B6-lpr = 2.03% ± 0.15%; P = 0.03; γδ+ T percentage in B6 = 0.23% ± 0.01% in B6-lpr = 0.35% ± 0.02%; P = 0.0002; 7 and 8 mice per group, respectively). When splenocytes from day 2 infected mice were cultured in vitro in the presence of Brefeldin A and IL-2 to examine spontaneous IFN-γ production, B6-lpr mice were found to have increased numbers of IFN-γ+ NK cells (Fig. 4E), IFN-γ+ CD8+ T cells (Fig. 4G), and IFN-γ+ CD3+ CD8− γδ− cells (most likely predominantly CD4+ T cells) compared to B6 mice. An observed increased number of IFN-γ+ γδ+ T cells was not statistically significant (Fig. 4F). Since there was an increase in multiple lymphocyte subsets that might control VACV infection in B6-lpr mice, B6-lpr mice were treated with either an NK cell-depleting antibody, a CD8+ T cell-depleting antibody, a CD4+ T- cell-depleting antibody, or an antibody against the γδ TCR. They were infected 1 day later, and liver and spleen virus titers were examined at day 3. No depletion tested resulted in increased virus titers in B6-lpr mice, suggesting that no single lymphocyte population was associated with the increased protection of B6-lpr mice (Table 1).
When doing these experiments, we noticed that the B6-lpr mice, whether freshly purchased from the Jackson Laboratory or bred in our facilities, weighed slightly more than wild-type B6 mice at the age of inoculation, and we wanted to make sure that the differences in viral titers seen at day 2 were not accounted for by the weight difference instead of the Fas expression difference. Our concerns over this point prompted us to both age and weight match the mice prior to experiments. The above tested mice enabled us to closely examine this issue with fairly large numbers of replicas from the various experiments. To statistically analyze whether weight matching was necessary, virus loads in the livers and spleens at day 2 postinfection were plotted against the weights of the mice before infection. The number of splenocytes at day 2 was also plotted against the virus loads in the livers and spleens at day 2 to see if any correlation was evident, as B6-lpr mice had an increased number of splenocytes as well as decreased virus loads compared to B6 mice (Fig. 2 to to44).
The initial weight did, indeed, negatively correlate with day 2 virus titers in the livers and spleens of B6 wild-type (Fig. 5A) or B6-lpr (Fig. 5B) mice, in that mice in either group that weighed more on the day of infection had a lower amount of virus in the liver and spleen 2 days later. This negative correlation between virus load and initial weight was typically strongest in the liver. These results suggested that careful weight matching, which we had employed, helped limit variability within our experiments. Because of this intentional weight matching, no significant differences were found in weights when comparing all naïve B6 to naïve B6-lpr mice used in the experiments. When the virus titers in all infected B6 wild-type and B6-lpr mice were compared, the B6-lpr mice were found to have significantly decreased virus titers in the liver (Fig. 5C) and spleen (Fig. 5D), as had been demonstrated in individual day 2 experiments (Fig. 3C and D). When initial weight was divided by virus titer to control for weight as a source of the decreased virus titer, our results confirmed that weight was not the reason for decreased titers in B6-lpr mice (Fig. 5E and F). In other correlation analyses, day 2 splenocyte numbers negatively correlated with liver and spleen virus titers in either B6 wild-type (Fig. 5G) or B6-lpr (Fig. 5H) mice. This negative correlation suggested that increased splenocyte number was indicative of better protection.
It was possible that leukocytes from B6-lpr mice were inherently more resistant to lymphocyte depletion by VACV infection because of decreased Fas-associated death, and this protection from death might preserve leukocyte viability and allow increased protection from infections. To test this hypothesis, adoptive transfer experiments at day 2 time points were examined. B6-lpr (Ly5.2+ Thy1.2+) splenocytes were labeled with DDAO-SE, and IgHa congenic wild-type (Ly5.2+ Thy1.1+) splenocytes were labeled with CFSE. These were mixed 1:1 and adoptively transferred into B6.SJL (Ly5.1+ Thy1.2+) mice. One day later, some mice were infected with VACV i.p. Similar percentages of donor DDAO+ B6-lpr (Ly5.2+ Thy1.2+) and donor CFSE+ IgHa wild-type (Ly5.2+ Thy1.1+) splenocytes were detected in uninfected host (Ly5.1+ Thy1.2+) mice, confirming the 1:1 mix of donor splenocytes (Fig. 6A). After infection with VACV, there remained equivalent numbers of DDAO+ B6-lpr (Ly5.2+ Thy1.2+) and CFSE+ IgHa wild-type (Ly5.2+ Thy1.1+) splenocytes detected in host (Ly5.1+ Thy1.2+) mice (Fig. 6A). In further analyses, B6 and B6-lpr γδ+ CD4− CD8− cells (Fig. 6B), NK+ CD4− CD8− cells (Fig. 6C), CD4+ cells (Fig. 6D), and CD8+ cells (Fig. 6E) survived infection similarly. These results suggest that a Fas-dependent cell death mechanism induced by viral infection was most likely not the reason for increased splenocyte number 2 days after infection in B6-lpr mice.
