RNase 7 participates in cutaneous innate control of Corynebacterium amycolatum

Nondiphtheria corynebacteria are typical members of the skin microbiota. However, in addition to being harmless inhabitants of healthy skin commensal skin-derived corynebacteria such as C. amycolatum occasionally also cause infections. This suggests that human skin must harbor adequate mechanisms to control the growth of corynebacteria on the skin surface. Here we show that keratinocytes are able to detect the presence of C. amycolatum leading to the epidermal growth factor receptor (EGFR)-dependent induction of the antimicrobial protein RNase 7. C. amycolatum-mediated induction of RNase 7 was also confirmed in a human 3D skin equivalent. The functional relevance of these findings was demonstrated by potent antimicrobial activity of RNase 7 against C. amycolatum and C. xerosis. In addition, the capacity of human stratum corneum to restrict the growth of C. amycolatum was significantly attenuated when RNase 7 was inactivated by a specific RNase 7-neutralizing antibody. Taken together, the interaction of RNase 7 with C. amycolatum indicates that RNase 7 may function as important effector molecule to control the growth of corynebacteria on human skin.


Induction of RNase 7 by Corynebacterium amycolatum in keratinocytes requires functional EGFR.
Several reports indicate that the EGFR is critically involved in the induction of RNase 7 by microorganisms 13,14,16 . Therefore we sought to determine whether the induction of RNase 7 by C. amycolatum was also dependent on the EGFR. To this end we incubated the keratinocytes with the selective EGFR inhibitor AG-1478. As shown in Fig. 2a,b AG-1478 diminished the induction of RNase7 gene expression and protein secretion in keratinocytes treated with C. amycolatum. In line with these experiments the use of EGFR-blocking antibody cetuximab also significantly decreased the C. amycolatum-mediated RNase7 gene and protein expression in  C. amycolatum-induced RNase 7 expression in keratinocytes is mediated by the epidermal growth factor receptor (EGFR). Human primary keratinocytes were treated for 24 h with living C. amycolatum with or without the selective EGFR inhibitor AG-1478 (10 µM). (a) Relative RNase7 gene expression was analyzed by real-time PCR and (b) RNase 7 protein secretion was measured by analysis of the supernatant using an RNase 7 specific ELISA. In a second experimental setup stimulation of the keratinocytes with living C. amycolatum was done for 24 h in the absence or presence of the EGFR blocking antibody cetuximab (20 µg/ml). (c) Relative RNase7 gene expression was analyzed by real-time PCR and (d) RNase 7 protein secretion was measured by analysis of the supernatant using an RNase 7-specific ELISA. Shown are means ± s.e.m. of nine separate stimulations (*p < 0.05; **p < 0.01, one-way analysis of variance ANOVA using Tukey's multiple comparison test). (e) Analysis of RNase 7 expression by immunostaining in a 3D skin equivalent. The 3D skin equivalent was left unstimulated or stimulated with living C. amycolatum for 24 h in the presence or absence of cetuximab (20 µg/ml). Bars represent 50 µM.
primary keratinocytes (Fig. 2c,d). Immunohistochemistry analysis of a 3D skin equivalent stimulated with living C. amycolatum revealed that the C. amycolatum-mediated RNase 7 induction was inhibited by treatment with cetuximab ( Fig. 2e).

RNase 7 exhibits antimicrobial activity against C. amycolatum and C. xerosis.
To determine whether RNase 7 is able to restrict the growth of corynebacteria we incubated C. amycolatum and C. xerosis with different concentrations of RNase 7 in a microdilution assay for 3 h. As shown in Fig. 3a,b RNase 7 dose-dependently inhibited the growth of C. amycolatum and C. xerosis. Concentrations lower than 1 µg/ml RNase 7 still inhibited the growth of the bacteria. To investigate whether the enzymatic activity of RNase 7 is necessary for the observed killing activity we used a mutated recombinant RNase 7 without ribonuclease activity. This ribonuclease-inactive RNase 7 variant showed similar activity against C. amycolatum as compared to recombinant wild-type RNase 7 (Fig. 3c) indicating that the ribonuclease activity of RNase 7 is not crucial for its antibacterial activity against C. amycolatum. RNase 7 contributes to the antimicrobial activity of human stratum corneum against C. amycolatum. To analyze the functional importance of RNase 7 in cutaneous defense against corynebacteria we used an RNase 7 blocking antibody which neutralized the antimicrobial activity of RNase 7 against C. amycolatum whereas an irrelevant control antibody had no influence (Fig. 4a). We then incubated a stratum corneum extract with C. amycolatum in the presence of the RNase 7 blocking antibody or in the presence of the irrelevant antibody. The stratum corneum extract was able to control the growth of C. amycolatum also in the presence of the irrelevant control antibody. In contrast, inactivation of RNase 7 in the stratum corneum extract by incubation with the RNase 7 blocking antibody led to an outgrowth of C. amycolatum (Fig. 4b). These data show that RNase 7 contributes to the capacity of human stratum corneum to control the growth of C. amycolatum.

