IL-1R Type 1–Deficient Mice Demonstrate an Impaired Host Immune Response against Cutaneous Vaccinia Virus Infection

If complexity of a cytokine network is an index of its biological importance, very few cytokines can compare with IL-1. The IL-1 family of cytokines includes three ligands (IL-1α, IL-1β, and the IL-1R antagonist [IL-1Ra]) and two forms of IL-1 receptors (IL-1R1 and IL-1R2). IL-1α and IL-1β are synthesized by cells throughout the body, including monocytes, macrophages, and neutrophils, as well as endothelial and epithelial cells. Both IL-1α and IL-1β signal through IL-1R1, whereas IL-1R2 is unable to transduce a signal and is generally considered to be a decoy receptor. IL-1Ra, which is produced in conjunction with IL-1α and IL-1β, also binds the IL-1R1 with high affinity but does not induce a signal, thus acting as an inhibitor of IL-1–mediated responses. After IL-1 binds to IL-1R1, an accessory protein (IL-1RAcP) joins with IL-1/IL-1R1 to form a signaling complex. This complex recruits intracellular adaptor molecules, including MyD88, IL-1R–associated kinase, and TNFR-associated factor 6, to activate a variety of signaling pathways, including NF-κB, AP-1, JNK, and p38 MAPK (1).

Human skin is a complex immune environment in which all elements of the IL-1 axis are represented. IL-1 is not only an important regulator of host immune defense, but it is also a potent contributor to inflammatory skin diseases, such as psoriasis and atopic dermatitis, which are associated with alterations in the balance of IL-1 proteins and receptors (2). The key role of IL-1 in host resistance to systemic viral infections such as i.v. adenovirus infection (3) and respiratory influenza virus infection (4) has been demonstrated in a variety of mouse models. However, the role of the IL-1 network in host immunity to cutaneous viral infection, particularly vaccinia virus (VACV) skin infection, remains incompletely characterized.

Similar to other poxviruses, VACV encodes proteins that interfere with important components of host antiviral defense, including soluble cytokines, chemokine binding proteins, and several cytokine and chemokine receptors (5). As IL-1 is a key initiator of the host defense against infection, linking the innate and acquired immune response systems, it is not surprising that VACV produces gene products that interfere with components of the IL-1/IL-1R1 pathway. As one example, VACV gene B15R encodes a glycoprotein that functions as a soluble IL-1R that binds IL-1β, but not IL-1α or IL-1Ra. Deletion of B15R from VACV has been shown to induce differential effects on infected mice depending on which inoculation route is used. For example, B15R-deficient VACV demonstrated increased virulence in intracranially infected mice and attenuated virulence in intranasally infected mice compared with wild-type (WT) VACV (6). However, deletion of B15R from VACV did not alter pathogenesis in an intradermal injection model of VACV infection (7). These authors observed a similar phenotype with the B15R-deficient VACV and control VACV intradermal injection and did not provide a mechanism. The outcome of applying B15R-deficient VACV to mice by skin scarification has not been reported and is worth exploring in future studies. The use of B15R-deficient VACV by skin scarification may produce different results than intradermal injection, based on previous studies from our laboratory that demonstrated significant differences in immune response between infections by intradermal injection versus skin scarification (8). Owing to the similarity of IL-1R1 to TLR, the conserved sequence in the cytosolic region of these proteins has been termed the Toll/IL-1R (TIR) domain. The proteins activated during signaling by IL-1R1 also participate in signaling by other receptors with TIR domains (9). VACV proteins A46R and A52R target multiple TIR adaptors and contribute to virulence. Immune responses to VACV lacking A46R or A52R genes were attenuated in several models of infection in mice (10–12).

Mouse models have been used to study the role of MyD88, an essential downstream adaptor involved in IL-1R1 signaling. In VACV intradermal infection, which is a more natural route than systemic routes of infection such as i.p. or i.v. injections, it was reported that MyD88 is required for induction of CD4 T cell and B cell responses and for control of virus replication at the inoculation site (13). MyD88 was also shown to be critical for CD8 T cell response after i.p. infection by VACV (14).