In light of these data suggesting that B6-lpr mice were better protected than B6 wild-type mice as early as day 2 after infection (Fig. 2A to C, 3C and D, and 5C and D), and because other data suggested that the NK cell percentages and γδ T-cell percentages were increased in B6-lpr mice infected with VACV as compared to B6 wild-type infected mice, uninfected B6-lpr and B6 mice were examined for differences in NK and γδ+ T-cell numbers, as these lymphocytes are thought to exert effector functions early in VACV infections (6, 28, 35, 51).
Increases in the NK1.1+ CD3− percentages (Fig. 7A), NK1.1+ CD3− numbers (Fig. 7B), γδ+ CD3+ percentages (Fig. 7C), γδ+ CD3+ numbers (Fig. 7D), and increases in the percentage of CD44+ CD4+ (Fig. 7E) and CD44+ CD8+ T cells (Fig. 7F) were found in 6-week-old uninfected B6-lpr mice as compared to uninfected B6 wild-type mice. The increased percentage of CD44+ CD4+ and CD44+ CD8+ T cells in uninfected B6-lpr mice resulted in increased numbers of both CD44+ CD4+ (B6 CD44+ CD4+ T-cell numbers = 2.3 × 106 ± 0.3 × 106, B6-lpr CD44+ CD4+ T-cell numbers = 8.1 × 106 ± 1 × 106; P = 0.0007; 5 mice per group) and CD44+ CD8+ (B6 CD44+ CD8+ T-cell numbers = 2.6 × 106 ± 0.3 × 106, B6-lpr CD44+ CD4+ T-cell numbers = 4.8 × 106 ± 0.8 × 106; P = 0.0425; 5 mice per group) T cells compared to B6 wild-type mice. There was also an increased percentage and number of CD8+ T cells that synthesized IFN-γ+ after a polyclonal anti-CD3ε stimulation in uninfected B6-lpr mice (Fig. 7G and H). The increased percentage of CD44+ CD4+ and CD44+ CD8+ T cells and increased IFN-γ produced by CD8+ T cells after a polyclonal stimulation was not surprising, as it has previously been described (5). In fact, in older (6- to 7-month-old) MRL animals with the lpr mutation, IFN-γ is produced constitutively, although this had not been noted in younger MRL-lpr mice (4 to 6 weeks old) such as the ones we used in our study (62). Although younger B6-lpr mice have not been described as having constitutive IFN-γ production, the amount of IFN-γ detected by intracellular cytokine staining after a 5-h incubation with GolgiPlug indicated that there was increased IFN-γ in splenic leukocytes in B6-lpr mice compared to B6 mice with CD3+ CD8+ cells, NK1.1+ CD3− cells, and CD3+ CD8− γδ− cells but not CD3+ γδ+ cells having significantly increased IFN-γ detected by intracellular staining (Fig. 7I). Because there was an increased number of innate-like lymphocytes in the spleens of uninfected B6-lpr mice (Fig. 7), with increased IFN-γ in leukocytes of uninfected mice (Fig. 7I), and an increased number of IFN-γ+ lymphocytes in B6-lpr mice after infection with VACV, in comparison to B6 mice (Fig. 4E to H), we hypothesized that increased numbers of cells with the capacity to make IFN-γ might result in increased protection in B6-lpr mice. This hypothesis was influenced by earlier work showing that antibody blockade of IFN-γ increases VACV titers and decreases survival after VACV infection (47), that mice that are deficient in IFN-γ have decreased survival after VACV infection (7), and that mice that have disrupted IFN-γR have decreased survival and increased viral loads after infection in comparison to wild-type controls (7, 20, 52).
In light of these studies and of our observation that there were increased numbers of IFN-γ+ NK cells, IFN-γ+ CD8+ T cells, and IFN-γ+ CD3+ CD8− γδ− T cells in infected B6-lpr mice compared to infected B6 mice, B6 and B6-lpr mice were treated with a MAb against IFN-γ and infected with VACV i.p. In three individual experiments, a trend was demonstrated that antibody against IFN-γ increased virus titers in B6 mice and more so in B6-lpr mice. The experimental results were next combined and plotted in Fig. 8. B6 or B6-lpr mice treated with an IFN-γ-blocking MAb had significantly increased viral loads in the liver and spleen compared to their genetic counterparts treated with an isotype control MAb (Fig. 8A and B). Wild-type B6 mice treated with an antibody against IFN-γ had similar amounts of virus in comparison to B6-lpr mice treated with an antibody against IFN-γ in the liver (Fig. 8A) and spleen (Fig. 8B). The fact that MAb to IFN-γ eliminated the differences in viral titers between B6 and B6-lpr mice suggests that the enhanced total IFN-γ-producing capacity of leukocytes from the B6-lpr mice accounts for the differences in the resistance of B6 and B6-lpr mice to VACV.
