Skin and nasal colonization of coagulase-negative staphylococci are associated with atopic dermatitis among South African toddlers

Background Skin colonization with coagulase-negative staphylococci (CoNS) is generally beneficial, but recent investigations suggest its association with flares and atopic dermatitis (AD) severity. However, this relationship remains unclear. Objective To assess patterns of staphylococcal colonization and biofilm formation in toddlers with and without AD from rural and urban South African settings. Methods We conducted a cross-sectional study of AD-affected and non-atopic AmaXhosa toddlers from rural Umtata and urban Cape Town, South Africa. CoNS isolates were recovered from lesional, nonlesional skin samples and the anterior nares of participants. Identification of the staphylococci was achieved by MALDI-TOF mass spectrometry. The microtiter plate assay assessed in-vitro biofilm formation. Results CoNS and S. aureus commonly co-colonized nonlesional skin among cases (urban: 24% vs. 3%, p = 0.037 and rural 21% vs. 6%, p<0.001), and anterior nares in urban cases (24% vs. 0%, p = 0.002) than the control group. S. capitis colonization on nonlesional skin and anterior nares was positively associated with more severe disease in rural (48.3±10.8 vs. 39.7±11.5, P = 0.045) and urban cases (74.9±10.3 vs. 38.4±13, P = 0.004), respectively. Biofilm formation was similar between cases and controls, independent of rural-urban living. Conclusion CoNS colonization is associated with AD and disease severity and may be implicated in AD exacerbations. Studies are needed to understand their underlying pathological contribution in AD pathogenesis.

a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 [22][23][24]. In vitro studies have shown that S. epidermidis may antagonize [25] or cooperate [26,27] with S. aureus in biofilm formation. These studies on the interactions of CoNS in mixed biofilms with S. aureus have primarily included healthy individuals [12,25,28] and are limited in patients with AD [26]. Moreover, no study has compared these interactions between AD patients and healthy individuals to assess whether they differ based on AD disease status.
We explored the skin and nasal staphylococcal colonization patterns in AD cases and healthy controls across urban and rural environments. We also evaluated biofilm formation of staphylococcal species, including S. aureus and CoNS interactions in mixed-species biofilms.

Study population
We recruited 220 AmaXhosa (same ethnolinguistic background) toddlers aged 9-38 months (mean, 22.5 months; standard deviation, 7.3 months) with and without AD from Umtata, South Africa, and Cape Town, South Africa. The duration of the recruitment was between February 2015 and May 2016. Toddlers with an AD diagnosis based on the United Kingdom Working Party diagnosis of atopic eczema [29] were recruited through the Red Cross Children's War Memorial Hospital (Cape Town, South Africa) and Nelson Mandela Academic Hospital (Umtata, South Africa). AD severity was measured, at the time of specimen collection, using the objective scoring of atopic dermatitis (SCORAD) index [30]. Similarly, aged toddlers without a clinical diagnosis of AD were recruited from the community early development centers in Cape Town and Umtata. Swabs were collected from the anterior nares and nonlesional skin (area with most normal-appearing skin-usually the back) in both cases and controls as previously described [31]. We also collected lesional skin swabs (i.e., most active area of eczematous skin with acute and or chronic changes) from only cases. Collected swabs were placed in STGG and stored at -80˚C for subsequent batch processing.

Bacterial isolation and species identification
To recover the CoNS isolates, 20μL of each specimen was inoculated on mannitol salt agar (National Health Laboratory Services [NHLS], South Africa), streaked for single colony growth, and incubated at 37˚C in ambient air for 48 hours. All morphologically distinct colonies were selected for further analysis. Species identification was performed by matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry. All identified CoNS isolates were stored at -20˚C in skim milk-tryptone-glucose-glycerine (STGG) medium (NHLS, South Africa) for further batch processing. S. aureus isolates included in the analyses within this study were recovered previously [31].

