Journal Logo

Original Articles

Temporal Variation of the Facial Skin Microbiome: A 2-Year Longitudinal Study in Healthy Adults

Hillebrand, Greg G. PhD; Dimitriu, Pedro PhD; Malik, Kausar MPhil, MBA; Park, Yulia PhD; Qu, Di PhD; Mohn, William W. PhD; Kong, Rong PhD

Author Information
Plastic and Reconstructive Surgery: January 2021 - Volume 147 - Issue 1S-2 - p 50S-61S
doi: 10.1097/PRS.0000000000007621
  • Free

Abstract

Skin is colonized by a diverse community of microbes including bacteria, fungi, and viruses collectively called the skin microbiome. These commensal microbes play a key role in skin barrier function, educating the immune system, and protecting us from pathogens while contributing to healthy-looking skin. There is high variation within and between people in the diversity, composition, and function of the skin microbiome depending on body site, age, gender, and lifestyle factors.1 It is also important to understand the dynamics and stability of the skin microbiome as we age. However, this temporal variation of the facial skin microbiome, especially over years and decades, is not well understood, likely because large base-size longitudinal studies are more difficult to execute compared with cross-sectional designs.2–5 We therefore set out to build on the prior longitudinal studies to better characterize how, and how much, the facial skin microbiome changes with age.

The pioneering longitudinal study of Grice et al.5 suggested that the temporal stability of the skin microbiome is body site dependent. More microbially diverse facial skin sites like the glabella and alar creases showed high temporal stability compared with dryer sites like the forearm. The study was limited by its small base size (n = 5) and short duration (4 months).

Flores et al.3 also profiled microbiome temporal variability at several body sites including forehead skin, over a 3-month period in 85 healthy, mostly white college-age adults. Using 16S amplicon sequencing, they observed significant intrasubject temporal variation in both diversity and community structure over the 3-month time period. The relationship between microbiome diversity and temporal stability was assessed at the same body site. Again, microbiome stability was related to diversity; individuals with more diverse forehead skin communities tended to be less stable over time compared with individuals with less diverse forehead communities suggesting that temporal variability, like composition, is highly personalized. This study was also limited by its short duration of only 3 months. Skin microbiome variability tended to be driven by the presence or absence of transient taxa on the skin surface. Interestingly, individuals with more stable forehead skin microbiomes had a greater abundance of Staphylococcaceae and Corynebacteriaceae.

Oh et al.2 collected skin microbiome samples at multiple skin sites for up to 2 years and used shotgun metagenomics to analyze microbiome composition at the strain level. They found that in the absence of significant perturbations such as exposure to antibiotics or illness, the bacterial and fungal composition of the microbiome remained remarkably stable; foot skin sites were least stable and sebaceous skin sites were most stable. This study was limited by its small base size (n = 12). The overarching theme from this prior work is that the skin microbiome, especially the facial skin microbiome is remarkably robust and resilient to external stress, at least in the short term in healthy people.

Even acute washing with antibacterial soaps and antiseptic treatments do not seem to modulate the skin microbiome in the hours postwash.6,7 Finally, skin inflammatory disease is clearly associated with an altered stability of the skin microbiome.8,9 Over a period of weeks, the lesional skin of atopic dermatitis patients shows higher intrasubject variability compared with nonlesional skin sites.10

In 2014, we initiated a longitudinal study aimed at carefully tracking the facial skin condition of 161 healthy adult men and women over an entire decade. We call the study the “Cinco de Mayo Study” because the first clinical site visit was on May 5 and subjects are invited back for annual measurements on or around May 5 of each year to control for seasonal variation in skin condition. In 2017, we added facial skin swabbing for 16S ribosomal RNA gene amplicon sequencing to track the facial skin microbiome. The longitudinal design is more powerful for understanding the influence of aging on the skin microbiome over years and decades compared with our previous large base size cross-sectional study design.1 This article reports on the microbiome results and their association with facial skin condition over the 2-year period from 2017 to 2019.