These experiments demonstrate that an altered immune makeup in B6-lpr mice associated with enhanced IFN-γ production leads to increased protection from VACV infections long before the onset of the full lymphoproliferative disorder typical of these mice. This enhanced resistance to infection was manifested as early as 2 days postinfection, when there was an elevation in the numbers of several different leukocyte populations previously shown to provide resistance to poxvirus infections: NK cells (6, 12, 28, 35), γδ+ T cells (51), and memory phenotype conventional αβ T cells (3, 47, 52, 70). We suggest, therefore, that the heightened state of activation occurring in FAS pathway-mutant individuals might actually provide enhanced resistance to certain viral infections.
This finding supports some very recent studies consistent with this hypothesis. In patients with chronic hepatitis B, the FAS-1377A polymorphism has been associated with a delayed onset of hepatocellular carcinoma (24). This polymorphism is in a specificity protein 1 (SP1) binding site (19), and this decrease in SP1 binding has been postulated to result in decreased Fas protein expression (55). More recently, the FAS-1377A polymorphism was associated with decreased HSV-2 seropositivity in humans (8). HSV-2, like VACV, is very sensitive to IFN-γ, perhaps explaining this phenomenon (36). This is in apparent contradiction with studies done in lpr or gld mice, in which inactivating mutations at FAS or FASL resulted in a worse outcome after HSV-2 infection (11, 21), but the lpr and gld mice are very nearly complete functional knockouts of FAS and FASL, while the FAS-1377A polymorphism is much less severe, perhaps explaining this discrepancy. Other studies demonstrated decreased incidence of HIV associated with a single-nucleotide deletion in FASLnt169 and increased recovery of CD4+ T-cell counts after antiretroviral treatment in HIV-infected individuals bearing specific combinations of polymorphisms in the FAS or FASL genes or promoter regions (40). The FAS-670A/A polymorphism has also been associated with viral clearance in patients with HCV (30). IFN-γ has been shown to directly inhibit HCV in virus replicon systems, and this mechanism could explain this association (9, 14). Alternatively, decreased expression of Fas might also limit inflammation in the liver during HCV infection, as Fas expression has previously been correlated with inflammation in liver samples of patients with chronic HCV (15). It is also possible that both mechanisms are working simultaneously.
In the mouse model, BALB/c-gld mice had better survival than their wild-type counterparts after high-dose influenza virus infections, and this was attributed to an increased CD8+ T-cell response thought to normally be restrained by FasL expression on dendritic cells (31). In other studies, MRL mice with the lpr mutation had increased survival, increased production of IL-1 and TNF-α by macrophages, increased neutrophil recruitment, and decreased bacterial outgrowth after Candida albicans infection (41). Even though these results and our current study collectively indicate that these mutations, some of which have been shown to predispose an individual to autoimmune disease, may confer a level of resistance to infectious diseases, the associated mechanisms of protection are different. Considering that these include pandemic-causing viruses like influenza and poxvirus (shown here), it is possible that these agents provided some selective pressure to maintain FAS and FASL mutations in the human population.
There has also been much work in mouse models that has demonstrated that inactivating mutations in FAS or FASL result in decreased control of pathogens (4, 11, 17, 18, 21, 22, 33, 43, 54, 60). Most of these studies suggest that the decreased control of the infecting pathogen was due to the lack of Fas-mediated cytotoxicity. With VACV, on the other hand, it has been shown that FasL was not required to resolve an intranasal VACV infection, suggesting that Fas-mediated cytotoxicity was not required to resolve VACV infection (25). This study did not describe increased protection in gld mice as our work has, but virus titers were measured in the lungs of infected animals only at days 6 and 12 (virus was cleared on day 12), and our results at day 6 mirror this finding (Fig. 1B). In fact, poxviruses encodes many inhibitors of apoptotic cell death (59), and two proteins expressed by VACV, FL1 (64, 65) and B13R (SPI-2) (29), have been shown to directly inhibit Fas-mediated cytotoxicity. Thus, Fas-mediated cytotoxicity is directly inhibited by VACV and does not appear to be involved in the clearance of VACV.
