Synergistic effect of immunomodulatory S100A8/A9 and ciprofloxacin against Pseudomonas aeruginosa biofilm in a murine chronic wound model

ABSTRACT

The majority of chronic wounds are associated with bacterial biofilms recalcitrant to antibiotics and host responses. Immunomodulatory S100A8/A9 is suppressed in Pseudomonas aeruginosa biofilm infected wounds. We aimed at investigating a possible additive effect between S100A8/A9 and ciprofloxacin against biofilms. Materials/methods: Thirty-two mice were injected with alginate-embedded P. aeruginosa following a third-degree burn. The mice were randomized into four groups receiving combination ciprofloxacin and S100A8/A9 or monotherapy ciprofloxacin, S100A8/A9 or a placebo and evaluated by host responses and quantitative bacteriology in wounds. In addition, in vitro checkerboard analysis was performed, with P. aeruginosa and ascending S100A8/A9 and ciprofloxacin concentrations. Results: S100A8/A9 augmented the effect of ciprofloxacin in vivo by lowering the bacterial quantity compared to the placebo arm and the two monointervention groups (P < 0.0001). S100A8 and 100A9 were increased in the double-treated group as compared to the monointervention groups (P = 0.032, P = 0.0023). Tissue inhibitor of metalloproteinases-1 and keratinocyte\chemokine chemoattractant-1 were increased in the double-intervention group compared to the S100A8/A9 group (P = 0.050, P = 0.050). No in vitro synergism was detected. Conclusion: The observed ciprofloxacin-augmenting effect of S100A8/A9 in vivo was not confirmed by checkerboard analysis, indicating dependence on host cells for the S100A8/A9 effect. S100A8/A9 and ciprofloxacin is a promising therapy for optimizing chronic wound treatment.

biofilm, flouroquinolone, Pseudomonas

INTRODUCTION

Chronic wounds have a multifactorial etiology, regarding both the initial genesis and the persistence of the wound. This is based on a complex interaction between several factors, including the host response, metabolism, comorbidities, pharmaceutical treatments, malignancy, age (Jockenhöfer et al. 2016; Moser et al. 2017) and phenotypic state (biofilm or planktonic) of one or several invading microorganisms (Percival, McCarty and Lipsky 2015). Pseudomonas aeruginosa is a prominent opportunistic pathogen, a notorious biofilm former and the sixth most common cause of hospital-acquired infections (Weiner et al. 2016). Biofilm formation results in non-mutational-based tolerance to host responses and antibiotic therapy compared to planktonic growth (Høiby et al. 2010) due to e.g. slow growth, upregulated beta-lactamase production, downregulation of outer membrane proteins (Pai et al. 2001), development of effective multidrug efflux pumps (Aires et al. 1999; Köhler et al. 1999; De Kievit et al. 2001; Hooper 2001; Hocquet et al. 2003; Sobel, McKay and Poole 2003), binding of antibiotics in the matrix and the possible transfer of a methylase gene from wound co-pathogens (Madsen et al. 2012; Cao et al. 2015; Cao et al. 2016). Biofilm infections are also a ‘hot spot’ for the development of traditional antibiotic resistance due to the increased mutation rate compared to planktonic growing bacteria and the prolonged selection pressure of ineffective antibiotic therapies further complicating treatment (Nesse and Simm 2018). Therapeutic regimens to prevent this development are pivotal.

In biofilm-infected wounds, the healing is halted in the inflammatory phase of healing, with some overlapping into the proliferative state. The influx of polymorphonuclear cells (PMNs) and peripheral monocytes transforming into tissue macrophages in the inflammatory phase help to remove debris and prevent a settlement of the bacteria in the wound bed by phagocytosis. The phagocytosis ability of these cells is decreased and cellular senescence in fibroblasts is induced, simultaneously with a transition to an inflammatory dormant and proteolytic environment in the chronic wound bed (Parmely et al. 1990). The matrix metalloproteinases (MMPs) and neutrophil elastases are proposed to be responsible for the proteolytic environment in which growth factors are degraded (Yager et al. 1997; Lazaro et al. 2016).