Assessment of in vitro mono-species and dual-species biofilm formation
Biofilm formation was assessed in vitro using the crystal violet microtiter assay (32). Staphylococcal isolates in STGG were streaked on 2% blood agar and incubated overnight at 37˚C in ambient air. A single colony was inoculated into 3mL 3% tryptic soy broth supplemented with 0.1% glucose (gTSB; Merck, South Africa) followed by incubation with shaking for 24 hours. Thereafter, the broth culture was diluted at 1:100 for mono-species biofilms and 100μL:100μL for dual staphylococcal biofilms in a flat-bottomed 96-well microtiter plate (Lasec, South Africa) previously described [26]. Then, the plates were statically incubated in ambient air at 37˚C for 24 hours. S. aureus ATCC 29213 and uninoculated gTSB were included as positive and negative controls, respectively. Planktonic bacterial suspensions were discarded, followed by rinsing the microtiter plates three times in sterile reverse osmosis (RO)-filtered water and dried at 60˚C for 1 hour. The formed biofilms were stained with 200μL 0.1% crystal violet and incubated at room temperature for 10 minutes and followed by discarding excess crystal violet. Thereafter, the stained biofilms were solubilized with 200μL 30% acetic acid and incubated at room temperature for 30 minutes. The absorbance was measured at 492nm using the Mindray MR 96A ELISA Microplate Reader (Vacutec, South Africa). Biofilm biomass was analyzed using OD readings as previously described (OD>OD negative control [OD c ], non-producer; OD c <OD<2�OD c , weak producer; 2�OD c <OD<4�OD c , moderate producer; and OD>4�OD c , strong producer) [32].

Statistical analysis
We used R studio version 4.0.4 to conduct statistical analyses. Species with less than ten isolates were arbitrarily considered rare in this cohort and classified as "rare CoNS." The Fisher's exact test was performed for comparisons between categorical variables. Differences in continuous dependent variables by a categorical variable were assessed using either the Wilcoxon ranksum test or the Kruskal-Wallis test. We used the Bonferroni method to adjust for multiple comparisons. All analyses were two-tailed. The log2 fold change method was used to compare the fold change of the biofilm biomass from dual CoNS-S. aureus biofilms compared to monospecies S. aureus biofilms. The effect on biofilm biomass was arbitrarily classified based on log2 fold change as follows: "no effect" (> -0.1 to <0.1); "weak, positive" (0.1 to <0.4); "moderate, positive" (0.4 to <0.9); "strong, positive" (>0.9); "weak, negative" (-0.1 to < -0.4); "moderate, negative" (-0.4 to < -0.9). Correlation between continuous variables was assessed using Pearson's correlation. A p-value of <0.05 was considered statistically significant.

Statement of ethics
The parent study was approved by the Human Research and Ethics Committee of the Faculty of Health Science, University of Cape Town (HREC/REF: 451/2014) and the Western Cape Provincial Child Health Research Committee. Additional ethical approval specifically for the study was obtained from the Human Research and Ethics Committee of the Faculty of Health Science, University of Cape Town (HREC/REC: 668/2020). No additional data was collected other than that approved in the parent study. Written informed consent and assent were provided by guardians and participants, respectively. All data obtained and generated during the study were kept confidential. This research was conducted in accordance with the Declaration of Helsinki. Table 1 describes the characteristics of the participants. We recovered 381 staphylococcal isolates from lesional skin (n = 112), nonlesional skin (cases: n = 117; controls: n = 60) and anterior nares (cases: n = 67; controls: n = 25) from cases and controls (Fig 1). These isolates  Table).

