FormalPara Key Summary Points

Why carry out the study?

The relationship between eczema herpeticum (EH) and a history of Staphylococcus aureus skin infections or immune responses to S. aureus has not been explored in a highly matched cohort of deeply phenotyped atopic dermatitis (AD) subjects while controlling for key confounders/covariates.

Despite the high prevalence of S. aureus skin colonization and/or infection in subjects with AD, little is known about the immunoglobulin (Ig) G immune response to S. aureus antigens in this patient population.

We hypothesized that IgG levels against S. aureus antigens would associate with a history of S. aureus skin infections and that a history of S. aureus skin infections would significantly associate with a history of EH.

What was learned from this study?

IgG against the S. aureus virulence factor SElX was significantly associated with a history of S. aureus skin infection and a history of EH in AD subjects.

Subjects with a history of S. aureus skin infection(s) have more than a sixfold increased odds of also having experienced a case of EH even after controlling for critical confounders.

Introduction

Atopic dermatitis (AD) is a complex disease in which Staphylococcus aureus skin colonization, epithelial barrier disruption, and type 2 immune polarization are all thought to contribute to disease severity, allergen sensitization, and enhanced susceptibility to cutaneous bacterial and viral infections. There are unique and AD-specific viral skin complications that can be serious and even life-threatening, including eczema herpeticum (EH, caused by herpes simplex [HSV] infections), eczema molluscatum (caused by molluscum contagiosum [MCV] infection), eczema coxsackium (caused by coxsackievirus A16), eczema vaccinatum (caused by vaccinia virus), and the newly identified eczema monkeypoxicum (caused by monkeypox virus) [1, 2]. A previous study reported that patients with AD with a history of S. aureus skin infections are at higher risk for developing EH, while another study found a high association between EH and skin colonization specifically with S. aureus strains that produce toxic shock syndrome toxin-1 (TSST-1) [3, 4].

The aim of this study was to explore the associations between S. aureus and viral skin infections in AD subjects after controlling for key confounders using a cohort of 112 AD subjects with and without a history of EH. We tested (1) whether an adaptive response to S. aureus antigens was associated with either a history of S. aureus skin infections or a history of EH; (2) whether a serum biomarker of skin barrier or epithelial damage (lactate dehydrogenase) was associated with these histories; and (3) whether type 2 biomarkers commonly measured as part of routine clinical management of AD (absolute eosinophils and total immunoglobulin [Ig] E) were associated with these histories.

Methods

Study Procedures

A subset of subjects (and their samples) collected as part of the National Institute of Allergy and Infectious Diseases (NIAID)-funded, Atopic Dermatitis Research Network (ADRN) were studied. This multi-center, clinical registry study was designed to examine factors associated with susceptibility of AD participants to cutaneous viral dissemination and bacterial colonization/infection as well as the biomarkers related to AD sub-phenotypes (ClinicalTrials.gov Identifier: NCT01494142 [ADRN02]). The ADRN02 protocol was approved at the institutional review board (IRB) at each academic center (Ann & Robert H. Lurie Children’s Hospital of Chicago; Boston Children’s Hospital; Children’s Hospital of Los Angeles; Mount Sinai School of Medicine; National Jewish Health; Oregon Health & Science University; University of Rochester Medical Center; IRB: RSRB00015368) and was performed in accordance with the Helsinki Declaration of 1964 and its later amendments. Written informed consent was provided by the participant or parent/legal guardian, and written consent was provided by the participant, as applicable, before participation. Detailed medical history and case reports, disease severity assessments, and serum were available from the biorepository. Serum tests including total IgE, complete blood counts, and the Immunocap™ Phadiatop test for aeroallergens had been previously performed [3]. We received IRB approval to perform additional serum assays on these samples (IRB: RSRB00007325).

Medical History Questions

Subjects were considered to have had a history of S. aureus skin infections if they answered “yes” during the medical history questionnaire to the question “Have you ever had a Staph infection?” and answered “skin” to the question “Was this staph infection of the skin, bloodstream, bones, or orthopedic hardware?”. Subjects were considered to have had a history of EH if they answered “yes” during the medical history questionnaire to the question “Have you ever been diagnosed with eczema herpeticum by a health care provider?” and answered “yes” to the question “Did you ever have the following skin condition?”, after being read a description of EH and being shown photographs of EH. Subjects were considered to have a history of cutaneous MCV or human papillomavirus (HPV) if they answered “yes” to either of the following questions: “Have you ever had a skin infection with molluscum contagiosum?”, after being shown photographs of molluscum contagiosum infections, or “Have you ever had a skin infection with HPV (also called warts)?”, after being shown a photograph of HPV infection.