Discussion
RNase 7 is an antimicrobial protein which is abundantly expressed in keratinocytes and characterized by a broad spectrum of antimicrobial activity [12][13][14] . This suggests that RNase 7 may play a major role in cutaneous defense. This is in concordance with recent functional studies reporting that RNase 7 contributes to the capacity of the human skin surface to inhibit the growth of S. aureus, E. faecium and P. aeruginosa 13,14,17 . In line with these observations it has been reported that high levels of RNase 7 may confer protection against S. aureus infection of the skin 18 . Since there is increasing evidence that AMP may also play an important role to control and shape the commensal microbiota we sought to determine whether RNase 7 may help to control cutaneous growth of corynebacteria, inhabitants of healthy normal skin. Our data show that RNase 7 exhibits potent antimicrobial activity against C. amycolatum and C. xerosis. The use of a recombinant ribonuclease-deficient RNase 7 mutant revealed that the enzymatic activity of RNase 7 was dispensable for its antimicrobial activity against C. amycolatum. This is in line with our previous study documenting that the activity of RNase 7 against the gram-positive bacterium Enterococcus faecium required no ribonuclease activity 13 .
Our data revealed that concentrations lower than 1 µg/ml RNase 7 already restricted the growth of the corynebacteria suggesting that RNase 7 may play an important role to control the growth of corynebacteria on the skin surface. In support of this hypothesis our results show that the capacity of stratum corneum to inhibit the growth of C. amycolatum was reduced when RNase 7 was inactivated by a blocking antibody.
This study demonstrates for the first time that keratinocytes are able to sense the presence of corynebacteria leading to an increased expression of RNase 7. The induction of RNase 7 by C. amycolatum in keratinocytes required the EGFR. This is in line with other reports demonstrating a crucial role of the EGFR for the microbial induced expression of RNase 7 in keratinocytes 14,[19][20][21] . It remains to be investigated if the bacteria directly activate the EGFR-pathway as it has been reported for the S. aureus protein A-mediated activation of the EGFR 22 or if the activation of the EGFR follows sensing by a pattern recognition receptor. In this regard a recent report demonstrated the upregulation of Toll-like receptor expression in corneal epithelial cells infected with Corynebacterium pseudodiphtheriticum 23 . This suggests that Toll-like receptors may be involved in the detection of corynebacteria in epithelial cells, a hypothesis which remains to be addressed in further studies.
As mentioned above corynebacteria are abundant commensals on the skin surface. Another abundant skin commensal is Staphylococcus epidermidis. Wanke et al. reported that S. epidermidis induced the expression of RNase 7 in keratinocytes and this induction required functional EGFR 21 . One may speculate that a permanent induction of RNase 7 by skin commensals such as coagulase-negative staphylococci and corynebacteria may increase cutaneous defense by providing constant antimicrobial activity on the skin surface. A failure to adequately induce RNase 7 may contribute to the increased cutaneous infection risk of cancer patients receiving anti-EGFR therapy 24 .
Taken together, our data indicate a novel role of RNase 7 to control the growth of corynebacteria on human skin. It remains to be shown whether a failure to adequately control the growth of C. amycolatum and other cutaneous corynebacteria may be associated with a higher risk for infections caused by these bacteria. In addition, it is an intriguing hypothesis that variations in the expression of corynebacteria-controlling AMP such as RNase 7 may influence body odor formation.
Organotypic 3D skin equivalent. The organotypic 3D skin equivalent was generated as previously described 14 . The 3D skin equivalent was stimulated with 20 µl C. amycolatum diluted in KGM2 (OD 600 of 0.2) as described above. In order to block the EGFR 20 µg/ml cetuximab was added to the culture medium in the external wells 45 min before stimulation. After stimulation for 24 h two biopsies were taken from each 3D skin equivalent using a 6 mm biopsy punch. One biopsy was embedded in paraffin for immunohistochemical analysis and the other biopsy was used for RNA isolation. The KGM2 medium in the external wells was harvested for ELISA.
RNA isolation and cDNA synthesis. Total RNA of the keratinocytes and the 3D skin equivalent was isolated using 500 µl of the RNA isolation reagent Crystal RNAmagic according to the manufacturer's protocol (BiolabProducts, Gödenstorf, Germany). The isolated RNA was dissolved in H 2 O and 0.5 µg total RNA was reversed transcribed to cDNA using an oligo (dT)18 primer and 50 Units Maxima Reverse Transcriptase (Thermo Fisher Scientific, Waltham, MA) according to the supplier's protocol.