The responsible interaction between host and pathogen is sparsely understood, since the subject has not received the attention it deserves and is often neglected in many clinical and experimental studies. This is despite more than 3% of the total healthcare expenditure in developed countries being spent on the treatment and care of these chronic wounds (Posnett and Franks 2008), not to mention the enormous impact on quality of life in this patient group (Situm, Kolić and Spoljar 2016).

Recent clinical observations of chronic venous ulcers and animal experimental findings of a suppressive effect of P. aeruginosa biofilms on host cell markers like S100A8/A9 (also named calprotectin), and the granulocyte colony stimulating factor (G-CSF) (Trøstrup et al. 2017b), have led us towards a better understanding of this host–pathogen interaction, and indicated possible adjunctive treatment options. Several chronic wound treatment regimens have been investigated, including wound debridement, topical therapy, including nano-therapeutics, antiseptics and antimicrobial agents, different wound dressings, wound coverage and additional adjunctive therapies, optimizing the control of comorbidities. Different topical substitutions of growth factors, such as platelet-derived growth factor (PDGF), human epidermal growth factor and G-CSF (Falanga et al. 1992; Marques da Costa et al. 1997; Wieman, Smiell and Su 1998; Da Costa et al. 1999) are also available on the market. The main limitation is that these compounds are partly or completely degraded before they reach the wound bed and thus possess a doubtful healing potential. More recently, the topical use of immune-modulating S100A8/A9 has been prophesized as holding a great potential in chronical wound care (Trøstrup et al. 2017a). This physiologically active heterodimer, comprising calgranulin A and B (S100A8/A9), is located in the cytosol of neutrophils and monocytes and is released in the phagocytosis process (Rammes et al. 1997; Roth et al. 2003). It consists of two subunits with different immunologic modes of action (Vogl et al. 2007). The heterodimer has been shown to stimulate the differentiation of keratinocyte by inhibiting the activity of telomerase and displays immune stimulatory, with neutrophil chemotaxis (Ryckman et al. 2003) and apoptosis (Ghavami et al. 2004), and antimicrobial properties both in vitro and in vivo in murine models (Trøstrup et al. 2017a). S100A8/A9 is released from the neutrophil cytoplasm in a normal inflammatory response in active wounds (Thorey et al. 2001) but is found to be decreased in chronic wounds (Trøstrup et al. 2011). When administered, it stimulates the host response (Trøstrup et al. 2017a). Based on recent findings that S100A8/A9 monotherapy has an effect on wound P. aeruginosa reduction after 5 days of treatment (Trøstrup et al. 2017a), we speculated whether a synergistic effect could be obtained by combining systemic ciprofloxacin with topical S100A8/A9 treatment.

MATERIALS AND METHODS

Ethics

The following description is approved by the Animal Ethics Committee of Denmark, and all animal handling was at all times performed in accordance with the National and European guidelines (2015–15-0201–00618).

In vivo study design

Forty BALB/c pathogen-free 12-week female mice (Janvier, CS 4105 Le Genest Saint Isle, Saint Berthevin, Cedex 53 941 France) were acclimatized in the laboratory (BRIC—Biotech Research & Innovation Centre, University of Copenhagen, Copenhagen, Denmark) one week prior to the experiment launch. Full-thickness necrosis was inflicted on 32 mice in a 1.7 × 1.7 cm (2.9 cm² 119) caudal area. The third-degree burns were performed in a reproducible and standardized method by hot air for 7 seconds (Calum, Høiby and Moser 2014). Analgesia with subcutaneous (s.c.) injections of buprenorphine (100 µL, 30 µg/mL) every eighth hour was administered 24 hours post burn procedure.

On the fourth day, post thermal damage, 100 µL of a 107 colony forming unit (CFU)/mL P. aeruginosa (strain PAO1, Iglewski strain) solution was injected under the eschar to create a monopathogenic biofilm infection. The bacteria were embedded in seaweed alginate, as previously described, to simulate biofilm embedment (Christophersen et al. 2012). The biofilm settlement was accomplished in 24 hours.