Sample Selection

Subjects were selected based on classification by history of EH (n = 56 EH+, n = 56 EH−) (Fig. 1). Subjects were matched by disease severity (Eczema Area and Severity Index [EASI] ± 6 points) and age (± 10 years) and, when possible, by race and gender (Table 1).

Fig. 1
figure 1

Figure created using BioRender (https://www.biorender.com/)

Flow chart of study design, matching criteria, and metrics. EASI Eczema Area and Severity Index, EH eczema herpeticum, Ig immunoglobulin, LDH lactate dehydrogenase.

Full size image
Table 1 Demographics for atopic dermatitis patients with or without a history of eczema herpeticum
Full size table

Biomarker Measurements

Serum Lactate Dehydrogenase

Serum lactate dehydrogenase (LDH) was measured by routine laboratory assay in the UR Medicine Labs at the University of Rochester Medical Center.

Serum IgG Measurements

Serum IgG measurements against 17 S. aureus antigens (Fig. 2) were determined utilizing arrayed imaging reflectometry (AIR). A previously described array termed StaphAIR consisting of 17 S. aureus antigens (Fig. 2) along with control proteins on the AIR platform was used for antigen-specific serum IgG measurements [5]. Briefly, S. aureus antigens were printed on amine-reactive silicon/silicon dioxide chips to create a microarray of probe spots, background blocked, and then stabilized for storage. Human serum was diluted 1:50 in Enhanced Assay Buffer (EAB: 25 mM Tris Base, 250 mM NaCl, 250 KCl, 3% w/v propylene glycol, 0.125% w/v Triton X-100, 1% w/v bovine serum albumin, pH 7.2 in nanopure water) with 20% fetal bovine serum (FBS) and incubated with the chips overnight with shaking at 4 °C. The chips were then washed with Assay Wash Buffer (AWB: 150 mM NaCl, 10 mM Na2HPO4, 10 mM NaH2PO4, 0.005% Tween-20, pH 7.2), incubated for 1 h with 5 µg/mL anti-human IgG in EAB with 20% FBS, washed again, and then dried with nitrogen gas. Chips were then imaged by AIR to detect molecular binding. Images were captured at 6, 11, 25, 50, 100, 250, 500 and 1000 ms CCD exposure times. Pixel intensity was quantified and applied to an intensity versus thickness calibration curve to calculate the thickness of IgG build onto the probes.

Fig. 2
figure 2

Figure created using BioRender (https://www.biorender.com/)

Schematic of Staphylococcus aureus antigen classes included in the StaphAIR assay. Amd Autolysin domain amidase, CHIPs chemotaxis inhibitory protein of Staphylococcus, ClfA staphylococcal clumping factor A, Gmd autolysin domain glucosaminidase, Hla alpha-hemolysin, Isd iron-regulated surface determinant protein, PVL Panton-Valentine leukocidin, SCIN staphylococcal complement inhibitory protein, TSST-1 toxic shock syndrome toxin-1

Full size image

Data Analysis

The thickness of IgG was calculated as described above and in previous studies [5]. Case reports, IgG measurements, and serum LDH measurements were imported into R for data manipulation, organization and creation of plots. Linear correlations and logistic regressions were performed in SAS OnDemand for Academics. The Bonferroni correction for multiple comparison was applied resulting in a p  < 0.001 being used as the threshold for significance in our logistic regression models [6].

Results

The StaphAIR assay was used to measure IgG against 17 S. aureus antigens in serum samples collected from 112 AD subjects. We selected S. aureus antigens representing five main classes of S. aureus proteins important for pathogenesis (see Fig. 2):

  1. 1.

    Iron acquisition proteins (iron-regulated surface determinant proteins IsdA, IsdB, IsdH)

  2. 2.

    Cell division proteins (autolysin domains glucosaminidase [Gmd] and amidase [Amd])

  3. 3.

