Interleukin-36β provides protection against HSV-1 infection, but does not modulate initiation of adaptive immune responses

Interleukin-36 (IL-36) represents three cytokines, IL-36α, IL-36β and IL-36γ, which bind to the same receptor, IL-1RL2; however, their physiological function(s) remain poorly understood. Here, the role of IL-36 in immunity against HSV-1 was examined using the flank skin infection mouse model. Expression analyses revealed increased levels of IL-36α and IL-36β mRNA in infected skin, while constitutive IL-36γ levels remained largely unchanged. In human keratinocytes, IL-36α mRNA was induced by HSV-1, while IL-1β and TNFα increased all three IL-36 mRNAs. The dominant alternative splice variant of human IL-36β mRNA was isoform 2, which is the ortholog of the known mouse IL-36β mRNA. Mice deficient in IL-36β, but not IL-36α or IL-36γ, succumbed more frequently to HSV-1 infection than wild type mice. Furthermore, IL-36β−/− mice developed larger zosteriform skin lesions along infected neurons. Levels of HSV-1 specific antibodies, CD8+ cells and IFNγ-producing CD4+ cells were statistically equal in wild type and IL-36β−/− mice, suggesting similar initiation of adaptive immunity in the two strains. This correlated with the time at which HSV-1 genome and mRNA levels in primary skin lesions started to decline in both wild type and IL-36β−/− mice. Our data indicate that IL-36β has previously unrecognized functions protective against HSV-1 infection.

. Protein sequence alignment of human and mouse IL-36 cytokines. (a) Human (h) and mouse (m) IL-36 cytokines were aligned using Clustal omega. *conserved residues: Gonnet PAM 250 matrix score > 0.5, 0 < Gonnet PAM 250 matrix score ≤ 0.5. Amino acid sequence of human IL-36β isoform 1, which diverge from the other family members, is shown in red. (b) Neighbor-joining phylogenetic tree showing relationships among IL-36 cytokines. The optimal tree with the sum of branch length = 2.69 is shown. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) are shown next to the branches. The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the Poisson correction method and are in the units of the number of amino acid substitutions per site. (c) Heat map showing percentage similarity between the human and mouse IL-36 cytokines. activation 29 . Subsequently, in vitro cleavage by neutrophil proteases has been demonstrated 30,31 ; however, in vivo processing remains to be documented.
The physiological function(s) of the IL-36 cytokines remain poorly understood. Due to increased IL-36α and IL-36γ skin expression and presence of loss-of-function mutations in the natural antagonist, IL-36R antagonist (IL-36Ra), the IL-36 cytokines are believed to play a pathogenic role in plaque and generalized pustular psoriasis 16,25,32 . One such mechanism may involve neutrophil recruitment 25,[32][33][34][35] . Furthermore, in vitro studies suggest that the IL-36 cytokines can stimulate maturation of dendritic cells and downstream polarization of naïve T cells towards IFNγ producing Th1 cells [36][37][38][39] . This could suggest a role in immune responses directed against microorganisms; yet, documentation of such a function has remained elusive 36,40 . Another intriguing aspect of the IL-36 cytokines is the presence of three genes encoding proteins acting on the same receptor as described above.
With the long-term goals of identifying normal physiological functions and the purpose of maintaining the IL-36 gene duplication during evolution, we have started comparing outcomes of challenging IL-36α −/− , IL-36β −/− and IL-36γ −/− mice. As part of these studies, we recently reported that inflammation induced by the antiviral drug imiquimod requires IL-36α, but neither IL-36β nor IL-36γ 41 . Interestingly, imiquimod is sometimes used to treat active HSV skin disease caused by strains resistant to acyclovir [42][43][44][45] . Extending upon our previous studies, we here examined the role of the individual IL-36 cytokines during HSV-1 infection using the flank skin mouse model as our experimental system. Unexpectedly, we found that mice deficient in IL-36β, but not IL-36α, developed more severe disease; however, IL-36β was not essential for initiation of adaptive immunity against HSV-1. Hence, IL-36β appears to have previously unrecognized functions that protect against the outcome of HSV-1 infection.

IL-36β deficient mice have increased mortality following HSV-1 infection.