METHODS

Study Design and Subjects:

This was a single-center, multivisit, longitudinal, IRB-approved study in self-reported healthy men and women, ages 28–64 (mean age = 47 in 2019). There were 3 site visits, each in early May 2017 (baseline), 2018 and 2019. Table 1 shows subject enrollment by year and gender. Of the 155 subjects who enrolled in 2017, 115 completed all three site visits. Ninety percent of the subjects were white. Subjects were asked to not shower, bath, or cleanse their face on the day of their site visit and to arrive without makeup. In this way, a “steady-state” skin microbiome and skin surface pH would be measured. After signing the informed consent statement, the subjects’ cheek and forehead were immediately swabbed, and cheek skin surface pH was measured (see below for details). The subjects then cleansed their face with a commercial facial cleanser, rinsed and dried and then acclimated in a controlled temperature and humidity environment for 20 minutes before proceeding to subsequent measurements.

Table 1. - Subject Demographics*
2017 (n) 2018 (n) 2019 (n) Longitudinal (n)
Men 66 64 56 51
Women 89 86 78 64
Total 155 150 134 115
*“Longitudinal” indicates totals of subjects who enrolled every year.

Skin Swabbing, DNA Extraction, Amplicon Sequencing, and Sequence Processing

Skin swabbing and amplicon sequencing are described in detail by Dimitri et al.1 Briefly, sterile BBL Culture Swabs EZ II were wetted with sterile buffer (50 mM Tris, 1 mM EDTA, 0.5% Tween 20 [pH 7.6]) and used to collect microbiome samples from the cheek and forehead in a standardized procedure. To minimize swabbing variation, the same operator conducting the swabbing at all time points. One swab was used to sample the skin microbiome in a trapezoidal area on the right and left cheek (Fig. 1, above). A separate swab was used to sample a rectangular area on the forehead (Fig. 1, above). Swabs were frozen at −20oC and then batched processed for extraction, amplification, and sequencing.

Fig. 1.
Fig. 1.:
Example front view facial image (woman age 31 in 2019). Standard 2 white light image (above, left). Standard 2 white light image depicting skin areas in blue that were swabbed for microbiome analysis (above, right). Red porphyrin image with cheek image analysis overlay (below, left). Red porphyrin image with forehead image analysis overlay (below, right).

To extract genomic DNA, clipped swabs were placed into a MoBio PowerMag Soil DNA Isolation Bead Plate. DNA was extracted following MoBio’s instructions on a KingFisher robot. Bacterial 16S rRNA genes were PCR-amplified with dual-barcoded primers targeting the V1–V3 region (27F 5′-AGAGTTTGATCCTGGCTCAG-3′, 519R 5′-ATTACCGCGGCTGCTGG-3′), as per a modified version of the protocol of Kozich et al.11 Amplicons were purified and normalized using a SequalPrep kit (Invitrogen, Eugene, OR). Purified amplicons were quantitated with a Qubit 2.0 fluorometer and pooled for Illumina sequencing. The 16S ribosomal RNA genes were sequenced on an Illumina MiSeq (v. 3 chemistry) using the dual barcoding protocol of Kozich et al.11 Bacterial sequences were processed and clustered into operational taxonomic units (OTUs) with the mothur software package (v. 1.39.5),12 following the recommended procedure (https://www.mothur.org/wiki/MiSeq_SOP; accessed July 2019). Paired-end reads were merged and curated to reduce sequencing error.13 Chimeric sequences were identified and removed using VSEARCH.14 The curated sequences were assigned to OTUs at 97% similarity using the OptiClust algorithm15 and classified to the deepest taxonomic level that had 80% support using the naive Bayesian classifier trained on the Greengenes taxonomy outline (version 13_8). A consensus taxonomy for each OTU was obtained by 50% majority rule. To assess the impact of contamination, we included a swab, an extraction kit-only, and a DNA template-free control in each 96-well plate.

Microbial Alpha and Beta Diversity

We excluded OTUs occurring in fewer than 3 samples with a count of less than 3 and calculated alpha (Shannon) and beta (Bray–Curtis dissimilarity) diversity indices with the R phyloseq package.