    Immune evasion proteins (chemotaxis inhibitory protein of Staphylococcus [CHIPS]; staphylococcal complement inhibitory protein [SCIN])

  4. 4.

    Superantigens (staphylococcal enterotoxin A [SEA], B [SEB], and C [SEC]; staphylococcal enterotoxin-like G [SElG], I [SElI], and X [SElX], and TSST-1)

  5. 5.

    Cytotoxins (leukocidin F [LukF] and S [LukS], and alpha-hemolysin [Hla]).

A previous study had found associations between specific S. aureus IgG levels and AD severity (self-administered EASI), including immunomodulatory proteins [7]. We sought to determine if there were significant associations between specific S. aureus IgG levels and two clinical AD phenotypes: history of S. aureus skin infections and history of EH, while controlling for disease severity (EASI) measured by AD experts.

Logistic regressions were run for each S. aureus antigen to determine the relative odds of having a history of S. aureus skin infection, or history of EH, associated with each 1 unit increase in IgG thickness. The Bonferroni correction was used to set a p value < 0.001 as the threshold for significance and inclusion into the models [6]. Of the 17 antigens measured, only IgG against SElX was significantly associated with both a history of S. aureus skin infections and a history of EH (Table 2). IgG against SElX was significantly greater in the EH+ group than the EH- group (Table 1).

Table 2 Staphylococcus aureus antigens and relative odds of history of S. aureus skin infection or history of eczema herpeticum as determined by logistic regression
Full size table

A previous study found that AD subjects with EH had more severe disease, greater type 2 immune polarization, and increased incidence of cutaneous infections with S. aureus or MCV [3]. We sought to test if the association of S. aureus skin infections and EH held true when controlling for confounders (factors influencing both independent and dependent variables) and covariates (correlated with independent variable). We controlled for the main confounder of disease severity in our study design, and subjects were matched into the EH+ and EH− groups based on AD disease severity which was measured by AD experts (EASI). We considered the inclusion of additional variables into the model based on their associations with history of EH, history of S. aureus infection, their correlations with EASI, and our understanding of AD.

Serum LDH is thought to be a marker of barrier dysfunction, as serum LDH correlates with two key measures of skin barrier integrity: transepidermal water loss and stratum corneum hydration [8]. This barrier biomarker was used because no physiological measures of barrier function were collected in the ADRN02 registry study. We found that LDH did not associate with a history of S. aureus skin or EH infections and that LDH was not significantly different between EH+ and EH− groups (Tables 1, 3). This is likely due to the fact that EH+ and EH− samples were matched for disease severity (EASI), which may explain the lack of differences in LDH. We did observe a significant correlation (Pearson correlation coefficient of 0.32, p = 0.0005) between LDH and EASI, further supporting that matching by EASI score will mask any differences in LDH among the groups and reconfirming the observation that barrier dysfunction is greater in those with more severe disease [9,10,11,12,13].

Table 3 Relative odds of history of S. aureus skin infection and history of eczema herpeticum infection as determined by logistic regression
Full size table

We also evaluated the importance of two markers of type 2 immunity, namely, serum total IgE and absolute eosinophil count in the circulation. IgE and absolute eosinophils have been shown to be elevated in AD patients with a history of EH, and IgE has been implicated as a risk factor for EH among AD patients [3, 14]. These measurements have also been shown to be associated with disease severity [3, 15, 16]. We observed a significant correlation (Pearson correlation coefficient of 0.36, p = 0.0001) between IgE (log10 transformed) and EASI, and between absolute eosinophil count and EASI (Pearson correlation coefficient of 0.27, p = 0.0047). Nevertheless, we observed a significant difference in total IgE (log10 transformed; Mann–Whitney test, p = 0.0136) and absolute eosinophils (log10 transformed; Mann–Whitney test, p = 0.0143) between the EH+ and EH− groups, even when matching by EASI. This suggests that our matching by disease severity did not fully correct for changes in these variables. When added to the model, serum total IgE (log10 transformed) changed the odds ratio (5.48, 95% CI 2.33–12.92) by > 10% and, therefore, it was considered to be a confounder and significant contributor to the model in this study (Table 4: model 3).