Using IL-36α knockout (KO) mice, we previously demonstrated that IL-36α is critical for the skin inflammation induced by the antiviral drug imiquimod 41 . In contrast, IL-36β −/− and IL-36γ −/− mice developed the same degree of skin inflammation as wild type mice 41 . Indirectly, this could suggest a role for one or more IL-36 cytokines in immune responses against viruses. Interestingly, when the IL-36s were discovered it was demonstrated that expression of one, likely IL-36α, was up-regulated in keratinocytes infected with HSV-1 14 ; however, the function was not explored. Thus, we examined progression of HSV-1 infection in wild type, IL-36α −/− , IL-36β −/− , and IL-36γ −/− mice by employing the flank skin HSV mouse model. We have previously reported that the IL-36 knockout (KO) mice have no spontaneous phenotypes precluding this study 41 . In the flank model, HSV-1 enters sensory neurons at the site of primary infection (Fig. 2a, yellow circle). From here it migrates through the neurons to the dorsal root ganglia (Fig. 2a, red arrows), where it can establish latency. Subsequently, viral replication, anterograde migration (Fig. 2a, blue arrows) and shedding leads to formation of secondary zosteriform lesions along the affected neurons (Fig. 2b, days 4-7, note the linear lesion pattern forming 5 days post-infection) 1 . Using this model, we found that IL-36α −/− and IL-36γ −/− mice exhibited the same mortality rate (Fig. 2c, median survival time, >16 days) as wild type mice. In contrast, the IL-36β −/− mice revealed significantly reduced survival (Fig. 2c, median survival time: 11 days). Fatal outcome in both wild type and IL-36β −/− mice was associated with progressive weight loss (Fig. 2d). The increased mortality rate of the IL-36β KO mice strongly suggest an important involvement of IL-36β in controlling the outcome of HSV-1 infection.
Lethal outcome is associated with bowel dysfunction syndrome in both wild type and IL-36β KO mice. It has been reported that mice infected vaginally with HSV-2 (strain 333) develop paralysis, constipation and bladder retention 46 . These outcomes could be the result of inflammation causing demyelination in the nervous system 47 . Using the flank model and HSV-1 (strain NS), we previously observed a bowel dysfunction syndrome evident by greatly enlarged colons and small intestines in moribund wild type and IL-1R1 KO mice 12 . Recently, the lethal outcome of vaginal HSV-1 infection in mice was linked to these phenotypes, as the virus spreads through the nervous system resulting in inflammation-mediated damage to the enteric neurons and toxic megacolon (constipation) 48 . Here we found that moribund wild type and IL-36β KO mice exhibited greatly enlarged small and large intestines (Fig. 3a). This bowel dysfunction syndrome likely explains the lethal outcome in the flank model. Viral DNA is present in brain, liver and lungs in moribund wild type and IL-36β KO mice. In humans HSV can disseminate to internal organs such as the brain, liver and lungs, where it causes organ damage leading to significant morbidity and mortality [2][3][4][5][6] . We previously showed that in the HSV-1 flank model the virus disseminates to internal organs such as the brain, liver and lungs 12 . Furthermore, we reported that IL-1R1 KO mice exhibit an increased mortality following HSV-1 infection 12 similar to that reported here for IL-36β (Fig. 2c); yet levels of viral genomic DNA in internal organs from moribund IL-1R1 deficient mice did not differ from those of wild type moribund mice 12 . Upon analyses of organs from moribund IL-36β KO mice, we found that HSV-1 genome copy numbers were similar to those found in moribund wild type mice (Fig. 3b). This could suggest that HSV-1 disseminates to the same organs in wild type and IL-36β KO mice.
IL-36β KO mice develop antibodies against HSV-1 at the same time as wild type mice. The adaptive immune system is essential for immunity against HSV 1,7,8 . In vitro studies have demonstrated that IL-36 up-regulates expression of MHC class II and CD83 on dendritic cells 37,39,49 ; thus, IL-36 may promote the development of adaptive immune responses in vivo. In our HSV-1 flank model, mice deficient in B and T cells (RAG1 KO mice) had a median survival time of 9 days (Fig. 2c). Given the timing of death observed with the RAG1 and IL-36β (Fig. 2c) KO mice, we hypothesized that IL-36β protects against HSV-1 by promoting the development of adaptive immunity towards the virus. Initial analyses of leukocyte populations in spleens and draining inguinal lymph nodes revealed no significant differences in the proportions of granulocytes, antigen presenting cells and lymphocytes in wild type and IL-36β KO mice in the presence or absence of infection ( Fig. 4a,b).