We investigated the roles of other IL-1 elements in VACV infection via skin scarification, the route used in humans to induce resistance to smallpox. In a previous study, we applied VACV by scarification to mice that overexpressed IL-1α in the epidermis (K14/IL-1α transgenic mice) and found that the mice demonstrated enhanced innate and adaptive immune responses (15). In the present study, we applied VACV by scarification to IL-1R1^−/− mice and found that these mice showed increased susceptibility, developing markedly larger lesions with higher viral genome copies in skin than did WT mice. IL-1R1^−/− mice also developed multiple satellite lesions in adjacent skin. This phenotype is very similar to eczema vaccinatum (EV) in humans, a severe side effect seen in atopic dermatitis patients following VACV scarification (16). Interestingly, the impaired immune response of IL-1R1^−/− mice did not reflect a systemic immune deficiency, because IL-1R1^−/− mice cleared virus from the skin and survived subsequent intranasal challenge with a lethal dose of VACV as well as immunized WT control animals. We further show that the impaired skin response of IL-1R1^−/− mice following VACV scarification was not caused by defects of T cell activation or T cell skin homing. Further study revealed that whereas VACV infection and replication after scarification were limited to the epidermal layer of WT mice, lack of IL-1R1 permitted extension of productive VACV infection to deep dermal layers of the skin in IL-1R1^−/− mice. We further demonstrate that IL-1R1^−/− mice recruited significantly more macrophages and monocyte-derived dendritic cells (mo-DCs) to the dermis after VACV scarification and that these cells were specifically vulnerable to VACV infection. We propose that the enhanced production of live virus in dermal macrophages and mo-DCs of IL-1R1^−/− mice increases the transmission of virus to adjacent skin areas, thus leading to larger skin lesions and satellite lesions seen in IL-1R1^−/− mice. This effect of impaired IL-1 signaling, leading to increased viral replication in recruited immune cells, may relate to the risk for EV seen in patients with atopic dermatitis and other inflammatory skin disorders.

IL-1R1–deficient mice and macrophage Fas-induced apoptosis (Mafia) transgenic mice were obtained from The Jackson Laboratory (Bar Harbor, ME) and bred in a biosafety level 1 facility at Harvard Medical School. Mafia mice were crossed with IL-1R1^−/− mice to produce Mafia/IL-1R1^−/− mice. C57BL/6J mice were purchased from The Jackson Laboratory and used as WT controls. All mice were handled in accordance with guidelines set out by the Center for Animal Resources and Comparative Medicine at Harvard Medical School.

Western Reserve (WR)–VACV (American Type Culture Collection, Manassas, VA) and recombinant VACV expressing GFP (rVACV-GFP, a gift from Dr. B. Moss, National Institutes of Health) were expanded in HeLa cells and titered in CV-1 cells using standard procedures (17). Modified vaccinia Ankara (MVA; ACAM3000MVA) was expanded and titered using DF-1 cells (American Type Culture Collection) (17). Mice were inoculated with 5 × 10⁶ PFU of VACV or MVA by scarification at ∼1 cm from the base of the tail as described (17). Lesions were measured with a millimeter ruler every 2–4 d following scarification. For challenge experiments, mice were anesthetized with isoflurane and inoculated intranasally with 2 × 10⁶ PFU of WR-VACV. Individual mice were weighed and checked for survival daily. Mice that lost >25% of original body weight were euthanized. Quantitative PCR was used to measure viral copy number in skin and various organs of infected mice as described (15). For analysis, inoculated skin was defined as the immediate area (0.5 cm in either direction) surrounding the inoculation site. One centimeter length of skin adjacent to the inoculated skin toward the tip of the tail was defined as adjacent skin. Measurement with a caliper was performed to guarantee the same location in all mice.

AP20187 ligand (Clontech, Palo Alto, CA) was used for in vivo macrophage depletion. It was dissolved in 4% (v/v) ethanol, 10% (v/v) polyethylene glycol 400, and 1.7% (v/v) Tween 20 and injected into Mafia mice (10 mg/kg) i.v. daily from day 0 to day 5 of VACV scarifications followed by AP20187 injections (1 mg/kg) every 3 d to maintain the depletion.

Blood was drawn from infected mice at the indicated time points, and VACV-specific Ab levels were determined as described (15).

Single-cell suspensions from whole skin were prepared as described (8). Cells from epidermal skin and dermal skin were generated as described previously with some modification (18). Briefly, the skin samples were floated, dermis side down, on a solution containing 2.4 U/ml grade II dispase (Roche, Indianapolis, IN) for 40 min at 37°C. Epidermis and dermis were gently teased away from each other and digested in HBSS skin digestion buffer, which contains 25% FCS, 1 mg/ml collagenase A, 40 μg/ml DNAse I (Roche), and 5 mM Ca²⁺ at 37°C for 30 min. The digested material was filtered through a 75-μm cell strainer to generate a single-cell suspension. Cells were rinsed twice with PBS prior to infection in media containing rVACV-GFP at multiplicity of infection (MOI) ranging from 1 to 20. The cells were incubated for 6 h at 37°C, 5% CO₂ with gentle shaking every 15 min in the first 2 h to ensure an even distribution of media in the plates. GFP⁺ VACV-infected cells were detected by flow cytometry.