Table 4 Relative odds of history of eczema herpeticum associated with subject characteristics, as determined by logistic regression
Full size table

History of cutaneous MCV or HPV infection was included as a co-variate, as a previous study found that patients with EH had a greater incidence of MCV infections [3]. Furthermore, we found a history of MCV or HPV to be significantly associated with history of S. aureus skin infection (Table 3). Lastly, we added IgG against SElX, as this significantly associated with both history of S. aureus skin infection and EH and was therefore treated as a confounder (Table 3). When adjusting for history of MCV or HPV infection, total IgE (log10 transformed) and IgG against SElX, those with a history of S. aureus skin infections had 6.60-fold increased odds of a history of EH compared to those without a history of S. aureus skin infections. (Wald Chi-Square type III p = 0.002, 95% CI 2.00–21.83) (Table 4: model 7).

Discussion

In this study we utilized a large, deeply phenotyped cohort of AD subjects with and without a history of EH to investigate associations between this serious AD-associated viral skin infection and a history of S. aureus infections, S. aureus immunoglobulin responses, and a biomarker of skin barrier disruption. Notably, this is the first study to measure IgG against 17 different specific S. aureus antigens in AD patients with and without a history of EH, and little is known about how antibody response against S. aureus antigens contribute to AD disease severity, barrier defects, or associated infections.

We found IgG against the S. aureus superantigen SElX to be significantly associated with history of S. aureus skin infection and history of EH. In addition to the traditional superantigen function of activating T cells, SElX is also able to bind neutrophils and disrupt IgG-mediated phagocytosis [17]. SElX is also unique because its gene is chromosomally located, as opposed to being located on a mobile genetic element like many other virulence factors [18, 19]. The chromosomal location of the SElX gene means that nearly all strains of S. aureus are able to produce SElX, which may increase the likelihood of developing an immune response to this antigen. When we quantified virulence factors produced by S. aureus strain USA300 in our previous study, the expression level of SElX was the highest among three superantigens, SElK, SElQ, SElX, secreted in the culture media of this strain [20]. We also observed that the five S. aureus virulence factors with the most significant IgG response associated with a history of S. aureus skin infection or history of EH were all immune evasion proteins (Table 2). These virulence factors (SElX, CHIPS, SCIN, LukS, LukF) are able to act on neutrophils and inhibit their immune function. Neutrophils are the first line of immune cells to respond to S. aureus infection and are able to secrete antimicrobial factors, produce reactive oxygen species, phagocytose bacteria, and release neutrophil extracellular traps [21]. The importance of neutrophils in clearing S. aureus infection has been made clear in patients with congenital and acquired diseases of impaired neutrophil number and/or function [22]. Arguably, antibodies capable of neutralizing the activity of these immune-evasion proteins would give the host more time to recruit immune cells to the site of infection to clear the bacteria.

Our study has a number of limitations. One limitation is that this is a cross-sectional study and that the history of S. aureus skin infection and EH were assessed at one point in time and not over time or at the time of an acute infection. Because of this, we are unable to address temporality. These medical histories are subject to recall bias, which could result in misclassification of history of S. aureus skin infection, history of MCV or HPV infection, and history of EH.

A limitation of the StaphAIR assay is that antibodies against many of these antigens are not commercially available and, therefore, we were unable to create standard curves that would allow us to determine the concentration of each of the antigen-specific antibodies in the serum. Instead, we have a semi-quantitative measurement of antibody build on the probes (measured as thickness) [5]. Importantly, for our purposes it was more important to compare values across samples rather than to know the absolute amount. While our platform included 17 S. aureus antigens, this is in no way exhaustive of all the potentially relevant S. aureus virulence factors important for human colonization and/or infection.

Conclusions

This study has demonstrated that there is a significant association between a history of S. aureus skin infection and history of EH, even after controlling for confounders and covariates such as disease severity, epithelial barrier disruption (using LDH as a surrogate), type 2 immunity, and history of MCV or HPV skin infection. Furthermore, we found that IgG against the S. aureus antigen SElX is associated with history of S. aureus skin infection and history of EH. These findings support that skin infection caused by S. aureus or an increased susceptibility of AD patients to S. aureus skin infection significantly impacts susceptibility to viral infection. Reducing S. aureus skin infections in patients with AD may help reduce the risk of developing serious viral skin complications, such as EH. A greater understanding of these mechanisms and the key S. aureus virulence factors involved may help better identify patients at a greater risk of EH.