In mice, antibodies can protect against lethal outcome of HSV infection 50 , including that caused by the NS strain used here 51,52 . Hence, we examined the progressive development of HSV-1 antibodies in wild type and IL-36β KO mice. HSV-1 gD has been reported to be the major antigen towards which early antibodies are directed during a primary HSV infection (see ref. 51 for refs.). In agreement with this, we observed potent reactivity towards an approximately 43 kDa protein band likely representing HSV-1 gD (Fig. 4c). Similar banding patterns were observed in wild type and IL-36β KO mice (Fig. 4c). Quantitative analyses (Fig. 4d) revealed appearance of increased antibody levels already at day 5 in both wild type and IL-36β KO mice (Fig. 4d). These levels increased up to day 9 when RAG1 −/− died; however, no statistically significant differences between wild type and IL-36β KO mice were detected (Fig. 4d). This outcome demonstrates that IL-36β plays no role in initiating a primary antibody response in the present model system.
Wild type and IL-36β KO mice develop gB(498-505) specific CD8 + cells at the same time. CD8 + cells can kill HSV-1 infected cells 8 , and start clearing the primary HSV-1 infection in the mouse flank skin model as early as day 5 post-infection 53 . Since IL-36β deficient mice developed antibodies at the same time as wild type mice, we next examined their ability to develop HSV-1 specific CD8 + cells. Overall levels of the major T cell populations were indistinguishable in wild type and IL-36β KO spleens and draining inguinal lymph nodes in both uninfected and infected mice (Fig. 5a). In mice, the majority of HSV-1 specific CD8 + cells are directed against the epitope gB(498-505) comprising the sequence SSIEFARL 54 . To address if the initiation of the development of these cells requires IL-36β, we examined levels of HSV-1 gB(498-505) specific cells 6 days post-infection. Since this is a very early time point, numbers of these cells were expected to be low. In both wild type and IL-36β KO mice, the percentage of CD8 + cells specific to gB(498-505) increased significantly in both the draining inguinal lymph nodes and the spleens post-infection (Fig. 5b,c); however, no significant differences between wild type Scientific RepoRts | 7: 5799 | DOI:10.1038/s41598-017-05363-4 and IL-36β KO mice were detected (Fig. 5b,c). This suggests that IL-36β does not regulate processes required for initiation of CD8 mediated cellular immunity.  6) in primary infections sites were found to be not statistically significantly different in the wild type and IL-36β KO mice on days 3 and 5. This could suggest that HSV-1 replicates in the skin at similar rates in both strains. Furthermore, the viral genome copy numbers started to decrease at the same time, i.e., leading to a significant reduction in HSV-1 genome copy numbers from day 5 to day 7 (Fig. 6). The timing of this decrease correlated with induction of early HSV-1 specific antibodies (Fig. 4). This provides further evidence that adaptive immune responses aimed at clearing the skin infection are equally well initiated in the wild type and IL-36β KO mice.

HSV-1 replication appears to progress similarly at the site of primary infection in wild type and IL-36β
IL-36β deficient mice develop more severe secondary zosteriform skin lesions. An interesting aspect of the flank model of HSV-1 skin infection is the re-dissemination of the virus from the dorsal root ganglion, through the sensory neurons, to the skin, where it causes the formation of secondary zosteriform lesions along the dermatome several days after primary infection (Fig. 2a,b). In IL-1R1 KO mice these secondary lesions appear at the same time as in wild type mice 12 . Furthermore, the lesions progress in size in a similar manner 12 . Interestingly, in the present study we found that these zosteriform lesions emerged approximately at the same time in wild type and IL-36β KO mice (Fig. 7, day 5) suggesting that the virus migrates through the neurons at the same rate. However, the secondary lesions progressed to become significantly larger in IL-36β KO mice than in wild type (Fig. 7, days ≥ 6). This may implicate IL-36β in mechanisms aimed at limiting wound progression as the virus reemerges from the neurons.