Risk factors and staphylococcal colonization patterns
The distribution of staphylococcal colonization in cases and controls by geographic location is shown in Fig 2. Colonization with only S. aureus, CoNS, or both was common on lesional skin in the rural and urban cohorts. In the urban cohort, 47% (16/34) of controls had no staphylococcal colonization on nonlesional skin compared to the nonlesional skin of cases (20% [9/ 45], P = 0.015). Co-colonization with CoNS species and S. aureus on nonlesional skin was more prevalent in cases from both rural (21% [21/58] vs. 6% [3/53], P <0.001) and urban (24% [9/45] vs. 3% [1/34], P = 0.037) settings than controls. Similarly, the colonization rates for only S. aureus or CoNS were higher on the nonlesional skin of cases than controls, independent of geographic location. Urban controls had no co-colonization with CoNS species and S. aureus in their anterior nares (Fig 2).

PLOS ONE
sampling sites revealed a higher colonization prevalence of S. hominis in the nonlesional skin than lesional skin (P = 0.016) and anterior nares (P = 0.0004) in urban cases. This trend was also observed with S. capitis nonlesional skin colonization compared to anterior nares (P = 0.027) but not AD lesional skin (P = 0.069) among urban cases.

Biofilm propensity of staphylococcal species
Due to the low numbers of certain species, we opted to report the overall S. aureus and CoNS biofilm phenotypes without stratifying by species. We observed an overall modest-to-high prevalence of strong biofilm-producing-staphylococci (43%-72%) regardless of disease status, sampling site, or geographic location (Fig 4). Furthermore, S. aureus and CoNS isolates that lack the ability to form biofilms were only identified in the rural cohort. For 61 S. aureus-CoNS pairs in co-culture biofilms, we assessed the change in the co-culture biofilm biomass compared to S. aureus mono-culture biofilm biomass to determine the effect of CoNS on S. aureus biofilm biomass (Fig 5). We observed no difference in the biofilm biomass of mixed  biofilms compared to mono-species S. aureus biofilms between cases and controls when the geographic location was not considered. Furthermore, there was no relationship between the biofilm biomass fold change in mixed biofilms and disease severity, independent of geographic location (Fig 6).

Discussion
This study examined staphylococcal colonization and biofilm formation in AD and healthy toddlers from the urban and rural settings in South Africa. Our results showed that AD cases are more commonly colonized by CoNS than controls, although this was mainly limited to nonlesional skin. Also, this study provided evidence that co-colonization with CoNS and S. aureus on nonlesional skin and anterior nares is associated with more severe disease. However, this was dependent on rural-urban living. Staphylococcal biofilm formation did not differ between cases and controls and might not be significant in staphylococcal pathogenicity in AD.
We observed that nonlesional skin colonization with S. capitis and S. haemolyticus was more common in urban and rural cases, respectively, compared to their control counterparts. This observation is similar to previous reports that showed increased abundance of these two CoNS species in cases compared to controls [15]. A recent randomized phase I clinical trial revealed that S. hominis is a protective CoNS, especially against S. aureus colonization, and Table 4. Mean values of objective SCORAD based on the presence or absence of CoNS in rural cases.

Species
Lesional skin Nonlesional skin Anterior nares Sample + μ ± SD Sampleμ ± SD p-value Sample + μ ± SD Sampleμ ± SD p-value Sample + μ ± SD Sampleμ ± SD p-value  Table 5. Mean values of objective SCORAD based on the presence or absence of CoNS in urban cases.