Facial Imaging

Facial images (front, right, and left views) were collected with the VISIA-CR 4.1 Imaging System (Canfield Scientific, Parsippany, NJ) equipped with a Canon EOS 6D camera under seven lighting modalities (Standard 1, Standard 2, Standard 3, Cross-Polarized, Parallel-Polarized, Red, and Blue EX). An example front view image is shown in Figure 1 (above, left). Regions on the left and right cheeks (Fig. 1, below, left) and forehead (Fig. 1, below, right) that matched the facial areas swabbed for microbiome analysis were masked and analyzed using the VAESTRO Image Analysis Toolkit (Canfield Scientific). The following facial features were quantified: cheek wrinkles, cheek visible spots, cheek brown spots, cheek and forehead porphyrins, cheek color evenness, and cheek L*, a*, b*. Images were also analyzed for cheek red blotchiness using an in-house algorithm. The location, size, and shape of the facial region analyzed in the image were the same as that swabbed for 16S amplicon sequencing.

Biophysical Measures

Skin barrier function (transepidermal water loss) was measured in triplicate on the left cheek and forehead using the VapoMeter evaporimeter according to the manufacturer’s instructions (Delfin Technologies, Finland). The average of the three measurements was used in statistical analysis.

The relative parameter for skin elasticity R7 (Ur/Uf) was measured on the left malar eminence using the Courage + Khazaka Cutometer MPA 580 (Courage + Khazaka Electronic, Köln, Germany) equipped with a 2-mm probe. The pressure value was set to 350 mbar and one measuring cycle (10 s) was performed, with a 5-second Suction Time/On-Time and a 5-second Relaxation Time/Off-Time. Measurements were in triplicate and the average of those 3 measurements was used in statistical analysis.

Skin pH measurements were made on the left cheek (about 3–4 cm below the left corner of the eye). Measurements were made with a Mettler-Toledo Seven2Go8 Pro equipped with an “InLab Surface” flat-surface electrode (Mettler-Toledo, Columbus, OH) and calibrated twice a day with pH 4 and pH 7 standard buffer.16 Three measurements were made at each skin site and the average of those 3 measurements was used in statistical analysis.

Stratum corneum hydration on the cheek was measured with the Corneometer according to the manufacturer’s instructions. Measurements were made in triplicate and the average of those 3 measurements was used in statistical analysis.

Body mass index (BMI) was calculated from the subjects’ height and weight collected using calibrated scales.

Statistics

The comparison of mean Shannon diversity between men and women at each yearly visit was carried out using the Kruskal–Wallis One Way ANOVA on Ranks with Dunn’s method for Pairwise multiple comparison procedures using SigmaStat computer software. A Pearson product moment correlation was performed to measure the correlation between subjects’ Shannon diversity index with demographic and skin aging and health parameters.

RESULTS AND DISCUSSION

Temporal Variation in Facial Microbiome Alpha Diversity and Composition

Figure 2 shows the box plots for median alpha diversity (Shannon Diversity Index, a measure of the richness and evenness of the skin microbiome) on the cheek and forehead for men and women in 2017, 2018, and 2019. Within any given year, men showed consistently lower diversity than women on both the cheek and forehead with statistically significant differences observed for the cheek in 2017. For both men and women, forehead diversity is consistently lower than cheek diversity with statistically significant differences observed for women in 2018 and men in 2017 and 2018. Because alpha diversity is both gender and facial site dependent, we compared microbiome temporal stability within gender and site. There was no significant difference in cheek or forehead median microbiome Shannon diversity for either men or women comparing 2017–2018, 2018–2019, or 2017–2019 (ANOVA, P > 0.05).

Fig. 2.
Fig. 2.:
Box plots of median and upper and lower quartiles for forehead (above) and cheek skin (below) microbiome Shannon diversity index by gender and year. *Men mean Shannon diversity index is significantly less than the woman mean Shannon diversity index (P < 0.005).

The stability of the facial skin microbiome diversity is better visualized in Figure 3, where each subject’s forehead Shannon diversity in 2017 is plotted against their own diversity in 2018 (Fig. 3, above) and 2019 (Fig. 3, below). Although there are outliers, subjects with low or high diversity in 2017 retained their low or high diversity, respectively, in 2018 and 2019. Subjects with higher diversity tended to have lower transepidermal water loss (TEWL) (i.e., better stratum corneum barrier function—data not shown).

Fig. 3.
Fig. 3.:
Forehead Shannon diversity index in for each subject in (above) 2017 versus 2018 and (below) 2017 versus 2019. The lines are linear regressions (above) r2 = 0.49 and (below) r2 = 0.37.