Levels of IFNγ-producing CD4 + cells are similar in wild type and IL-36β KO zosteriform
lesions. IFNγ-producing CD4 + cells provide protective immunity against HSV infections in peripheral epithelial tissues 55,56 . Since IL-36 has been shown to promote differentiation of naïve T cells into IFNγ-producing Th1 cells 37, 57 , we hypothesized that the larger zosteriform lesions could be due to reduced levels of these cells in the skin. Significantly increased levels of IFNγ-producing CD4 + cells were found in skin surrounding early secondary lesions compared to uninfected and primary infection sites (Fig. 8). However, no statistically significant differences in IFNγ-producing CD4 + cell levels were found when comparing wild type to IL-36β KO mice (Fig. 8). This suggests that IL-36β KO mice have an uncompromised capacity to activate and/or recruit IFNγ-producing CD4 + cells.

IL-36 expression is induced during in vivo infection in mice.
To possibly explain the specific role of IL-36β, but not IL-36α and IL-36γ, in protection against HSV-1 infection (Fig. 2c), we examined in vivo IL-36 Organs were collected from moribund mice, homogenized and HSV-1 genome copy numbers determined by QPCR. Each symbol represents a single mouse. Data is pooled from two independent experiments. No statistical significant differences were observed between wild type and IL-36β deficient mice. expression in the skin during HSV-1 infection (Fig. 9a-c). Expression of the IL-36α and IL-36β mRNAs was dramatically increased (approximately 20-40-fold) in primary lesions starting at day 3, and levels remained elevated through day 7 (Fig. 9a,b). In contrast, the IL-36γ was only modestly induced (2-3-fold) by day 7 (Fig. 9c). Induction of the IL-36 mRNAs appeared delayed compared to viral replication as measured through levels of No statistically significant differences were observed between wild type and IL-36β deficient mice (#) despite significant increases in anti-gD titers in both strains (**p < 0.01 (compared as indicated); ***p < 0.001). Each symbol represents a single mouse. HSV-1 gD mRNA (Fig. 9d). Furthermore, the IL-36 mRNA levels remained elevated by day 7 (Fig. 9a-c), as viral levels decreased (Fig. 9d). The HSV-1 gD mRNA levels, including the observed decrease in the HSV-1 gD mRNA levels at day 7, correlate with the observed HSV-1 DNA genome copy numbers (Fig. 6). This could suggest that the induction of IL-36 expression is secondary to the infection, i.e., not a direct response to HSV-1.
Next the relative expression of the three IL-36 mRNAs was examined (Fig. 9e). Analyses in uninfected skin revealed that the IL-36γ mRNA is constitutively expressed at approximately 10-fold higher levels than the IL-36α and IL-36β mRNAs (Fig. 9e). However, due to the strong induction of the IL-36α and IL-36β mRNAs (Fig. 9a,b), a trend towards the IL-36β mRNA being the dominant form at day 3 may be detected.

IL-36α, but not IL-36β and IL-36γ, expression is induced in human keratinocytes during in vitro
HSV-1 infection. Previously, we showed that human primary keratinocytes have high levels of pre-formed IL-1α that are released upon HSV-1 infection 12 . In contrast, unstimulated keratinocytes express only very low levels of IL-36β mRNA; however, the expression is increased approximately 10-fold upon poly(I:C) treatment 22 .  Here we found that HSV-1 infection induced expression of IL-36α (Fig. 10a). However, we observed no significant increase in the IL-36β and IL-36γ mRNAs (Fig. 10b,c).

Cytokines induce expression of all three IL-36 mRNAs. The delayed in vivo induction of the IL-36
mRNAs could suggest secondary responses. Expression of the IL-36 mRNAs can be induced by cytokines such as IL-1, IFNγ (IL-36β only) and TNFα 21,28 . We previously showed that these cytokines can induce chemokine expression in keratinocytes in additive and synergistic manners 58 . Here, we found that IL-1β and TNFα significantly induced all three IL-36 mRNAs (Fig. 10d); sometimes with synergistic effects (IL-36α mRNA 24 hours, Fig. 10d). In some (not shown), but not all (Fig. 10d), experiments we observed increased expression of the IL-36β mRNA in response to IFNγ. Overall, the IL-36α and IL-36γ mRNAs appeared the most dramatically induced with IL-36γ exhibiting the highest levels of expression (Fig. 10d).