Fluorochrome-conjugated anti-mouse Abs to NK1.1 (FITC), CD3 (PE), CD11b (FITC), Gr-1 (PE), F4/80 (PE), CD4 (FITC), and CD8 (PE) (all purchased from BD Biosciences, San Diego, CA) were used for cell surface staining. To detect VACV-specific CD8⁺ T cells, single-cell suspensions from inguinal lymph nodes (ILN), spleen, and skin of immunized mice were stained with PE-conjugated B8R-pentamer (ProImmune, Oxford, U.K.) as previously described (17). To detect regulatory T cells, the cells were stained with Abs to mouse Foxp3 (allophycocyanin) as well as CD4 (FITC), CD25 (PE), and CD3 (PerCP) (eBioscience, San Diego, CA) according to the manufacturer’s protocol. For intracellular IFN-γ staining, cells were first stimulated with 10 μM B8R_20–27 peptide at 37°C in 5% CO₂ in the presence of protein transport inhibitor mixture (eBioscience) for 5 h, and then stained with Abs to mouse IFN-γ (allophycocyanin) in addition to CD8 (FITC) and CD3 (PerCP) from BD Biosciences. FITC-conjugated IFN-α and IFN-β from PBL Assay Science and IFN-γ and TNF-α from BioLegend were used to detect cytokine production by mo-DCs/macrophages in inoculated skin. The single-cell suspensions made from VACV-infected skin were incubated with protein transport inhibitor mixture for 4 h prior to staining. The stained cells were acquired on a FACSCanto (BD Biosciences) flow cytometer and analyzed with FlowJo software (version 6.4.7; Tree Star, Ashland, OR). Lymphocytes were gated on CD19⁻ cells for B8R-pentamer staining.

Inoculated skin from mice scarified with rVACV-GFP (n = 5–7 per group) was harvested and digested in HBSS skin digestion buffer at 37°C for 30 min. Cells were filtered, stained with PE-conjugated anti-CD45 mAb (BD Biosciences), and sorted (BD FACSAria) into GFP⁺CD45⁺ and GFP⁻CD45⁺ subsets. Sorted cells were freeze-thawed three times, sonicated three times, and the titer of vaccinia PFU per sample was determined by standard plaque assay on CV-1 monolayers (19).

Single-cell suspensions were prepared from ILN and spleens of immunized mice and stimulated as previously described (17). Supernatants were collected at 40 h and cytokine concentrations in the supernatant were measured by ELISA (BD Biosciences).

Inoculated skin was harvested from three to four mice per group at each time point. To detect the presence of rVACV-GFP, skin was preserved in formalin, embedded in paraffin, sectioned, and stained with anti-GFP Ab (clone JL-8; Clontech) as previously described (8). To detect F4/80 or late viral Ag B5R expression, skin samples were fixed in formalin, embedded in paraffin, sectioned (10 μm), and treated with 1× target retrieval solution (DakoCytomation). Sections were incubated with purified rat anti-mouse F4/80 Ab (AbD Serotec, Raleigh, NC) or rat MAb19C2 (gifts from B. Moss, National Institutes of Health), followed by anti-rat HRP-conjugated Ab (Envision detection kit; Dako). Images were obtained with a Nikon E600 microscope and a Nikon EDX-35 digital camera.

The significance of observed differences between indicated groups was assessed by an unpaired Student t test. Tests were two-tailed with a confidence interval of 95%. A p value <0.05 was considered significant. Statistical analyses were performed with Prism software (v4.0; GraphPad Software, La Jolla, CA).

IL-1R1^−/− mice and WT mice were scarified with 5 × 10⁶ PFU of WR-VACV at the base of the tail and monitored for the development of pox lesions every 2–4 d. IL-1R1^−/− mice started to exhibit larger skin lesions than did WT mice 7 d after VACV scarification and developed multiple satellite lesions at approximately day 10. Both IL-1R1^−/− and WT mice ultimately survived VACV inoculation, and skin lesions resolved 3–4 wk after scarification (Fig. 1A, 1B, left). To investigate whether the larger skin and satellite lesions were associated with virus replication and dissemination, we measured the viral copy number in inoculated tail skin and skin adjacent to the pox lesion using PCR to quantify VACV genome copies per microgram of DNA in the skin (20–22). As illustrated in Fig. 1C, IL-1R1^−/− mice possessed 10-fold more VACV copies in inoculated skin than did WT mice at 3 d following scarification and comparable viral copies at later time points. The viral load difference between WT and IL-1R1^−/− mice was even more dramatic in skin distal to the scarification site. IL-1R1^−/− mice contained at least 100-fold more viral genome copies in skin adjacent to the site of VACV inoculation at day 3 than did WT mice and 10-fold more viral genome copies at day 6. At day 10, viral copies could not be detected in adjacent skin in 50% of WT mice whereas 100% of IL-1R1^−/− mice contained significant numbers of viral genome copies in adjacent skin. Thus, it appeared that IL-1R1^−/− mice have a defect in controlling local VACV replication and dissemination that contributed to the formation of larger and more numerous skin lesions.