Next, we examined the magnitude of the change in microbiome composition within each skin site by calculating the mean difference in Bray–Curtis pairwise dissimilarity among samples from the same subject from year 1 (2017) to year 2 (2018) to year 3 (2019). Bray–Curtis dissimilarity is a statistical method used to quantify the compositional dissimilarity between groups. Intrasubject dissimilarity was slightly higher in the cheek than on the forehead between consecutive years (Fig. 4). Over the 2-year period of the study (i.e., from 2017 to 2019), the forehead microbiome remained more stable and the cheek microbiome became more unstable, that is, the intrasubject dissimilarity increased.

Fig. 4.
Fig. 4.:
Box plot of median intrasubject Bray–Curtis pairwise distances on the cheek and forehead from 2017 to 2018, 2018 to 2019, and 2017 to 2019.

To examine the relationship between alpha diversity and temporal variability, we fitted linear models using mean Shannon index values as our metric of alpha diversity (calculated over the 3-year period) and mean Bray–Curtis dissimilarities as our metric of stability. A positive relationship was observed between forehead and cheek microbiome alpha diversity and temporal variability (Fig. 5).

Fig. 5.
Fig. 5.:
Correlation between mean Bray–Curtis dissimilarity and Mean Shannon diversity Index on the cheek and forehead.

The Skin Microbiome and Facial Skin Condition

Because microbiome alpha diversity has been proposed as a measure of human gut health, we measured the strength of the association between an individual’s skin microbiome alpha diversity and 16 parameters of skin condition across all subjects during each year of the study. Table 2 shows the Pearson product moment correlation coefficients for each pairwise comparison. Of the 16 parameters, 3 were consistently statistically significantly associated with Shannon alpha diversity every year of the study: skin barrier function as measured by TEWL, facial pore density, and follicular porphyrins. Each was negatively correlated with alpha diversity. Subjects with a leakier stratum corneum barrier, higher pore density, or more porphyrins tend to have lower alpha diversity. Because follicular porphyrins were the most strongly correlated with alpha diversity, we focused our attention on the relationship between the temporal stability of the microbiome and follicular porphyrins.

Table 2. - Pearson Product Moment Correlation Coefficients for the Association of Cheek Microbiome Shannon Diversity with Demographic and Skin Parameters*
Parameter 2017 ( n ) P 2018 ( n ) P 2019 ( n ) P
Age 0.246 (153) 0.002 0.284 (149) <0.001 0.148 (134) 0.087
BMI 0.236 (140) 0.003 0.166 (149) 0.043 nd nd
Cheek TEWL −0.313 (140) 0.0001 −0.206 (149) 0.012 −0.182 (132) 0.037
Surface pH nd nd 0.041 (149) 0.616 nd nd
Stratum corneum hydration nd nd nd nd 0.103 (130) 0.243
Elasticity (R7) 0.0966 (152) 0.236 0.113 (148) 0.170 0.052 (134) 0.552
Cheek hyperpigmentation 0.043 (155) 0.594 0.031 (149) 0.876 0.008 (134) 0.926
Brown hyperpigmentation 0.201 (155) 0.012 0.134 (149) 0.103 0.142 (134) 0.103
Pores −0.373 (155) <0.001 −0.273 (149) <0.001 −0.195 (134) 0.024
Cheek wrinkles 0.051 (155) 0.528 0.098 (149) -0.235 0.011 (134) 0.897
Cheek porphyrins −0.558 (155) <0.001 −0.531 (149) <0.001 −0.307 (134) <0.001
Color evenness −0.191 (155) 0.017 0.046 (149) 0.580 0.149 (134) 0.085
Cheek redness (a*) 0.041 (155) 0.612 −0.171 (149) 0.037 0.039 (134) 0.651
Cheek yellowness (b*) 0.069 (63) 0.356 0.097 (149) 0.241 0.122 (134) 0.160
Cheek lightness (L*) 0.069 (155) 0.880 0.141 (149) 0.086 0.049 (134) 0.574
Cheek red unevenness 0.071 (150) 0.389 0.137 (143) 0.102 0.061 (129) 0.492
nd, not done.
*Correlation coefficients with P values <0.05 are are boldface.