IL-36β isoform 2 is the major IL-36β isoform expressed in human keratinocytes. Interestingly, two alternative splice variants of human IL-36β mRNA have been sequenced (Fig. 1). Isoform 2 represents the ortholog of the known mouse IL-36β mRNA (Fig. 1), whereas isoform 1 is substantially distinct from the other IL-36 cytokines (Fig. 1). Isoform specific analyses of human keratinocyte mRNA revealed that the IL-36β isoform 2 mRNA was expressed at dramatically higher levels (>50-fold at all time-points and treatments) than IL-36β isoform 1 (Fig. 10d). This suggests that most of the IL-36β expressed by keratinocytes will have activity similar to those of IL-36α and IL-36γ.

Discussion
A role for the IL-36s in immunity against microorganisms has been suggested by their induction in epithelial cells in response to several pathogen associated molecular patterns in vitro 21,22 as well as HSV-1, rhinovirus and influenza in vivo 14,19,20 . However, the physiological functions of the IL-36s during infections remain poorly understood. Using KO mice for each individual IL-36 cytokine, we found that IL-36β, but not IL-36α or IL-36γ, is critically involved in protecting against the outcome of HSV-1 skin infection. We specifically observed that IL-36β KO mice develop larger skin lesions (Fig. 7) and have decreased survival (Fig. 2c). This suggests an essential role for IL-36β in controlling the outcome of the viral infection.
The IL-36s can activate dendritic cells in vitro [36][37][38][39]57 and in vivo in an experimental model of psoriasis, a chronic inflammatory skin condition 59 . This could suggest that the IL-36s play a role in activating the adaptive immune response. Interestingly, we find that IL-36β KO mice start to clear the viral skin infection as quickly as the wild type mice (Fig. 6, day 7; Fig. 9d, day 7). Furthermore, we were unable to detect differences in levels of antibodies (Fig. 4c,d), gB(498-505) specific CD8 + (Fig. 5b,c) and IFNγ-producing CD4 + cells (Fig. 8) in wild type and IL-36β KO mice. This suggests that IL-36β does not play a significant role in initiating adaptive immunity against the examined immunogens in the presently used model. Since phenotypic differences are observed between wild type and IL-36β KO mice, IL-36β must have other essential physiological functions during HSV-1 infection. Although IL-36β did not contribute to the development of adaptive immunity in the present studies, the IL-36 cytokines may still be involved in these responses. Our initial working hypothesis was that the IL-36 cytokines could overcome HSV-1 immune evasion that inhibits the function of IL-1β 10,11 and presumably IL-18 by similarity. IL-1α and IL-33 are related cytokines that are highly expressed in keratinocytes and play a significant role in immunity against HSV infection 12,60 . The presence of these cytokines could be enough to initiate adaptive immunity in the present HSV infection model. Further studies involving mice unable to signal via all these cytokines will be required to test this possibility. In addition, IL-36β may play a role in recruitment of immune cells (discussed below) and maturation of the immune response. We here examined early immunity against the dominant antibody gD (Fig. 4d) and CD8 + gB (Fig. 5b,c) immunogens. However, over time the  (Fig. 2b) were excised and pooled. Number of IFNγ-producing CD4 + cell were determined using ELISpot assays. Representative data (means ± SD) from one of two experiments are shown. # p > 0.05; *p < 0.05; **p < 0.01.
immune response develops to enhance epitope affinity and diversity. Future analyses of such responses may reveal if IL-36β regulates these processes. Similar analyses of IL-36α and IL-36γ functions should also be performed in conjunction with quantification of IL-36 protein levels and proteolytical activation.