Both IL-1 R1^−/− and WT mice showed negligible weight loss on day 2 postinfection and then gained weight thereafter (Fig. 1B, right), indicating that neither strain experienced severe systemic effects. We also investigated whether the impaired cutaneous response of IL-1R1^−/− mice to VACV scarification reflected a globally compromised immune response or a skin-specific immune dysfunction by measuring viral genome copies present in peripheral organs. The amount of viral DNA in ILN, spleen, lung, and liver was comparable between IL-1R1^−/− and WT mice at all time points measured up to 2 wk after scarification (Fig. 1D). Although VACV preferentially replicates and accumulates in the ovaries after i.p. infection (23), viral DNA cannot be detected in the ovaries of WT mice after skin scarification and appeared only sporadically in the ovaries of IL-1R1^−/− mice (Supplemental Fig. 1). These results indicate that the impaired host response of IL-1R1^−/− mice to VACV scarification appeared to be restricted to skin.

Ag-specific T cell proliferation and Ab production are presumed to be critical components of the immune response to primary viral infection. To understand the role of these mechanisms in the observed response to VACV scarification, we measured VACV-specific Ab and T cell cytokine production in IL-1R1^−/− and WT mice following VACV scarification. IL-1R1^−/− mice produced modestly higher titers of VACV-specific IgG and IgG2a in comparison with WT mice, indicating a Th1 response bias. The level of VACV-specific IgG1 was similar between IL-1R1–deficient and WT strains, demonstrating that there was no difference in Th2 response (Supplemental Fig. 2A). Eight weeks after VACV scarification, we challenged IL-1R1^−/− and WT mice with a lethal dose of WR-VACV by intranasal inoculation to confirm the establishment of systemic immunity. All of the immunized IL-1R1^−/− and WT mice survived intranasal viral challenge, whereas naive, unimmunized IL-1R1^−/− and WT control mice both lost body weight at the same rate and died within 1 wk after intranasal challenge (Supplemental Fig. 2B). Taken together, these results show that both IL-1R1^−/− mice and WT mice are able to achieve fully functional systemic immunity to VACV following skin scarification.

To assess the T cell immune response in scarified mice, we used B8R-pentamer staining to enumerate VACV-specific CD8⁺ T cells. We found that IL-1R1^−/− mice possessed similar numbers of VACV-specific CD8⁺ T cells in both spleen and ILN compared with WT mice at 7 d after VACV scarification (Fig. 2A). VACV-specific cytokine production was assessed in in vitro restimulation assays utilizing cells from spleen or ILN. CD8⁺ T cells from both spleen and ILN of IL-1R1^−/− mice and WT mice produced comparable amounts of IFN-γ following VACV-peptide stimulation, as measured by intracellular cytokine staining (Fig. 2B). Consistent with the results of intracellular cytokine staining, spleen and ILN cells from IL-1R1^−/− mice produced similar levels of VACV- specific IFN-γ compared with WT mice 7 d after scarification. Trace amounts of IL-4 and IL-10 were also produced by spleen and ILN cells from both strains of mice (Fig. 2C).

To address the possibility that an attenuated T cell response could be masked by greater Ag levels in IL-1R1^−/− mice due to higher virus counts in VACV-infected skin, we scarified mice with MVA, a strain of VACV that does not replicate in mammalian cells. At 7 d after MVA scarification, IL-1R1^−/− mice demonstrated a functional T cell response and produced similar amounts of IFN-γ, IL-4, and IL-10 as did WT mice (Fig. 2D). Collectively, these data indicated that the impaired cutaneous response of IL-1R1^−/− mice following VACV scarification was not caused by a defect of T cell activation or additional Ag due to virus replication.

To assess whether the impaired cutaneous response of IL-1R1^−/− mice might indicate a defect in T cell homing to skin following VACV scarification, we performed flow cytometry studies on T cells recovered from inoculated skin of VACV-scarified IL-1R1^−/− mice and WT mice 7 d after scarification. Both IL-1R1^−/− and WT mice displayed substantial numbers of CD8⁺ T cells in inoculated skin (WT, 66.7 ± 1.25%; IL-1R1^−/−, 63 ± 2.23%, p > 0.05) with a similar fraction of B8R-pentamer staining cells within the CD8⁺ T cell subpopulation of each strain (WT, 14.1 ± 1.5%; IL-1R1^−/−, 15.7 ± 1.3%, p > 0.05). Similar percentages of CD4⁺Foxp3⁺ regulatory T cells in WT and IL-1R1^−/− mice (WT, 1.09 ± 0.05%; IL-1R1^−/−, 0.93 ± 0.14%, p > 0.05) excluded the possibility that the impaired cutaneous response in IL-1R1^−/− mice was caused by enhanced recruitment of regulatory T cells to the site of infection. Furthermore, substantial IFN-γ production in response to VACV peptide stimulation was observed in CD8⁺ T cells isolated from inoculated skin (Fig. 3). These results suggested that general T cell homing and VACV-specific CD8 ⁺ T cell recruitment to skin were operational in IL-1R1^−/− mice following VACV scarification and that the functions of VACV-specific CD8⁺ T cells in the inoculated skin were not defective or suppressed.