Our initial studies with B6-lpr mice evaluated i.n. and i.p. infections with VACV and found B6-lpr mice had decreased morbidity, lower viral titers, and increased numbers of CD4+ and CD8+ T cells 6 days postinfection (Fig. 1 and and2).2). We originally hypothesized that T-cell resistance to Fas-mediated AICD would have preserved T-cell viability and enhanced the clearance of virus. AICD is a phenomenon whereby continued stimulation of the TCR's results in T-cell apoptosis (50, 57, 61), and it has been demonstrated that continued stimulation of virus-specific CD8+ T cells through their TCR can result in their disappearance or apoptosis (39, 44). AICD in most cases is linked to Fas expression, as T cells from B6-lpr (32, 48) and B6-gld have defects in in vitro-induced AICD (49). Studies in the influenza model suggested that T cells under conditions of high-dose infection were eliminated by FasL expressed on dendritic cells, as gld mice resisted that process (31). In our study, the increased number of CD4+ and CD8+ T cells at day 6 after VACV infection in B6-lpr mice may have in part been due to differences in AICD, but this is clearly not the only reason. It is also likely that the day 6 T-cell numbers reflected differences in the early control of viral titers prior to the development of the T-cell response. High levels of VACV suppress T-cell responses, and reduced control of VACV in wild-type mice early in infection may have released more virus to suppress the T-cell response. B6-lpr mice i.p. infected for 2 days were indeed found to have an increased number of splenocytes, CD8+ T cells, NK cells, and γδ+ T cells compared to infected wild-type B6 mice (Fig. 4). An increase in splenocytes, as demonstrated by our results in Fig. 5G and H, negatively correlated with virus titers in both B6 and B6-lpr mice, and it was possible that the increased number of splenocytes in B6-lpr mice was because of decreased cell death after VACV infection. However, there were higher numbers of some of these cell types in mice even before infection (Fig. 7), so the numbers found shortly after infection seem mainly to reflect the differences in the starting populations in the uninfected mice. Additionally, all lymphocyte subsets examined in B6 mice adoptively reconstituted with B6 and B6-lpr splenocytes and subsequently infected with VACV survived similarly (Fig. 6), arguing that AICD was not controlling effector cell numbers at that stage of infection.
Increased protection instead seemed to be due to increased IFN-γ production, as more of the NK cells, CD8+ T cells, and CD4+ T cells in infected B6-lpr mice produced IFN-γ (Fig. 4E-H). Importantly, treatment with MAb to IFN-γ elevated VACV titers in both B6-lpr and wild-type B6 mice and took away their statistically significant differences in titers (Fig. 8). IFN-γ is very important in controlling poxvirus infections (7, 20, 27, 47, 52), and separate studies have shown that VACV can be controlled by IFN-γ-producing γδ+ T cells (51), CD8+ and CD4+ T cells (3, 47, 52, 70), and NK cells (6, 28, 35). This suggests that several different types of leukocytes could be cooperatively or redundantly producing IFN-γ in response to infection in B6-lpr mice, thus providing increased protection. In fact, selective depletion of either of these populations did not eliminate protection (Table 1), but depletion of IFN-γ did (Fig. 8). This seems in apparent contradiction with other work that demonstrated slightly decreased levels of IFN-γ in BALB/c-gld mice 16 to 23 days after infection with Typanosoma cruzi (33) or in C3H-gld mice 2 to 4 weeks after infection with Borrelia burgdorferi (53), but those experiments were performed in gld mice at time points later than those in our experiments and examined IFN-γ production only by cultured splenic CD4+ T cells.
Consistent with our results, Huang et al. (18) observed an increased amount of IFN-γ in cultured splenic supernatants of MRL-lpr mice 11 weeks after Leishmania major infection. Although the MRL-lpr mice did not have reduced parasite burden compared to controls at that late time point, the lesion size of MRL-lpr mice was significantly decreased 3 to 5 weeks after infection. Thus, the pathogen, duration of infection, and age of mice when infected all seem to be crucial in predicting whether a mutation in either FAS or FASL might lead to a better outcome after infection, and the mechanism of protection associated with a FAS or FASL mutation also may vary depending on these factors.
Our results with VACV suggest that mice prone to develop autoimmune disorders may be better protected against viruses by virtue of their activated immune system and its ability to rapidly produce antiviral cytokines months before the mice develop pathogenic signs of a lymphoproliferative disorder. This might explain recent studies where resistance to some human viruses is observed in patients with certain polymorphisms in FAS or FASL. These polymorphisms may, in some cases, actually allow enhanced protection after infection with particular pathogens.
This work was supported by United States National Institutes of Health (NIH) research grants R37-AI-017672 and U19-AI-057330 and training grant T32 AI-007349, all sponsored by the National Institute of Allergy and Infectious Diseases. The work described represents our views and not necessarily the views of the NIH.
Published ahead of print 21 March 2012