Species Lesional skin Nonlesional skin Anterior nares
Sample + μ ± SD Sampleμ ± SD p-value Sample + μ ± SD Sampleμ ± SD p-value Sample + μ ± SD Sampleμ ± SD p-value improved AD disease scores [33]. The prevalence of S. hominis is generally reduced in AD lesions, especially in severe AD, compared to nonlesional skin of both cases and controls [15,34,35]. In contrast, some studies noted an increase in its prevalence in AD [36] and no difference in another investigation [37]. In our study, colonization with S. hominis on nonlesional skin was comparable between cases and controls, independent of geographic location. However, lesional skin was less frequently colonized with S. hominis than nonlesional skin of cases as previously noted [34], although this was limited to urban participants. These findings support the notion that S. hominis is primarily protective and is depleted in AD. Moreover, the differences in observation based on location suggest a potential influence of the environment in colonization dynamics. Rural and urban living has been shown to strongly influence the diversity of the skin bacteriome, particularly in early life [38]. This may be due to distinct environmental exposures in rural and urban settings and degrees of interaction with the environment [39]. Specifically, rural living is associated with more diverse skin bacteriomes than urban counterparts [40]. Therefore, we posit that rural living supports the acquisition of a more diverse skin bacteriome [41] and encourages the survival of "protective CoNS" despite AD disease. In contrast, these "protective CoNS" are limited in individuals living in urban areas, and their absence is particularly experienced in AD cases where S. aureus growth is uninhibited [42].  The prevailing inflammatory and immunological profile in AD has also been shown to influence staphylococcal colonization, with Th2-and Th17-dominance shown to promote S. aureus growth, especially in severe disease [43]. In contrast, CoNS generally encourages antiinflammatory skin responses [44], and inflammation in AD inhibits CoNS growth [22]. Moreover, the prevailing immunological profiles in AD have been shown to differ based on rural and urban living, with rural living frequently associated with microinflammation compared to urban living [15,45,46]. These observations possibly explain the differences in CoNS colonization and their association with severity between rural and urban toddlers. Collectively, these findings highlight that although some CoNS species may be commensals that are beneficial on healthy skin, they can adapt and increase AD lesions consequently contributing to disease severity. In contrast, other CoNS, such as S. hominis, are inhibited. The mechanisms by which this adaption, or lack thereof, occur are not clear and therefore warrant further study.
Blicharz et al. [47] showed that adult patients with AD co-colonized with S. aureus and CoNS on the skin and anterior nares had lower IgE levels than those colonized only with S. aureus. This trend negatively correlates with AD severity [48,49], suggesting that S. aureus and CoNS co-colonization relates to less severe disease. However, this might depend on the co-colonizing CoNS species or strains. A positive correlation between disease severity and the simultaneous colonization of CoNS and S. aureus has been demonstrated [34]. We observed that cases co-colonized with S. aureus and CoNS on lesional, nonlesional skin, or anterior nares had higher disease severity scores. However, this observation depended on geographic location. It has been suggested that the concurrent increase in S. aureus and CoNS colonization reflects a compensatory effect of the CoNS in response to S. aureus proliferation in AD, in particular severe AD [3]. In contrast, CoNS may cooperate with S. aureus promoting its deleterious effects in AD [13,50]. Nonetheless, our analysis is limited to correlating the simultaneous staphylococcal growth and AD severity. Therefore, future studies are needed to assess the clinical relevance of this co-colonization and how it contributes to AD flares and disease severity. Data on the relationship between the colonization of specific CoNS species on AD skin and disease severity is inconsistent [34,43,51]. Our study observed that colonization with S. capitis on nonlesional skin and anterior nares in rural and urban cases was positively associated with higher severity scores than those not colonized. This observation is consistent with Edslev et al. [34] who noted a positive association between S. capitis colonization and AD severity. However, an absence of correlation has also been demonstrated [43]. Our findings suggest that S. capitis in this cohort directly contributes too AD pathogenesis through mechanisms not been described in the literature. Furthermore, like previous reports, we observed no relationship between S. hominis and severity scores independent of colonization site and geographic location [43]. Our findings contrast other reports that reported a negative correlation between S. hominis and disease severity [34]. These conflicting findings could also indicate differences at the host level [52], which may differ across ethnicities, age groups, or geographies, affecting bacterial colonization in skin diseases [53] and their association with disease severity. Overall, our results suggest that the relationship between CoNS and AD severity depends on the colonizing CoNS species.
The disrupted skin barrier in AD lesions exposes the underlying matrix components to which staphylococci bind thereby promoting staphylococcal adhesion to the epidermis [20,54]. Staphylococcal binding is followed by biofilm formation, which leads to persistent colonization [55]. Patients with AD are frequently colonized by strong biofilm-producing S. aureus and S. epidermidis strains, which is associated with more severe disease [26,39]. We noted a modest-to-high prevalence of strong biofilm-producing staphylococcal isolates, which did not differ between cases and controls, regardless of sampling site and geographic location. CoNS may antagonize or synergize with S. aureus in mixed biofilm cultures [12,27]. However, the link between these interactions and AD pathogenesis remains poorly understood and limited to patients with mild AD [26]. Compared to mono-S. aureus biofilms, we found no overall difference in the fold change of mixed biofilms of co-colonizing S. aureus and CoNS in cases and controls. Also, this was not associated with disease severity independent of geographic location. Collectively, these findings suggest that the ability to forms biofilms and the outcome of these interactions are innate features of CoNS and S. aureus and may not relate to disease pathogenesis. Of note, the AD skin environment may alter the ability of staphylococci to form biofilms in vivo. For example, alkaline pH, a common feature of AD skin [56], and the cathelicidin LL-37 [57] hamper staphylococcal biofilm formation in vitro [58]. Moreover, S. aureus and S. epidermidis grow differently in dual-species biofilms at high pH in vitro [59], affecting their cooperation in biofilm growth and, consequently, biofilm biomass. In contrast, other host factors, including the pro-inflammatory environment in AD, can augment staphylococcal biofilm formation [22]. The differential effect of various AD disease features on staphylococcal biofilm formation warrants the need for in vivo studies. These investigations would include studies using murine [60] or ex vivo human skin [61] models of healthy and AD skin to elucidate the dynamics of staphylococcal biofilm formation in AD and how these contribute to disease parameters.