The 32 mice were randomized into four groups and initiated into treatment once daily 24 hours post alginate injection: Group (1) ciprofloxacin and S100A8/A9 treated; Group (2) ciprofloxacin monotherapy; Group (3) S100A8/A9 monotherapy; and Group (4) saline controls.

Intervention groups receiving ciprofloxacin were treated with a dose of 500 µL (2 mg/mL) ciprofloxacin (Fresenius Kabi, Islands Brygge 57, 2300 KBH S, Denmark) injected subcutaneously on the abdomen. Mice receiving recombinant murine S100A8/A9 (Cloud Clone, Katy, TX, USA) were administered 250 µL (4 µg/mL) subeschar injections. A fifth group of six mice was used as non-infected burned controls (sham controls) receiving only saline. Lastly, two unharmed healthy mice were kept as background controls. CFUs were counted manually and the result was adjusted for the dilution factor in each homogenate.

Cytokines and growth factors

The impact of cytokines and growth factor on wound tissue was measured by the LUMINEX® 200TM platform (Luminex Corp., Austin, TX, USA). The bead-based multiplex assays (Biotechne R&D Systems, 614 McKinley Place NE, Minneapolis, MN, USA) included measurements of PDGF subunit B (PDGF-BB), tissue inhibitor of metalloproteinases-1 (TIMP-1), vascular endothelial growth factor (VEGF), chemokine/keratinocyte chemoattractant (CXCL1/KC), G-CSF, interleukins (IL-1β, IL-10, IL-17α) and S100 calcium-binding proteins (S100A8 and S100A9).

Digital photoplanimetry

All 40 mice were photographed on Days 1, 3 and 4 post-infections. In order to standardize this procedure with no potential skin traction, the mice were temporarily sedated in an isoflurane inhalation box while photographed. Digital image analysis was performed using Image J (Bethesda, MD, USA).

In vitro study design

The synergy between ciprofloxacin and S100A8/A9 in P. aeruginosa eradication was assessed by the checkerboard method. Microtiter trays with cation-supplemented Mueller-Hinton broth were prepared for checkerboard analysis. Each microtiter well was inoculated with 100 µL of a 107 CFU P. aeruginosa (strain PAO1, Iglewski strain). Ascending concentrations were placed in the wells. For ciprofloxacin (Fresenius Kabi, Islands Brygge 57, 2300 KBH S, Denmark) and recombinant murine S100A8/A9 (Cloud Clone, Katy, TX, USA), the concentrations were 0.0, 0.06, 0.12, 0.25, 0.5, 1.0, 2.0 mg/L and 0.0, 3.1, 6.25, 12.5, 25.0, 50.0 µg/mL, respectively.

Each well held a unique concentration combination of the two substrates. The tubes were incubated overnight and minimal inhibitory concentrations (MICs) were registered. A 4-fold MIC reduction compared to the monotherapies, ciprofloxacin or S100A8/A9, was considered a synergism indicator.

Statistics

Non-parametric statistical methods were employed when D'agostino-Pearson omnibus analysis did not identify a Gaussian distribution. These included the Mann–Whitney test for comparing groups individually and the Kruskal–Wallis by ranks for multiple comparisons. Statistical analyses were performed directly on the fluorescence index (mean of duplicates), subtracting the fluorescence background level on all samples, to avoid any faulty computerized estimated values. Wound tissue densities (mean ± standard deviation) were calculated as log10CFU/g wound.

A level of P ≤ 0.05 was considered significant. Statistics were performed employing Microsoft Excel (Microsoft, version 15.41, WA, USA) and GraphPad Prism (version 7.02, CA, USA). The results from the checkerboard analysis were analyzed by the fractional inhibitory concentration (FIC). An FIC ≤ 0.5 indicated synergism between the two substrates. An FIC between 0.5 and 1.0 was considered to be non-synergistic or non-additive. An FIC value from 1 to 4 was defined as indifferent, while >4 was considered an antagonistic regime.