Temporal Stability of Cutibacterium acnes and Porphyrins

We were interested in understanding the temporal stability of the common skin commensal organism Cutibacterium acnes (formerly Propionibacterium acnes) and its association with follicular porphyrins. Figure 6 shows each subject’s cheek C. acnes relative abundance in (Fig. 6, above) 2017 versus 2018 and (Fig. 6, below) 2017 versus 2019. In general, subjects with low or high abundance of C. acnes in 2017 maintained low or high abundance, respectively, over the 2-year period. However, the decrease in the correlation coefficient and slope of the linear regression fit over the 2-year period suggests that intrasubject microbe abundance becomes more variable and declines with age.

Fig. 6.
Fig. 6.:
Temporal stability of cheek Cutibacterium acnes relative abundance for each subject in (above) 2017 versus 2018 and (below) 2017 versus 2019. The lines are linear regressions: (above) r2 = 0.64, (below) r2 = 0.45.

Cheek area follicular porphyrins were revealed by fluorescence imaging and objectively quantified using image analysis. Figure 7 shows each subject’s cheek porphyrin density (i.e., area fraction) in (Fig. 7, above) 2017 versus. 2018 and (Fig. 7, below) 2017 versus 2019. With a few exceptions, subjects with high or low levels of cheek skin porphyrins in 2017 maintained their high or low levels, respectively, in 2018 and 2019. We then looked at the intrasubject association between follicular porphyrins and C. acnes abundance. Figure 8 shows that 2017 cheek C. acnes abundance correlates strongly with 2017 cheek porphyrin density (similar results were observed for 2018 and 2019). Individuals with low C. acnes abundance show very low cheek porphyrin. However, not all subjects having high levels of C. acnes had high levels of porphyrins. A substantial fraction of the subject population showed high C. acnes abundance but very low porphyrins.

Fig. 7.
Fig. 7.:
Temporal stability of cheek porphyrins for each subject in (above) 2017 versus 2018 and (below) 2017 versus 2018. The lines are linear regressions: (above) r2 = 0.82, (below) r2 = 0.79. The red highlighted data point is subject 003 is shown in Figure 9.
Fig. 8.
Fig. 8.:
Association of cheek Cutibacterium acnes abundance with cheek porphyrins for each subject in 2017. The line is the linear regression (r2 = 0.32).

A few subjects in our study showed a dramatic change in facial skin porphyrins over the two-year period of our study. For example, Figure 9 shows white light and porphyrin-enhanced images of Subject 003 in 2017, 2018, and 2019 (this subject’s data point is highlighted in red in Fig. 7). There was a dramatic drop in porphyrins (Fig. 10, left) from 2017 to 2018 and a partial return to 2017 levels in 2019. The year-to-year change in relative abundance of C. acnes mirrored the change in porphyrins (Fig. 10, right). A personal interview was conducted with this subject to try and identify potential causes for the dramatic change in C. acnes and porphyrin levels. No obvious changes in health status, lifestyle, or other potential factors could be identified.

Fig. 9.
Fig. 9.:
Subject 003. Standard 2 white light and porphyrin images in 2017, 2018, and 2019.
Fig. 10.
Fig. 10.:
Cheek porphyrins (left) and cheek C. acnes abundance (right) for subject 003 in 2017, 2018, and 2019.

DISCUSSION

The composition of the human microbiome is highly variable depending on host and environmental factors including age, gender, body site, and lifestyle factors.1 Body site is the primary driver and at any one body site, there is tremendous variation in diversity and composition from person to person.1,4,5,17,18 However, a given body site on a single individual may also change with time. It is known that this temporal variation in the skin microbiome is relatively small relative to the variation between individuals.4 It is also known that the temporal variation in the skin microbiome is body site dependent with some sites having higher and some lower variance.5 Finally, temporal variation is highly personalize with some individual’s microbiomes being more stable than others, even at the same body site.3 The prior studies on microbiome temporal stability were limited by either short durations or small base sizes.

In this study, we characterized the temporal variation of the skin microbiome in a large population of healthy adults over a 2-year period. In addition, we used noninvasive imaging and biophysical methods to measure facial skin condition with the aim of correlating changes in the microbiome to changes in skin phenotype. We call our longitudinal study the “Cinco de Mayo Study” because skin measurements are always in early May of each year. In this way, seasonal and other environmental factors that might influence variation the skin microbiome and skin condition parameters are minimized.