The specific role of IL-36β, but not IL-36α and IL-36γ, in protection against lethal outcome of the HSV-1 infection (Fig. 2c) is intriguing. Using a vaginal model of HSV-1 infection, the lethal outcome of the disease in mice has recently been linked to dissemination of the virus to the enteric nervous system, resulting in inflammation mediated nerve damage, constipation and toxic megacolons 48 . In the HSV-1 flank skin infection model used here, moribund mice exhibit a gastro-intestinal phenotype (Fig. 3a) suggesting a similar loss of peristalsis leading to death. Furthermore, HSV-1 DNA can be detected in several vital organs such as the brain, liver and lungs (Fig. 3b). Interestingly, neurons can express IL-36β, and astrocytes and microglia have the receptor for IL-36, IL-1RL2 26,27 . Hence, it is possible that IL-36β protects against lethal outcome of HSV-1 infection (Fig. 2c) by modulating the ability of the virus to spread and/or replicate in the nervous system. The cellular source(s) of IL-36 and induction pathways in the skin, and possibly the nervous system, remain to be determined. It is conceivable . At the indicated time-points primary or mock lesions were collected at the center using 4 mm punch biopsies. IL-36 (a-c) and HSV-1 gD (d) mRNA levels determined by real-time PCR using GAPDH as the housekeeping gene against which individual mRNA levels were standardized. Fold changes in IL-36 mRNA levels were calculated against uninfected skin. Fold changes in HSV-1 gD mRNA levels were calculated against levels at day 1. Data shown is pooled from three independent experiments. (e) IL-36α (open bars), IL-36β (blue bars) and IL-36γ (red bars) mRNA levels were recalculated (data from a-c) relative to levels of the IL-36α mRNA in uninfected mice. *p < 0.05; **p < 0.01; ***p < 0.001; ND, not detected. that the IL-36s act in local microenvironments, and thereby regulate very different physiological responses to infections. Further studies will be required to establish these mechanisms and functions.
The IL-36s promote neutrophil recruitment in different types of psoriasis 25,41 . Neutrophils play an important role in guiding CD8 + cells to sites of influenza infection in the lungs 61 , and HSV-1 skin lesions are well known to be rich in neutrophils (see Wojtasiak et al. 62 for refs). However, intradermal administration of IL-36α does not promote recruitment of CD8 + cells 39 , and removal of neutrophils does not affect the outcomes of HSV-1 skin infection 62 . Hence, neutrophils may have distinct anti-viral functions in the lungs that are not observed in the skin. While this manuscript was being prepared, it was reported that mice deficient in the receptor for IL-36, IL-1RL2 (also known as IL-36R), are protected from the lethal outcome of lung influenza virus infection 20 . The reduced mortality was associated with decreased lung inflammation, e.g., neutrophil infiltration, and tissue damage 20 . Furthermore, the virus induced inflammation was IL-36α dependent 20 . The latter is in agreement with our studies in the imiquimod skin inflammation model, where neutrophils are recruited in response to increased IL-36α expression 41 . However, the former is the opposite outcome, compared to the increased mortality seen here in the HSV-1 infection model in the absence of IL-36β signaling (Fig. 2c). The distinct phenotype in the HSV model used here may be linked to a unique role of IL-36β in the nervous system as discussed above. However, IL-36β also appears to have a distinct role in the skin. If the observed skin lesions (Figs. 2b and 7) were caused by inflammation induced by IL-36β, we would expect IL-36β KO mice to have smaller lesions than wild type; however, larger lesions are in fact observed (Fig. 7). In vitro studies with keratinocytes may reveal if IL-36β can modulate innate immunity in these cells. Additional studies of other viruses may also determine if these are functions specific to HSV or neurotropic pathogens, or if they represent universal protective mechanisms.
IL-36γ was shown recently to promote wound healing in the gut in experimental models of inflammatory bowel disease 63,64 . IL-36 could have a similar function in the skin during HSV-1 infection. Such a wound healing property could explain, at least in part, why lesions are larger in IL-36β KO mice than wild type (Fig. 7), and is supported by the extended expression of IL-36 at day 7 post-infection (Fig. 9a-c), when levels of virus are already declining (Figs 6 and 9d). It is also conceivable that IL-36 expression (Fig. 9a-c) is induced by wound healing responses and not the infection per se. This could explain the paradox that the IL-36β mRNA is induced in vivo (Fig. 9b), but not during an in vitro infection (Fig. 10b). However, since IL-36β deficient lesions progress to become larger (Fig. 7), not merely heal poorly, IL-36β must have additional activity aimed at restricting the virus or lesion progression. This could involve an innate immune mechanism that restricts the ability of HSV-1 to spread laterally within the proliferating layer of keratinocytes in the skin. Further studies of how IL-36β controls HSV-1 infection, and IL-36 promote wound healing will be needed to determine the specific mechanism(s) involved. In summary, IL-36β clearly plays a critical role in controlling the outcome of HSV-1 infection (Figs. 2 and 7); however, further studies will be necessary to define the mechanisms whereby IL-36β acts, and how it is activated and released from cells.