The observation that IL-1R1^−/− mice demonstrated enhanced susceptibility to VACV cutaneous infection without evidence of defects in the adaptive immune system prompted us to examine the innate immune response in the skin following scarification. For these studies, we used rVACV-GFP for scarification, and the infection of skin cells was monitored by immunostaining for GFP. On day 3 following scarification, VACV infection was largely restricted to the epidermis of WT mice, whereas IL-1R1^−/− mice demonstrated disseminated VACV infection involving both dermal cells and epidermal cells (Fig. 4A). Because the expression of GFP in rVACV-GFP is driven by the P7.5 early-later promoter, this staining does not address whether the infection has the potential to produce progeny virus (24). To distinguish productive infection from abortive infection, we also measured the expression of the late viral protein B5R in VACV scarified skin (19). B5R⁺ cells were found in the epidermis of both WT and IL-1R1^−/− mice, but only IL-1R1^−/− mice demonstrated significant numbers of B5R⁺ cells in the dermis, suggesting productive infection of these cells (Supplemental Fig. 3). We separated epidermal skin from dermal skin at serial time points following VACV scarification and measured viral genome copies in inoculated skin and adjacent skin (Fig. 4B). The number of viral copies per microgram of DNA seen in inoculated epidermis from WT mice was consistent with that in full-thickness skin (Fig. 1C). IL-1R1^−/− mice, in contrast, showed 10-fold higher levels of viral genome copies per microgram of DNA in the epidermal skin at day 4 compared with WT mice, and slightly more at day 6. The difference of viral load in the epidermis was even more pronounced in adjacent skin, with IL-1R1^−/− mice showing 30-fold more viral copies per microgram of DNA than did WT mice at day 4 and day 6. Interestingly, VACV DNA could not be detected in the inoculated or adjacent dermis of WT mice at any time point. Although significant VACV DNA was detected in dermis at the inoculation site of IL- 1R1^−/− mice on day 4 after scarification, it was cleared by day 6 and absent in adjacent skin. In summary, VACV infection and replication were limited to the epidermal layer in WT mice, whereas deficiency of IL-1R1 permitted VACV infection with both early and late gene expression in the dermal layer and extension of viral infection to skin cells distant from the inoculation site.

Skin contains both intrinsic stromal cells (primarily keratinocytes and fibroblasts) and bone marrow–derived hematopoietic cells. To investigate which cells were susceptible to VACV infection in IL-1R1–deficient mice, we cultured primary epidermal and dermal cells from naive WT and IL-1R1^−/− mice and infected them with GFP-VACV in vitro. Infection was monitored by the expression of GFP using flow cytometry. CD45 staining was used to differentiate hematopoietic cells (CD45⁺) from skin stromal cells (CD45⁻). We found that both epidermal and dermal stromal cells and hematopoietic cells from IL-1R1^−/− mice and WT mice were equally susceptible to infection by VACV in vitro, with the number of infected cells increasing in proportion to the MOI (Fig. 5A).

We next investigated the phenotype and kinetics of cells recruited to VACV-scarified skin. We collected epidermal and dermal cells from IL-1R1^−/− and WT mice at different time points after VACV scarification and measured the percentage of NK cells, neutrophils, macrophages, and CD8 ⁺ T cells in the infected skin. VACV-scarified IL-1R1^−/− and WT mice initially had very few NK cells (CD3⁻NK1.1⁺) present in skin. The proportion of NK cells in the skin increased at day 2 after VACV infection and peaked at day 4, with most NK cells found in the dermis. There was no significant difference in the percentage of NK cells between IL-1R1^−/− and WT mice. Naive IL-1R1^−/− and WT mice had few neutrophils (CD11b⁺Gr-1⁺) in the epidermis or dermis. The proportion of neutrophils increased after VACV scarification, with 2-fold more neutrophils in IL-1R1^−/− dermis than in WT dermis at day 4. The difference in the presence of dermal macrophages (CD11b⁺F4/80⁺) between IL-1R1^−/− and WT mice was also striking. IL-1R1^−/− mice contained 2-fold more macrophages in dermis at both day 2 and day 4 as compared with WT mice. In contrast to innate effector cells, the infiltration of CD8⁺ T cells to the skin did not appear until day 4 and was more pronounced in the epidermis of both strains and was not significantly different in IL-1R1^−/− mice than in WT mice (Fig. 5B).