Microbiome alpha diversity and composition was relatively stable from 2017 to 2018 to 2019 on both the cheek and forehead (Figs. 2 through 4) in general agreement with Oh et al.2 However, the cheek skin microbiome showed higher diversity, but lower temporal stability compared the forehead skin microbiome. We also found that individuals with higher alpha diversity on either the cheek or forehead tend to have less stable facial skin microbiomes over the 2-year period (Fig. 5). Flores et al. also found that microbiome alpha diversity on the forearm is predictive of temporal stability.3

Microbiome alpha diversity negatively correlated with TEWL, pore density, and follicular porphyrins at every time point (Table 2). These 3 skin parameters are highly interdependent and correlated with each other. Individuals with higher pore density showed higher porphyrins (r2 = 0.51, P < 0.001). A higher follicular pore density provides more sebaceous niches for fatty acid-loving anaerobes like C. acnes which is well-known to synthesize and secrete porphyrins.19 Interestingly, pore density also significantly correlates with TEWL (r2 = 0.198, P = 0.01). The surface of skin is not flat. Rather, it is covered with deep invaginations of hair follicles and sweat ducts that tremendously increases its surface area.20 This is particularly important on the face. The more pores per square centimeter, the higher the skin surface area which likely helps drive transepidermal water loss.

Across our subject population, C. acnes relative abundance and level of follicular porphyrins are relatively stable over the 2-year period (Figs. 6 and 7). Follicular porphyrins are primarily produced by C. acnes.21 However, porphyrin synthesis by C. acnes is strain dependent22 with some strains producing substantially more and some less. Indeed, Figure 8 suggests that having C. acnes is necessary, but not sufficient, for having follicular porphyrins. Acne-associated type IA-2 strains of C. acnes produce significantly higher levels of porphyrins compared with type II strains. It has been proposed that porphyrins play a role in skin inflammation and the virulence mechanisms of C. acnes in progression of acne. Coproporphyrin III, produced by C. acnes, modestly induces IL-8 production by a keratinocyte cell line in vitro suggesting that coproporphyrin III is proinflammatory.23

Although most of the subjects in our study were free of any significant acne, many subjects had very high levels of porphyrins on their entire face and yet had very healthy-looking skin. Subject 003 in Figure 10 is a good example of how changes in C. acnes abundance and corresponding level of follicular porphyrins is not associated with significance changes in visible skin inflammation. It is interesting that porphyrins never appear around the eyes, presumably because of a lack of sebum production in the periocular skin. Acne is also very rare around the eye area. The role of porphyrins in the biology of the skin bacteria and their role in skin health and disease needs further research.

C. acnes has been implicated as a cause of prosthetic joint infection24 with suggestions that the organism should be eradicated before surgery. However, the use of antibiotics to eradicate C. acnes before surgery has had limited success which underscores the resilience of the skin microbiome. The beneficial role that C. acnes plays in maintaining healthy skin microbiome harmony, for example, in suppressing the growth of methicillin-resistant Staphylococcus aureus via propionic acid,25 would also suggest caution in attempting to manage C. acnes colonization in preparation for surgery. Future work should carefully examine the robustness of the skin microbiome to presurgical skin preparation methods on the face, shoulders, and other body sites relevant to cosmetic and orthopedic surgery. In addition, more precise sampling methods should be employed to dissect the skin into its various unique microbial compartments, like the dermal sebaceous glands versus the surface stratum corneum.

The healthy cohort of subjects in this study exhibited a relatively stable skin microbiome over the 2-year period. However, it is known that this is not the case for certain skin disease states. For example, atopic dermatitis is a chronic inflammatory skin disease characterized by recurring flares of erythema, edema, scaling, and itch affecting 15%–30% of children and approximately 5% of adults in industrialized countries.26,27 More than 90% of AD patients are colonized on the lesional skin by S. aureus,28 and S. aureus colonization and biofilm formation are directly associated with disease flares and remission.8,29S. aureus resides as a normal friendly commensal in up to two-thirds of healthy individuals.30 Although the 16S primer used in our study (V1–V3) does not detect S. aureus,31 it is likely that most of the healthy subjects in our study are colonized to some degree with S. aureus and maintain colonization from year to year as was observed with C. acnes and S. epidermidis (data not shown). It is becoming clear that the transition from a healthy to disease microbiome in AD may not involve colonization with new transient microbes. Rather, disease flares likely involve expression of virulence genes in resident S. aureus.32 Said another way, the commensal skin microbes are generally stable, and beneficial, but some have a dual “Jekyll and Hyde” personality. The factors that trigger the transition and the mechanisms of bacterial virulence are attractive targets for preventing and treating skin inflammation and maintaining a stable healthy skin microbiome.33,34