In vivo HSV-1 infections. Male and female mice were bred in house in a pathogen-free animal facility and used for experiments. Mice were matched for age (7-12 weeks) and sex in each individual experiment. The infection protocol was obtained from Dr. Friedman 52 . Mice were denuded the day before infections by sequential shaving and epilating cream application. Cream was removed by rinsing with water. Scratch inoculations were performed with 1.5 × 10 6 PFU HSV-1 on the right flank. Mice were photographed next to a ruler every second day. Image pixels were converted to mm using Image J (https//imagej.net/Welcome) and the depicted ruler. Skin lesions were outlined and sized in Image J. Blood was collected by cardiac puncture immediately following euthanasia and serum stored at −20 °C.
Cell culture and treatments. Pooled human neonatal primary keratinocytes (Thermo Fisher Scientific, Carlsbad, CA) were maintained in Defined Keratinocyte Serum Free Medium supplemented with 50 μg/mL gentamicin (Invitrogen). HaCaT cells (provided by Dr. Meenhard Herlyn, Wistar Institute, Philadelphia, PA) were maintained in Dulbecco's modified Eagle's medium (Thermo Fisher Scientific) supplemented with 10% (vol./vol.) FBS and 50 μg/ml gentamicin. Cells were treated with cytokines as previously described 58 . Western blotting. HaCaT cells infected with HSV-1 were detached from growth area and loaded on a single-well NuPAGE 10% Bis-Tris gel (Thermo Fisher Scientific) using standard SDS-PAGE reducing conditions. Proteins were transferred to a PVDF membrane and approximately 2 mm wide strips cut from the top to the bottom. Sera from HSV-1 infected mice were diluted 1:250 in PBS, 1% BSA, 0.05% tween 20, and individually probed on a strip each overnight. Strips were washed and developed with anti-mouse Ig conjugated with horseradish peroxidase (GE Healthcare Life Sciences, Pittsburgh, PA) and enhanced chemiluminescence.

HSV-1 genome copy numbers in tissue.
Mouse tissue samples were immediately frozen on dry ice following euthanasia. Primary infection sites were collected using 4-mm biopsy punches (Miltex, York, PA). Tissue was ground into a powder over dry ice with autoclaved mortar and pestle. DNA was extracted from tissue using the Qiagen DNeasy Blood & Tissue kit according to manufacturer's instructions (Qiagen, Valencia, CA). Viral genome copy numbers were determined using quantitative real-time PCR with 500 ng DNA per reaction and the following primer/probe set: 5′ CGACCAACTACCCCGAT 3′ (forward primer), 5′ CACTATGACGACAAACAAAATCAC 3′ (reverse primer), and VIC-CAGTTATCCTTAAGGTCTC-MGBNFQ (probe). TaqMan Gene Expression Master Mix (Applied Biosystems, Foster City, CA) was used for PCR amplification on an Applied Biosystems StepOnePlus Real-Time PCR System. A DNA standard was generated using the above primer pairs and an aliquot of NS. The generated PCR product was cloned into the pGEM-T Easy vector (Promega, Madison, WI) and a single cloned confirmed by full length sequencing. The plasmid was linearized with PstI restriction enzyme and quantified by nanodrop, after which genome copy numbers/μl were calculated. Serial dilutions (10-fold) were used to generate a standard curve from 1 × 10 7 copies to 1 copy.

Cloning, expression and purification of recombinant HSV-gD protein.
The HSV-1 gD gene sequence (GenBank accession number: SBS69688) was amplified from DNA isolated from infected tissue using the primers 5′ CATGGGGTCCGCGGCAAATATG 3′ and 5′ GAGGACGGCTGGTCGTCTTCC