We examined the morphology of VACV-infected GFP⁺ cells in the dermis of IL-1R1^−/− mice 3 d after scarification and found that these cells were dendritic in character. The areas of GFP staining closely corresponded to the distribution of F4/80⁺ macrophage lineage cells. In contrast, very few F4/80⁺ dendritic macrophages were detected in the dermis of WT mice (Fig. 5C). FACS analysis was used to quantify VACV-infected cells and showed that IL-1R1^−/− mice contained ∼2-fold more infected (GFP⁺CD45⁺) leukocytes than did WT mice (WT, 6.44%; IL-1R1^−/−, 13%) at day 3 after VACV scarification. Most CD45⁺GFP⁺ cells were CD11b⁺CD207⁻ (>95%), which indicates they are not Langerhans cells. Based on CD64 (high-affinity IgG receptor FcγR1) and Ly6C expression, these cells are a mixture of mo-DCs (CD64^highLy6C^{low to high}) and macrophages (CD64^highLy6C^low) (Fig. 5D) (25). Type 1 IFNs (IFN-α and IFN-β) and Th1 cytokines such as IFN-γ and TNF-α have been found to play essential roles in controlling VACV infection (19, 26). Therefore, we measured production of these cytokines by CD11b⁺CD207⁻CD64^high cells directly isolated from VACV-inoculated skin without additional stimulation. We found that CD11b⁺CD207⁻CD64^high cells from WT mice produced more IFN-α, IFN-β, IFN-γ, and TNF-α compared with those from IL-1R1^−/− mice (Fig. 5E). The lower levels of these antiviral cytokines might make the macrophages of IL-1R1^−/− mice more susceptible to VACV infection. In fact, we demonstrated that these IL-1R1^−/− cells were infected and able to generate new virus progeny by day 5 after scarification. We sorted CD45⁺GFP⁺ cells from infected skin after VACV scarification and performed plaque assays. At day 3, similar virus levels were observed for WT and IL-1R1^−/− mice (9900 ± 458.3 PFU of virus/10⁴ GFP⁺ leukocytes [WT] versus 9400 ± 655.7 PFU of virus/10⁴ GFP⁺ leukocytes [IL-1R1^−/−]). At day 5, most infected leukocytes (≈80%) are still CD11b⁺CD207⁻CD64^high, a phenotype indicative of mo-DCs or macrophages. CD45⁺GFP⁺ cells from IL-1R1^−/− mice produced significantly more virions per cell compared with WT mice (23,400 ± 1,587 PFU/10⁴ GFP⁺ leukocytes [WT] versus 85,500 ± 8,679 PFU/10⁴ GFP⁺ leukocytes [IL-1R1^−/−]) (Fig. 5F), indicating that only IL-1R1^−/− cells were productively infected at this time. In contrast, virus was never recovered from the GFP⁻ leukocytes in the same inoculated skin samples. Taken together, these data indicate that IL-1R1^−/− mice recruited significant amounts of mo-DCs/macrophages to inoculated skin after VACV scarification and that these cells produced fewer antiviral cytokines and were more susceptible to VACV infection compared with WT mice.

Because IL-1R1^−/− mice contained significantly more macrophages in inoculated skin after VACV scarification and these cells were vulnerable to VACV infection, we propose that this response aids in the transmission of virus to surrounding skin cells, thus leading to the higher viral load in the adjacent skin. To prove this hypothesis, we crossed IL-1R1^−/− mice with Mafia transgenic mice possessing an inducible Fas suicide/apoptotic system as well as GFP driven by the mouse colony stimulating factor 1 receptor promoter. Administration of the dimerizing reagent AP20187 was shown to induce apoptosis of GFP ⁺ monocytes, macrophages, and DCs in Mafia mice. In this model, neutrophils express low levels of GFP and will not be depleted by AP20187 treatment (27, 28). In our studies, repeated injections of AP20187 depleted >90% GFP⁺CD45⁺ leukocytes in inoculated skin of both Mafia and Mafia/IL-1R1^−/− mice after VACV scarification (Fig. 6A, left). Without AP20187 treatment, Mafia/IL-1R1^−/− mice contained significantly more CD11b⁺CD207⁻CD64^high mo-DCs/macrophages in VACV-scarified skin compared with Mafia mice. After AP20187 treatment, >90% of mo-DCs/macrophages were depleted in both mouse strains (Fig. 6A, right). We harvested inoculated skin and adjacent skin 3 and 6 d after VACV scarification and measured viral genome copies. For Mafia mice, AP20187 treatment did not affect viral load at day 3, although significantly more viral genome copies were measured in inoculated skin at day 6. In contrast, AP20187-treated Mafia/IL-1R1^−/− mice possessed 10-fold fewer VACV copies in inoculated skin on day 3 and day 6 compared with mock-treated Mafia/IL-1R1^−/− mice. The difference in viral load was more dramatic in adjacent skin where AP20187-treated Mafia/IL-1R1^−/− mice contained 100-fold fewer VACV copies at both times points compared with mock-treated Mafia/IL-1R1^−/− mice (Fig. 6B). We therefore conclude that depletion of mo-DCs/macrophages can reduce viral load in Mafia/IL-1R1^−/− mice after VACV scarification and confirmed that macrophages/mo-DCs are involved in dissemination of VACV in scarified Mafia/IL-1R1^−/− mice.