Limitations

This study employed 16S ribosomal amplicon sequencing for high-level analysis of the shifts of the microbiome at the phylum and species level and does not allow for strain level resolution. This study only looked at 2 facial skin sites, the cheek and forehead. Temporal stability is body site-dependent so results at other body sites will likely differ from the results observed here. The 2-year period over which this study was conducted may not reflect longer term changes in the skin microbiome. The subjects were mostly Caucasian which minimizes any potential effect of ethnic variation on temporal stability but also limits the extrapolation of conclusions to other ethnic groups.

CONCLUSIONS

The results from this 2-year longitudinal study add to our understanding about the stability of the facial skin microbiome. Despite the rigors of daily hygiene, UV exposure, human, and animal contact, the facial skin microbiome is amazingly robust. The great majority of our facial skin microbes reside in the follicles and pores which act as personal microbial reservoirs allowing the skin surface to be repopulated quickly after cleansing or other perturbations. The Cinco de Mayo Study is an ongoing study. We plan to continue monitoring the skin microbiome in this cohort in the coming years.

PATIENT CONSENT

Patients and subjects provided written consent for the use of their images.

REFERENCES

1. Dimitriu PA, Iker B, Malik K, et al. New insights into the intrinsic and extrinsic factors that shape the human skin microbiome. mBio. 2019;10.
2. Oh J, Byrd AL, Park M, et al.; NISC Comparative Sequencing Program. Temporal Stability of the Human Skin Microbiome. Cell. 2016;165:854–866.
3. Flores GE, Caporaso JG, Henley JB, et al. Temporal variability is a personalized feature of the human microbiome. Genome Biol. 2014;15:531.
4. Costello EK, Lauber CL, Hamady M, et al. Bacterial community variation in human body habitats across space and time. Science. 2009;326:1694–1697.
5. Grice EA, Kong HH, Conlan S, et al.; NISC Comparative Sequencing Program. Topographical and temporal diversity of the human skin microbiome. Science. 2009;324:1190–1192.
6. SanMiguel AJ, Meisel JS, Horwinski J, et al. Antiseptic agents elicit short-term, personalized, and body site-specific shifts in resident skin bacterial communities. J Invest Dermatol. 2018;138:2234–2243.
7. Two AM, Nakatsuji T, Kotol PF, et al. The cutaneous microbiome and aspects of skin antimicrobial defense system resist acute treatment with topical skin cleansers. J Invest Dermatol. 2016;136:1950–1954.
8. Kong HH, Oh J, Deming C, et al.; NISC Comparative Sequence Program. Temporal shifts in the skin microbiome associated with disease flares and treatment in children with atopic dermatitis. Genome Res. 2012;22:850–859.
9. Zheng Y, Wang Q, Ma L, et al. Shifts in the skin microbiome associated with diaper dermatitis and emollient treatment amongst infants and toddlers in China. Exp Dermatol. 2019;28:1289–1297.
10. van den Munckhof EHA, Niemeyer-van der Kolk T, van der Wall H, et al. Inter- and intra-patient variability over time of lesional skin microbiota in adult patients with atopic dermatitis. Acta Derm Venereol. 2020;100:adv00018.
11. Kozich JJ, Westcott SL, Baxter NT, et al. Development of a dual-index sequencing strategy and curation pipeline for analyzing amplicon sequence data on the MiSeq Illumina sequencing platform. Appl Environ Microbiol. 2013;79:5112–5120.
12. Schloss PD. A high-throughput DNA sequence aligner for microbial ecology studies. PLoS One. 2009;4:e8230.
13. Huse SM, Welch DM, Morrison HG, et al. Ironing out the wrinkles in the rare biosphere through improved OTU clustering. Environ Microbiol. 2010;12:1889–1898.
