IFN-γ action in the media of the great elastic arteries, a novel immunoprivileged site (original) (raw)
Viral tropism due to anatomic specificity of immune function. We considered two hypotheses to explain the restriction of γHV68-induced arteritis to the great vessels. γHV68 might infect only SMCs of the elastic media, leaving other vascular sites uninfected. Alternatively, γHV68 could infect multiple vascular sites, but be efficiently cleared from all sites other than the elastic media. To distinguish between these hypotheses, we infected IFN-γR–/– mice with β-galactosidase (β-gal) expressing γHV68 mutants that induce chronic elastic arteritis (8, 9). At various times, the aorta with its major branches and attached heart, liver, and kidneys were dissected out as a unit and stained for β-gal activity to localize regions of viral infection (Figure 1 and 2).
Localization of γHV68 vascular infection over time. IFNγR–/– mice were infected with either M1.LacZ or γHV68 v-cyclin.LacZ. The aorta and its major branches were resected and stained en bloc for β-gal activity. Infection with wild type γHV68 resulted in no comparable staining.
Seven to fifteen days after infection, γHV68 was present in both the media and adventitia of the great vessels as shown by evaluation under the dissecting microscope (Figure 1b and Figure 2, regions a and b). Together with infection of a peripheral vessel adjacent to the kidney (Figure 1a and Figure 2, region c), this proves that γHV68 tropism is not due to selective infection of SMCs in the elastic media.
Quantitation of clearance of γHV68 from different vascular structures over time. IFN-γR–/– mice were infected and analyzed as in Figure 1 at days 7, 9, 11, 13, 15, 17, 19, 21, 42, and 43 after infection. n, number of mice. The number of aortas with blue staining in the adventitia (Adv) or media (M) is reported for the aortic base (a) or the descending aorta (b), which was representative of the entire aorta. Staining of a peripheral vessel on the posterior surface of both kidneys is also reported (c). For staining at late times compared with staining at days 7–15, A_P_ < 0.0001; B_P_ > 0.05; C_P_ = 0.04. For medial clearance compared with adventitial clearance, D_P_ < 0.0001; E_P_ = 0.006; F_P_ = 0.1.
Infection with γHV68 was efficiently cleared from both the adventitia of the great vessels and a peripheral vessel. Whereas 20 of 28 mice had bilateral staining of a peripheral vessel 7–15 days after infection (Figure 1a and Figure 2, region c), clearance was evident 17–21 days after infection and complete by 42–43 days (P < 0.0001 for late times versus days 7–15). Similarly, between 17 and 21 days after infection, adventitial staining was cleared significantly (Figure 1b compared to Figure 1, c and d, and Figure 2, regions a and b; P < 0.0001 for late times versus days 7–15). In contrast, infection of the elastic media at the base of the aorta persisted through days 42–43. In other studies, persistence of medial infection and disease of the great vessels has been seen up to 105 days after infection (1). Clearance of the media of the thoracic aorta also lagged behind clearance of either the adventitia or a peripheral vessel (Figure 1 and 2).
Clearance of adventitial infection was immune mediated because it did not occur in B cell– and T cell–deficient RAG1–/– mice (not shown). Failure of efficient immune clearance of γHV68 from the elastic media was associated with a lack of infiltration of CD4 T cells (not shown), CD8 T cells, and macrophages into the media (Figure 1, e and f, and Figure 3) despite the presence of active virus infection of medial smooth muscle cells (refs. 1 and 2), data not shown, and Figure 1f). The only inflammatory cells consistently present in the media were CD11b-positive (Mac1-positive) cells with the morphology of neutrophils (Figure 3, c and d).
Localization of inflammatory cells in arteritic lesions. Shown are low- and high-power views of immunohistochemistry on parallel aortic frozen sections from an IFN-γR–/– mouse infected with γHV68 4 weeks after infection. Dark-brown staining represents Ab binding. (a and b) CD8 T cells are present in the intima and adventitia, but not the media. Similar staining was observed with an Ab to CD4 (not shown). (c and d) Staining for CD11b shows Mac1-positive cells (macrophages and neutrophils) in all three layers of the aorta. (e and f) Staining for F4/80-positive macrophages is limited to the intima and adventitia. (g and h) Negative control with rat IgG. The area of highest background staining is shown in h for comparison with specific staining in panels b, d, and f. I, intima; M, media; Adv, adventitia.