Strengths and limitations
The strengths of this study include the following. Firstly, a unique rural and urban cohort provided insights into how staphylococcal colonization differs between cases and controls in toddlers across geographies with distinct environmental exposures. Secondly, although S. epidermidis is the most dominant and studied skin CoNS in AD, we expanded our analyses to other CoNS species on the skin and anterior nares. Thus, the analyses provided a comprehensive characterization of the colonization and biofilm propensity of the staphylococci and how they associate with disease severity in early childhood AD. Nonetheless, our study has some limitations. These include using culture-dependent approaches to describe staphylococcal colonization patterns and biofilm formation. These methods may underrepresent the staphylococcal community due to reliance on bacterial viability and observable colonies formed on solid media. Moreover, we focused our analyses on the presence/absence of staphylococci from the skin and nasal samples. This does not evaluate differences in the abundance (based on colony-forming units) of the staphylococci and its correlation with measures of diseases. The cross-sectional design prevents the assessment of the causality between staphylococcal colonization and biofilm phenotypes with AD severity. Although all cases had used emollients before sampling, we could not account for the timing of emollient administration before sampling. This scenario may have affected the patterns of staphylococcal colonization presently observed [62] and may explain the modest prevalence of CoNS species on lesional skin compared to other studies. Moreover, the presently found differences were by chance due to our present small sample size coupled with the multiple levels of stratification by disease status, geographic location, and sampling site. Therefore, we advocate for future studies which will investigate CoNS colonization in AD in a larger cohort. Despite these limitations, our study provides a unique knowledge on skin and nasal staphylococcal colonization, biofilm propensity, their association with disease severity in early childhood AD, and how these factors differ based on rural-urban living.

Conclusion
In summary, although CoNS colonization is generally beneficial on healthy skin, we show that CoNS, particularly S. capitis, are associated with AD severity. This observation could play a role in the pathogenesis and exacerbation of AD through mechanisms not yet fully described in the literature. Furthermore, this may be dependent on their interactions with S. aureus. Our study highlights that while S. aureus remains the most studied Staphylococcus species in the pathogenesis of AD, CoNS may contribute to AD exacerbations. It also highlights the need for further in vivo studies on specific CoNS species or strains and how they contribute to AD exacerbation. Understanding the dynamics of the pathological role of CoNS will allow their targeting by therapeutic strategies aimed at countering complications from CoNS-associated infections in AD.
Supporting information S1