Real-time PCR analysis. Quantitative real-time PCR was done in a StepOnePlus Real Time PCR
System (Applied Biosystem, Carlsbad, CA) as previously described 25 using SYBR Premix Ex Taq II (TaKaRa Bio, Saint-Germain-en-Laye, France) and cDNA corresponding to 10 ng total RNA as template. The following intron-spanning primers were used: RNase 7: 5′-GGA GTC ACA GCA CGA AGA CCA-3′ (forward primer) and 5′-CAT GGC TGA GTT GCA TGC TTG A-3′ (reverse primer) and the house keeping gene RPL38 (ribosomal protein L38): 5′-TCA AGG ACT TCC TGC TCA CA-3′ (forward primer) and 5′-AAA GGT ATC TGC TGC ATC GAA-3′ (reverse primer). Standard curves were generated for each primer pair using serial dilutions of template cDNA. Relative RNase7 gene expression is given as a ratio between expression of RNase7 and RPL38 gene expression.
ELISA. Secreted RNase 7 protein levels in the cell culture supernatants were measured by a specific RNase 7 ELISA as previously described 13 . The detection range of the RNase 7 ELISA was between 0.3 ng/ml and 40.0 ng/ml.
Immunostaining. The organotypic 3D skin equivalent was embedded in paraffin and immunostaining was performed as described 26 . Briefly, a self-generated goat anti-RNase 7 antibody 13 was used followed by biotinylated rabbit anti-goat IgG antibody (DakoCytomation, Glostrup, Denmark) and avidin/biotinylated enzyme complex (Vectastain ABC-AP staining-kit, Vector laboratories, Peterborough, UK) and a red alkaline phosphatase substrate (Red AP; Vector Red, Vector laboratorie, Burlingame, Ca). Slides were counter stained with hematoxylin and mounted with Eukitt (Poly(butyl methacrylate-co-methyl methacrylate); O. Kindler, Freiburg, Germany). The use of pre-immune serum instead of an antibody served as negative control.
Antimicrobial Assay. Corynebacteria were grown in BHI medium overnight at 37 °C until reaching an OD 600 of 0.2. This culture was diluted 1:1000 in 10 mM sodium phosphate buffer containing 2% BHI medium and 0.1% BSA (bovine serum albumin, Sigma-Aldrich). 25 µl of this bacteria solution was mixed with 25 µl 10 mM sodium phosphate buffer containing different concentrations of recombinant RNase 7 or stratum corneum extract prepared as previously described 13 . The use of human stratum corneum derived from the heel was approved by the Ethics Committee at the Medical Faculty of the Christian-Albrechts-University, Kiel, Germany (A104/06) in concordance with the Declaration of Helsinki protocols and donors have given informed consent. Incubation was carried out at 37 °C for 3 h followed by plating serial dilutions on BHI agar plates to analyze colony forming units (CFU) after overnight incubation at 37 °C. In some experiments incubation was performed in the presence of a specific RNase 7 blocking antibody (0.5-0.1 mg/ml) or an irrelevant antibody as described 13,14 .
To investigate whether the ribonuclease activity of RNase 7 is responsible for its antibacterial activity against C. amycolatum we used a recombinant mutant RNase 7 without ribonuclease activity as previously described 13 . Data Availability. The datasets generated during the current study are available from the corresponding author on reasonable request.