In this study, we investigated the extent and specific patterns of the dysbiosis of the skin microbiome in acne patients compared to age-matched healthy individuals, using quantitative and qualitative methods. C. acnes was the most dominating species in both the healthy skin and the acne microbiome, with decreased relative abundance on acneic skin. Importantly, the C. acnes population size was 2.2-fold decreased on acneic compared to healthy skin, showing that C. acnes does not ‘overgrow’ on acneic skin.
Staphylococcus was the second most abundant genus after Cutibacterium in terms of relative abundances on healthy and acneic skin; its relative abundance significantly increased. In contrast to C. acnes, the staphylococcal population size did not decrease on acneic skin; thus, the quantitative C. acnes to staphylococci ratio dropped drastically (3-fold). On every staphylococcal cell there were about 34 C. acnes cells in the healthy cohort, but only about 11 C. acnes cells in the acne cohort. Thus, the acne microbiome is more enriched in staphylococci compared to the healthy skin microbiome.
Interestingly, the alpha diversity of staphylococci decreased on acneic compared to healthy skin. A significant decrease was determined for S. hominis and S. warneri. Previous studies showed that both S. hominis and S. warneri can protect the cutaneous barrier from opportunistic pathogens, such as S. aureus [26, 27]. In contrast, the relative abundance of S. aureus increased, with four acne patients having S. aureus as dominating staphylococcal species on cheek skin. Also, the S. epidermidis population size, the largest among skin-resident staphylococci, slightly increased in the acne cohort. The reduction of C. acnes and the relative increase of staphylococci might be a consequence of the disturbed microenvironment of the skin, with (partially) disrupted sebaceous follicles. Specific staphylococcal species or strains apparently have better colonization and survival properties on acneic skin.
The association of staphylococci with acne has not been investigated in great depth. A few studies suggested a role of S. epidermidis in acne [28–30]. One experimental study used human skin explants and cocultured them with bacterial suspensions containing different ratios of S. epidermidis and C. acnes. S. epidermidis-predominating cultures had a stronger IL-6 inducing potential than cultures in which C. acnes predominated [29]. Thus, a Staphylococcus to C. acnes ratio increase, as detected in our study, might be associated with an increased inflammatory potential of the skin microbiome. A closer look into the S. epidermidis population seems necessary, as strains can be very heterogenous, including strains with roles in maintaining skin homeostasis (e.g., by colonization resistance), and those that are opportunistic pathogens [7, 31]. Thus, the determination of the acne-associated S. epidermidis population at strain-level resolution and its comparison to healthy skin would be an interesting future perspective.
Besides the population size and relative abundance decrease of C. acnes, we found that the observed C. acnes community dysbiosis in acne patients compared to healthy individuals was characterized by (i) a decrease of C. acnes diversity, and (ii) a SLST class switch, with an increase of the relative abundances of classes A (IA1), C (IA1), E (IA1) and F (IA2) and a decrease of classes H (IB), K (II) and L (III). However, the changes were statistically significant only for the SLST classes H, K, and L. Thus, together with the quantitative data, this suggests that the C. acnes dysbiosis is mainly characterized by a loss of type IB, II and III strains. These strains seem less well equipped compared to type IA strains to survive in the altered microenvironment of acneic skin. In the future, it would be interesting to quantify the different SLST classes on acneic and healthy skin to verify this assumption. Overall, these results are largely in agreement with previous studies that reported a relative increase of IA1 strains and a relative decrease of IB and II strains in the acne microbiome [8–15]. However, direct comparisons with previous studies are difficult, since often different typing schemes were used, such as SLST, MLST and ribotyping [1, 16, 32, 33]. The SLST typing method has only been sparsely used so far, despite its relatively high-resolution power [16]. To our knowledge, no culture-independent study using the SLST amplicon scheme on acne skin samples was carried out so far. A few culture-dependent studies that used the SLST scheme found an increase of A-, C- and F-class C. acnes strains and a decrease of K-class strains [11, 12, 34].
Isotretinoin is usually a very efficient drug to treat acne [21]; it is one of the most commonly used drugs for patients with moderate to severe acne. It is well known that isotretinoin leads to reduced sebum production due to the cytotoxic effect on sebocytes [35, 36]. Altered sebum availability has likely a profound effect on the skin microbiota. Indeed, here we could identify as main effects of isotretinoin treatment: (i) a drastically changed microbiome with a strongly increased alpha diversity, and increased relative abundances of bacteria of other genera such as Streptococcus, Corynebacterium and Micrococcus; (ii) a strong quantitative reduction (36.9-fold) and a strong relative decrease of C. acnes abundance, (iii) a moderate quantitative reduction (3.6-fold), and a moderate relative increase of staphylococcal abundance, in particular characterized by an increase of S. aureus. Other studies have seen similar results regarding the reduction of C. acnes and the increased alpha-diversity upon isotretinoin treatment [37–39], albeit most of these studies did not quantitatively assess the microbiome. Here we could show that specifically C. acnes is targeted by this treatment; while C. acnes was 11.9-fold more abundant than staphylococci before treatment, staphylococci and C. acnes had similar absolute quantities after treatment. Thus, after treatment there were almost as few C. acnes left as staphylococci on patient skin. The staphylococcal community was strongly altered after treatment, with a relative decrease of S. epidermidis and an increase of S. aureus. Thus, the post-isotretinoin skin seems vulnerable to the colonization of unfavorable microorganisms, possibly due to the absence of C. acnes. Thus, a skin microbiome recovery therapy after isotretinoin treatment, e.g. by skin probiotics, might accelerate the restoration of the normal skin microbiome with its protective function.
This study has some limitations. The sample size of isotretinoin-treated acne patients (n = 14) that we could follow was relatively small due to recruitment problems. Moreover, quantifications were done for C. acnes and staphylococci, but not for the entire bacterial community. However, since C. acnes was dominating in relative abundances in the acne and healthy skin microbiomes, it is likely that the entire bacterial community was less abundant in acneic compared to healthy skin. The acne cohort in this study was heterogeneous in age, and it could not be excluded that acne patients have tried other anti-acne products with antimicrobial ingredients before they came to the dermatology clinic. We could only determine C. acnes population changes after isotretinoin treatment in a few patients, since the C. acnes abundance levels were too low upon treatment to amplify the SLST fragment. A recent study looked at such phylotype differences of C. acnes in metagenomic data on casts from facial follicles of acne patients treated with isotretinoin [40]. The authors reported that isotretinoin could selectively alter C. acnes strain diversity in SLST A and D classes.