Failure to clear the elastic media in normal mice. While it was possible that IFN-γ was required for effective immune function in the elastic media, we hypothesized that inefficient clearance of medial infection reflected a fundamental property of the elastic media. This hypothesis makes two predictions: (a) it should be possible to induce chronic elastic arteritis in wild-type mice, which have functional IFN-γ responses, with viral doses high enough to establish medial infection; and (b) transient depletion of IFN-γ in wild-type mice during acute infection should induce chronic elastic arteritis by fostering medial infection.
Consistent with our hypothesis, increasing the dose of γHV68 in wild-type mice to greater than 5 × 107 PFU γHV68 resulted in arteritis persisting 5.5 to 10 weeks (Figure 4). In contrast, IFN-γR–/– mice developed chronic arteritis when infected with 50-fold less virus (106 PFU) (1). While arteritis persisted, none of the ten wild-type animals evaluated at 6 weeks had detectable γHV68 in spleen, liver, or lung (sensitivity of plaque assay 50 PFU/organ), demonstrating that persistence occurs in the great vessels despite clearance of other organs. Furthermore, transient depletion of IFN-γ increased susceptibility of wild-type mice to chronic arteritis. Seventeen percent (5 of 29) of control mice infected with 5 × 107 PFU of γHV68 had chronic elastic arteritis 6 weeks after infection. In contrast, 67% (6 of 9) of mice transiently depleted of IFN-γ developed chronic disease (Figure 4a; P = 0.0043). At twofold to fivefold lower infectious doses, disease was seen in 50% of transiently depleted mice, but not in controls (Figure 4a; P = 0.0062).
Incidence and severity of chronic elastic arteritis in wild-type mice infected with high doses of γHV68 with or without transient depletion of IFN-γ. Mice were infected with the indicated doses of γHV68 with or without depletion of IFN-γ and evaluated for arteritis on H&E-stained sections 5.5 to 10 weeks after infection. (a) The incidence of arteritis in various groups is presented. Seven of 19 mice infected with 17 to 20 × 107 PFU γHV68 died within 7 days. Twelve of 14 mice infected with 34 to 50 × 107 PFU died within 7 days. Data for transient depletions are from two independent experiments. Data for no depletions are from 4–5 independent experiments. A_P_ = 0.0062; B_P_ = 0.0043. (b) Lesion at the aortic base of a 129Ev/Sv mouse sacrificed 10 weeks after infection with 108 PFU γHV68. Lesions in transiently IFN-γ–depleted 129Ev/Sv mice have similar histology (lesion scores for 25 mice from both groups combined = 2.0 ± 0.2; P < 0.0001 for scores compared with IFN-γR–/– mice, P = 0.001 compared with chronically depleted wild-type mice). (c) High-power view of boxed region in b. (d) Lesion at the aortic base of a chronically IFN-γ–depleted 129Ev/Sv mouse sacrificed 6 weeks after infection with 5 × 107 PFU γHV68 (lesion scores for four mice = 4.0 ± 0.7). Arteritis in IFN-γR–/– mice has similar histology (lesion scores for 16 mice = 5.0 ± 0). (e) High-power view of boxed region in d. The black lines show the boundaries of the media. Adv, adventitia; M, media; I, intima; L, lumen; V, aortic valve. ND, not determined.
We scored the severity of lesions after γHV68 infection of wild-type mice, wild-type mice transiently or chronically depleted of IFN-γ, and IFN-γR–/– mice (Figure 4, b–e). All arteritic lesions had viral antigen detectable in the media by immunohistochemistry (not shown and ref. 1 and 2). Wild-type and transiently IFN-γ–depleted wild-type mice had mild lesions with mononuclear infiltrates in the aortic intima and/or adventitia, but no medial neutrophilic infiltrates or necrosis (Figure 4, b and c). In contrast, IFN-γR–/– and chronically IFN-γ–depleted wild-type mice had severe lesions with intense medial neutrophilic infiltrates and/or medial necrosis (Figure 4, d and e).