IL-1 is a potent inflammatory cytokine that links innate and adaptive immune response mechanisms. In a previous study, we found that transgenic mice that overexpressed IL-1α in the epidermis (K14/IL-1α) displayed enhanced innate and adaptive immune responses to VACV inoculation by skin scarification (15). In the present study, we found that mice lacking IL-1R1 showed an impaired cutaneous immune response to VACV infection by skin scarification, resulting in an enhanced primary lesion size and the development of satellite lesions. Interestingly, IL-1R1^−/− mice were able to generate an effective VACV-specific Ab response, as well as both CD8⁺ effector and Foxp3⁺ regulatory T cells responses, following VACV scarification. Further studies revealed that the observed response in IL-1R1^−/− mice correlated with the disruption of innate immune response mechanisms. VACV infection was found to be restricted to the epidermal layer in WT mice inoculated by skin scarification, whereas inoculation of IL-1R1^−/− mice resulted in the infection of both epidermal and dermal layers of the skin and enhanced accumulation of VACV-infected dermal macrophages and monocyte-derived DCs in the dermis. These macrophage lineage cells produced live virus and were thus able to support increased VACV replication and dissemination to skin cells distant from the inoculation site. Depletion of the mo-DCs and macrophages from Mafia/IL-1R1^−/− mice significantly reduced viral genome copies in VACV-scarified skin at early time points.

IL-1R1^−/− mice have been studied in a variety of bacterial and viral infectious disease models, but the precise role of IL-1R1 in the balance between innate and adaptive immune systems remains incompletely characterized. In bacterial infection models, inhibition of the innate immune response is the predominant effect seen from disruption of the IL-1 signaling pathway (29, 30). In contrast, viral infection models have shown a more profound influence of IL-1 signaling in the adaptive immune response. For example, IL-1R1^−/− mice were more susceptible to respiratory influenza virus infection (influenza virus strain PR8) and showed both a markedly reduced IgM response and impaired recruitment of CD4⁺ T cells to the site of infection (4). IL-1R1^−/− mice infected with lymphocytic choriomeningitis virus also exhibited reduced expansion and proliferation of virus-specific CD8⁺ T cells and an inability to clear lymphocytic choriomeningitis virus compared with WT control mice (31).

Because the immune response to VACV is of high interest, due in part to its use for immunization against smallpox, a number of studies have been performed addressing the influence of innate and adaptive immune mechanisms on the containment of virus and the development of systemic immunity. In an intradermal VACV infection model, in which mice are inoculated with virus in the ear skin, the area was infiltrated first by macrophages, followed by granulocytes and then by lymphocytes (32, 33). Effective vaccination against smallpox in humans has historically been achieved by inoculation of live VACV into the skin via scarification. In an ear scarification model, VACV was shown to infect both keratinocytes and migratory monocytes, which were then eliminated by infiltrating CD8⁺ T cells (34). The positive influence of IL-1 on innate and adaptive immune responses to VACV scarification was shown in our previous study using K14/IL-1α transgenic mice that overexpress IL-1α in the epidermis (15). These mice demonstrated enhanced VACV-specific T cell proliferation and VACV-specific Ab production in response to scarification, as compared with WT mice, along with increased recruitment of T cells and DCs to the site of inoculation and more rapid clearance of virus from the skin by innate immune cells. In the present study using IL-1R1^−/− mice, we found that production of VACV-specific Ab after scarification was not altered, and that both T cell cytokine production and T cell homing to skin was unaffected in these mice. Our interpretation of these results is that the impaired cutaneous response to VACV scarification seen in IL-1R1^−/− mice reflects an alteration in the innate immune response.

Although the larger lesion phenotype seen in IL-1R1–deficient mice is similar to the effect we reported previously in TNFR1^−/− mice (17), the mechanism appears to be different. In the prior study, we showed that the skin stromal cells of TNFR1^−/− mice could not upregulate CRAMP, an antimicrobial peptide that provides critical antiviral activity, in response to VACV infection and thus could not effectively control the early stages of viral proliferation and distribution (35). Contrary to our results in TNFR1-deficient mice, IL-1R1^−/− mice show enhanced CRAMP expression relative to WT mice (data not shown).

Immunostaining of mouse skin inoculated with recombinant VACV by scarification showed that VACV infection and replication were limited to the epidermal layer in WT mice, yet extended deep into the dermal layer in IL-1R1^−/− mice. Whereas the dermis of WT mice had no detectable viral DNA at any point after VACV scarification, a significant number of viral genome copies appeared in the dermis of IL-1R1^−/− mice at day 4 after VACV scarification. The localization of VACV in the dermis of IL-1R1^−/− mice correlated with the presence and distribution of F4/80⁺ macrophage-like cells, which could be productively infected with vaccina.