14. Rognes T, Flouri T, Nichols B, et al. VSEARCH: a versatile open source tool for metagenomics. PeerJ. 2016;4:e2584.
15. Westcott SL, Schloss PD. OptiClust, an improved method for assigning amplicon-based sequence data to operational taxonomic units. mSphere. 2017;2:e00073-17.
16. Parra JL, Paye M; EEMCO Group. EEMCO guidance for the in vivo assessment of skin surface pH. Skin Pharmacol Appl Skin Physiol. 2003;16:188–202.
17. Findley K, Oh J, Yang J, et al.; NIH Intramural Sequencing Center Comparative Sequencing Program. Topographic diversity of fungal and bacterial communities in human skin. Nature. 2013;498:367–370.
18. Oh J, Byrd AL, Deming C, et al.; NISC Comparative Sequencing Program. Biogeography and individuality shape function in the human skin metagenome. Nature. 2014;514:59–64.
19. Barnard E, Johnson T, Ngo T, et al. Porphyrin production and regulation in cutaneous propionibacteria. mSphere. 2020;5:e00793-19.
20. Gallo RL. Human skin is the largest epithelial surface for interaction with microbes. J Invest Dermatol. 2017;137:1213–1214.
21. Pagnoni A, Kligman AM, Kollias N, et al. Digital fluorescence photography can assess the suppressive effect of benzoyl peroxide on Propionibacterium acnes. J Am Acad Dermatol. 1999;41(5 Pt 1):710–716.
22. Johnson T, Kang D, Barnard E, et al. Strain-level differences in porphyrin production and regulation in Propionibacterium acnes elucidate disease associations. mSphere. 2016;1:e00023-15.
23. Schaller M, Loewenstein M, Borelli C, et al. Induction of a chemoattractive proinflammatory cytokine response after stimulation of keratinocytes with Propionibacterium acnes and coproporphyrin III. Br J Dermatol. 2005;153:66–71.
24. Patel MS, Singh AM, Gregori P, et al. Cutibacterium acnes: a threat to shoulder surgery or an orthopedic red herring? J Shoulder Elbow Surg. 2020;29:1920–1927.
25. Wang Y, Dai A, Huang S, et al. Propionic acid and its esterified derivative suppress the growth of methicillin-resistant Staphylococcus aureus USA300. Benef Microbes. 2014;5:161–168.
26. Williams H, Flohr C. How epidemiology has challenged 3 prevailing concepts about atopic dermatitis. J Allergy Clin Immunol. 2006;118:209–213.
27. Rudikoff D, Lebwohl M. Atopic dermatitis. Lancet. 1998;351:1715–1721.
28. Leyden JJ, Marples RR, Kligman AM. Staphylococcus aureus in the lesions of atopic dermatitis. Br J Dermatol. 1974;90:525–530.
29. Di Domenico EG, Cavallo I, Bordignon V, et al. Inflammatory cytokines and biofilm production sustain Staphylococcus aureus outgrowth and persistence: a pivotal interplay in the pathogenesis of atopic dermatitis. Sci Rep. 2018;8:9573.
30. Aryee A, Edgeworth JD. Carriage, Clinical microbiology and transmission of Staphylococcus aureus. Curr Top Microbiol Immunol. 2017;409:1–19.
31. Meisel JS, Hannigan GD, Tyldsley AS, et al. Skin microbiome surveys are strongly influenced by experimental design. J Invest Dermatol. 2016;136:947–956.
32. Nakamura Y, Takahashi H, Takaya A, et al. Staphylococcus Agr virulence is critical for epidermal colonization and associates with atopic dermatitis development. Sci Transl Med. 2020;12(551):eaay4068.
33. Bhardwaj AK, Vinothkumar K, Rajpara N. Bacterial quorum sensing inhibitors: attractive alternatives for control of infectious pathogens showing multiple drug resistance. Recent Pat Antiinfect Drug Discov. 2013;8:68–83.
34. Nakagawa S, Hillebrand GG, Nunez G. Rosmarinus officinalis l. (rosemary) extracts containing carnosic acid and carnosol are potent quorum sensing inhibitors of staphylococcus aureus virulence. Antibiotics. 2020;9:149.
Copyright © 2020 by the American Society of Plastic Surgeons