Cellular sites of IFN-γ action that determine incidence and severity of arteritic lesions. We performed reciprocal bone marrow transfers to identify the cellular sites of IFN-γ action that determine the incidence and severity of chronic elastic arteritis. Wild-type (129Ev/Sv) and IFN-γR–/– recipient mice were lethally irradiated and reconstituted with bone marrow cells from either 129Ev/Sv or IFN-γR–/– donors (donor → recipient: 129Ev/Sv → 129Ev/Sv; IFN-γR–/– → IFN-γR–/–; 129Ev/Sv → IFN-γR–/–; IFN-γR–/– → 129Ev/Sv). After 8 weeks, reconstituted mice were infected with 107 PFU of γHV68, and arteritis was evaluated over 12 weeks. Controls demonstrated that the procedure neither protected from, nor predisposed to, arteritis. Thus, 10 of 13 of the IFN-γR–/– → IFN-γR–/– mice developed severe arteritis, whereas 0 of 11 of the 129Ev/Sv → 129Ev/Sv mice developed lesions (Figure 5, groups A and D). The IFN-γ–receptor status of the recipient determined susceptibility to arteritis since 9 of 18 of the 129Ev/Sv → IFN-γR–/– mice, but 0 of 19 of the IFN-γR–/– → 129Ev/Sv mice developed disease (Figure 5, groups B and C). Lesions in IFN-γR–/– → IFN-γR–/– mice were very severe (Figure 5, group A; average lesions score 4.7 ± 0.3), while lesions in the 129Ev/Sv → IFN-γR–/– mice were much less severe, lacking medial neutrophilic infiltrates and necrosis (Figure 5, group B; average lesion score 1.7 ± 0.2). Thus, the presence of IFN-γ receptor on hematopoietic cells determined lesion severity even when somatic cells lack the IFN-γ receptor.
IFN-γ prevents chronic elastic arteritis by effects on somatic cells but regulates the nature of the pathology by effects on hematopoietic cells. Reciprocal bone marrow reconstitutions were performed between IFN-γR–/– and 129Ev/Sv mice, and arteritis was evaluated for 12 weeks after γHV68 infection. Numbers above bars represent the number of mice within a group. Boxed numbers represent average lesion scores for a group. Data are pooled from four independent experiments. For incidence of disease, P = 0.0001 comparing group A with D, and P = 0.0004 comparing group B with C. For severity of disease, P < 0.0001 comparing groups A and B. NA, lesion scores not applicable.
IFN-γ blocks γHV68 infection and replication in SMCs and non-SMCs from the aorta. Since γHV68 replicates in vascular SMCs within arteritic lesions (1, 2) and IFN-γ prevents arteritis, we tested the hypothesis that IFN-γ has antiviral effects in primary aortic cells. SMCs and non-SMCs isolated from mouse aortas were treated with IFN-γ for 48 hours and then infected with γHV68. IFN-γ reduced virus induced cytopathic effect (Figure 6, a and b). Using dual-label immunofluorescence, we found that IFN-γ decreased the percentage of both non-SMCs and SMCs expressing viral antigen (P < 0.005; Figure 6, c, d, and g). IFN-γ treatment significantly decreased the percentage of cells with intranuclear viral capsids as determined by EM (Figure 6, e, f, and g). Thus IFN-γ has antiviral effects in both SMCs and other primary cells derived from the aorta.
IFN-γ inhibits γHV68 infection of primary aortic cells. IFN-γ–treated and untreated intimal/adventitial and medial cultures were evaluated 3 days after infection. In c and d nuclei were stained blue, muscle actin was stained red, and viral antigen was stained green. Uninfected cultures and infected cultures stained with control Ab’s demonstrated no viral antigen staining. (a and b) Phase-contrast microscopy of infected intimal/adventitial cultures with or without IFN-γ. (c) Representative field of infected medial cultures without IFN-γ treatment. (d) Representative field of infected medial cultures treated with IFN-γ. (e) EM of a cell with nuclear capsids from an infected, untreated medial culture. (f) EM of a cell without nuclear capsids from an infected, IFN-γ–treated medial culture. (g) Multiple fields were evaluated by dual immunofluorescence for viral antigen and muscle actin (three experiments) or by electron microscopy (two experiments). IF, immunofluorescence. The numbers above the bars represent the number of cells counted. For comparing results with or without IFN-γ treatment by IF, A_P_ < 0.005, B_P_ < 0.0002. For comparing results with or without IFN-γ by EM, A_P_ = 0.0024, B_P_ = 0.0013.