Macrophages have been shown to serve different functions in two models of VACV infection. In the VACV intradermal infection mouse model, macrophages can infiltrate the site of infection and directly inhibit the spread of the virus. In this case VACV replication was completely confined to the ear skin until clodronate liposomes were used to deplete macrophages, and replicating VACV was observed to disseminate to the ovaries (28). In primary human macrophages in vitro, the cells are permissive to VACV infection and can be a vehicle for VACV replication and dissemination. These macrophages can produce extracellular enveloped virions for long-range VACV dissemination and form virion-associated structures contributing to cell–cell spread (36). Although the coexpression of CD11b and F4/80 has often been used to identify macrophage-lineage cells, macrophages cannot be accurately distinguished from conventional DCs (CD11b⁺ DCs) and mo-DCs in skin. A combination of cell-surface markers has recently been described to differentiate these cell populations in healthy and inflamed skin (25). After excluding Langerhans cells that are CD11b⁺CD207⁺ by gating, the remaining CD11b⁺CD207⁻ cells can be further subdivided into CD11b⁺ DCs (CD64⁻MerTK⁻CCR2⁺Ly6C⁻), monocytes (CD64^lowMerTK^{− to low}CCR2⁺Ly6C^high), mo-DCs (CD64^{− to +}MerTK^{− to low}CCR2⁺Ly6C^{low to high}), and macrophages (CD64^highMerTK⁺CCR2^lowLy6C^low) . Using this expanded panel of markers to characterize the macrophage-like cells in VACV-infected skin, we found that the infected leukocytes (CD45⁺GFP⁺) cells were almost exclusively CD11b⁺CD207⁻ 3 d after VACV scarification in both WT and IL-1R1^−/− mice. The pattern of CD64 and Ly6C expression suggested it is a mixed population of both mo-DCs and macrophages. In IL-1R1^−/− mice, the enhanced infiltration of mo-DCs/macrophages and their susceptibility to infection may lead to increased dissemination of VACV to adjacent skin and cause the larger skin lesions and satellite lesions seen in IL-1R1^−/− mice, but not WT mice. Depletion of mo-DCs/macrophages with AP20187 in Mafia/IL-1R1^−/− mice dramatically reduced viral genome copies in VACV-scarified skin at the early stage compared with untreated control Mafia/IL-1R1^−/− mice. Multiple satellite lesions were found on the control Mafia/IL-1R1^−/− mice 10 d after VACV scarification whereas very few AP20187-treated Mafia/IL-1R1^−/− mice developed satellite lesions sporadically (data not shown). These data indicate that mo-DCs and macrophages play important roles in cutaneous VACV infection. The observation that T cell responses are unaffected in IL-1R1^−/− mice allows for eventual effective control of viral replication and the development of protective immunity. Infection of skin leukocytes after VACV scarification has been reported previously, but the phenotype of infected leukocytes was somewhat different from that in the present study. The previous study indicated that most VACV-infected leukocytes were inflammatory monocytes, along with minor populations of DCs, Ly6G⁺ neutrophils, and Ly6G⁺monocytes (34). The discrepancy between these two studies may reflect the different location of skin scarification (tail versus ear) and different combinations of surface markers used for flow cytometry.

As both IL-1α and IL-1β use IL-1R1 to mediate their biological activities, the impaired cutaneous response to VACV in IL-1R1^−/− mice may reflect a requirement for either IL-1α or IL-1β, or both, for optimal anti-VACV immune response. Production of IL-1α and IL-1β is largely segregated in the different anatomic layers in skin. IL-1α is constitutively expressed by epithelial cells, whereas IL-1β is more abundant in hematopoietic cells (1, 37, 38). To investigate the role of IL-1α versus IL-1β, we performed preliminary studies in which mice deficient in IL-1α or IL-1β were inoculated with VACV by scarification. In these studies we found that both IL-1α^−/− and IL-1β^−/− mice developed larger skin lesions and higher virus counts in adjacent skin compared with WT controls (Supplemental Fig. 4). The skin lesions observed in IL-1α^−/− mice were more dramatic than those of IL-1β^−/− mice, and IL-1α^−/− mice exhibited more satellite lesions. This phenotypic difference suggests that these two IL-1R1 ligands may have distinct roles in the regulation of the skin immune response to VACV scarification. Further studies are needed to clarify the influence of IL-1α and IL-1β in restricting VACV replication in mo-DCs/macrophages in vivo.

In conclusion, elements of the IL-1 network, including IL-1α, IL-1β, and IL-1R1, are essential for the optimal host response to VACV cutaneous infection. Our study may lead to new therapeutic strategies for treatment of EV and contribute to risk assessment for viral infections in patients receiving IL-1 blocking Abs for treatment of chronic inflammatory disorders.

We thank Dr. Niro Anandasabapathy for helpful discussions and for critically reviewing the manuscript.

The authors have no financial conflicts of interest.