Vaccinia virus binds to the scavenger receptor MARCO on the surface of keratinocytes

Abstract

Patients with altered skin immunity, such as individuals with atopic dermatitis (AD), can have a life-threatening disruption of the epidermis known as eczema vaccinatum after vaccinia virus (VV) infection of the skin. Here, we sought to better understand the mechanism(s) by which VV associates with keratinocytes. The class A scavenger receptor known as MARCO (macrophage receptor with collagenous structure) is expressed on human and mouse keratinocytes and found to be abundantly expressed in the skin of patients with AD. VV bound directly to MARCO, and overexpression of MARCO increased susceptibility to VV infection. Furthermore, ligands with affinity for MARCO, or excess soluble MARCO, competitively inhibited VV infection. These findings indicate that MARCO promotes VV infection and highlights potential new therapeutic strategies for prevention of VV infection in the skin.

Figure 1. MARCO expression in atopic dermatitis

(a,b) Skin biopsy samples taken from a normal donor and an AD donor at non-lesional and lesional sites were analyzed for Keratin-14 (K14), which is predominantly expressed by basal keratinocytes, and human MARCO expression, and compared to goat IgG (gIgG) and mouse IgG (mIgG) controls. Scale bars = 50μm. Images in (b) are one set of representative images from an experiment with 2 normal skin donors and 2 AD donors (see Supplementary Fig. S1). (c) NHEK were treated with 50ng/ml IL-4 and/or IL-13 for 24 hours before RNA isolation and qPCR for MARCO expression. Error bars indicate SEM, n=3. Differences between treatment conditions were not significant, as determined by One-way ANOVA with Tukey post-tests.

Figure 2. Vaccinia binds directly to human and mouse MARCO

Indicated concentrations of VV were added to plates coated with human MARCO (hMARCO) (a), mouse MARCO (mMARCO) (b), human OLR1 (d) and human MSR1 (e). Virus bound to the immobilized protein was quantified by ELISA. Error bars indicate SEM, n=3, with non-linear regressions plotted as solid lines. c, Plates coated with hMARCO were incubated with 6×10⁶ PFU/ml VV in the presence of Poly(I). Bound virus was quantified by ELISA. Each individual data point is plotted. Data in (c) was transformed using a logarithmic x-axis, and the half-maximal inhibitory concentration (IC₅₀) value and the 95% confidence interval for this value were calculated using Graphpad Prism. Poly(I) molarity was calculated based on the molecular weight of each individual nucleotide in the polymer, with average molecules ranging in length up to 1000 nucleotides. f, indicated amounts of HIV-1 pseudovirus (PsV) were added to plates coated with hMARCO and bound virus quantified by ELISA. Error bars indicate SEM, n=2, with non-linear regression plotted as a solid line. a–f, data presented are from one representative experiment of two independent experiments.

Figure 3. Overexpression of MARCO increases VV infection in keratinocytes

Age-matched female Wild-type (WT) and MARCO^−/− mice were infected with VV by skin scarification, and six days after infection wounds were (a) biopsied for histology and (b) sizes were quantified. a, Representative images of the skin of uninfected WT and MARCO^−/− mice and skin from the edges of the wounds of mice infected with VV are displayed. Scale bars = 50μm. b, Wound sizes were quantified with ImageJ, n=12. c, HaCat keratinocytes stably overexpressing MARCO (labelled [MARCO]), or control cells (pcDNA3) were infected with VV. Plaques were stained and quantified 24 hours after infection, n=3. d, HaCat keratinocytes stably overexpressing MARCO (labelled [MARCO]), or control cells (pcDNA3) were infected with 1000 PFU/ml VV for 24 hours before quantification of viral DNA, n=3. b–d, error bars indicate SEM, all comparisons were made using two-tailed Student’s T-tests, ns = not significant, ** P < 0.01, *** P < 0.001. All data are from representative experiments repeated at least two times.

Figure 4. Scavenger receptor antagonists prevent vaccinia infection

HaCat keratinocytes were infected with 100 PFU/well VV in the presence of the indicated concentrations of (a) Poly(I:C), Poly(I), and Poly(C), or (b) dextran sulfate (Dxs). Plaques were quantified 24 hours after infection. a, Poly(I:C)/Poly(I)/Poly(C) molarities were calculated and normalized based on the molecular weight of each individual nucleotide in the polymer, with average molecules ranging up to 1000 nucleotides in length. a, two-way ANOVA with bonferroni post-tests was used to compare Poly(I) and Poly(I:C) treatment to Poly(C) treatment at each concentration, *** P < 0.001. c, HaCat were pretreated with TLR ligands for 24 hours before infection with VV. HaCat (d) and NHEK (e,f) were infected with VV after Poly(I:C) pretreatment for 24 hours. c–e, plaques were quantified 24 hours after infection. f, Viral mRNA was quantified by qPCR 8 hours after infection. c–f, statistical significance was determined using One-way ANOVA with Tukey post-tests, * P < 0.05, ** P < 0.01, *** P < 0.001, nd, not detectable. g, VV was pretreated for 1 hour at 37°C with 50μg/ml bovine serum albumin (BSA) or recombinant human MARCO before adding to NHEK to assess viral plaque formation. A two-tailed Student’s T-test was used to determine statistical significance, *** P < 0.001. a–g, error bars indicate SEM, n=3. h, 50μl of PBS, or PBS containing 5mg/ml Poly(I) or Poly(C) were applied to depilated back skin of age-matched wild-type female mice for 15 minutes prior to infection with VV. Wound sizes were quantified at 6 days post-infection. Error bars indicate SEM, n=4, statistical significance determined using One-way ANOVA and a Bonferroni post-test, * P < 0.05.