RESULTS

Quantitative bacteriology of P. aeruginosa biofilm infected wounds

Analyses of the bacterial wound densities showed a significantly reduced bacterial load in the group receiving a combination of S100A8/A9 and ciprofloxacin compared to the saline-treated control group (P = 0.0003). Mean values (log10(CFU/g) ± standard deviation) were 7.46 ± 0.95 and 9.27 ± 0.51, respectively. A reduction in bacterial load, although less pronounced, was found in the group receiving ciprofloxacin monotherapy (8.31 ± 0.23) compared to saline-treated controls (P = 0.002). However, S100A8/A9 treatment alone was non-significant compared to the untreated controls (P = 0.66) (Fig. 1 and Table 1). Interestingly, there was a significant reduction in CFU in the group treated with combined ciprofloxacin and S100A8/A9 compared to the ciprofloxacin monotherapy group (P = 0.0027).

Figure 1.

Quantitative bacteriology of P. aeruginosa biofilm wounds (log10CFU/g wound 4 days post-infection).

A significantly lower CFU was found in the group treated with both ciprofloxacin and S100A8/A9 compared separately to the individual monotreatment groups and the comparator arm (P = 0.0003, P = 0.027 and P = 0.0003). Mean ± SD was respectively 7.460 ± 0.9501, 9.171 ± 0.3785, 8.1310 ± 0.2260 and 9.272 ± 0.5089 with a significant difference between all four groups found with ANOVA, P < 0.0001.

Table 1.

Overview of significant P-values in the analyses of quantitative bacteriology (in CFU per wound) and signaling proteins. A level of P < 0.05 was considered significant. Abbreviation: ns., non significant.

		Quantitative bacteriology	S100A8	S100A9	IL-1β	IL-10	KC	TIMP-1	VEGF	G-CSF
Multiple comparison		<0.0001	0.033	0.0023	0.0009	ns.	0.047	ns.	ns.	ns.
Unpaired two-sample comparison
Groups being compared
Ciprofloxacin + S100A8/A9
	S100A8/A9	0.0003	0.011	0.001	0.0015	ns.	0.05	0.05	ns.	ns.
	Ciprofloxacin	0.0027	ns.	ns.	ns.	ns.	ns.	ns.	ns.	ns.
	Biofilm controls	0.0003	0.002	0.0004	ns.	0.0011	0.037	0.0037	ns.	ns.
	Sham controls	Not relevant	0.019	ns.	0.023	0.025	ns.	ns.	ns.	<0.0001
S100A8/A9	Ciprofloxacin	<0.0001	ns.	0.045	0.008	0.02	ns.	ns.	ns.	ns.
	Biofilm controls	ns.	ns.	ns.	ns.	0.013	ns.	0.018	ns.	ns.
	Sham controls	Not relevant	ns.	0.011	0.0002	ns.	ns.	ns.	ns.	<0.0001
Ciprofloxacin	Biofilm controls	0.002	0.05	0.013	ns.	0.0005	ns.	ns.	ns.	ns.
	Sham controls	Not relevant	ns.	ns.	0.0043	0.007	ns.	ns.	0.033	0.0009
Biofilm Controls	Sham controls	Not relevant	ns.	0.0048	0.023	0.011	ns.	ns.	0.03	0.0069

The impact of ciprofloxacin and S100A8/A9 treatment on S100A8 and S100A9 levels in wounds

Multiple comparisons showed all three intervention groups to have significantly different levels of both S100A8 (P = 0.033) and S100A9 (P = 0.0023). The levels of S100A8 and S100A9 were highest in the wounds from the group treated with both ciprofloxacin and S100A8/A9 followed by the level in the ciprofloxacin monotherapy group.

When the groups were analyzed in an unpaired two-sample manner, the double-treated group was significantly different in both S100A8 and S100A9 levels as compared to the group treated with S100A8/A9 alone (P = 0.011 and P = 0.0010) and to the biofilm saline control group (P = 0.0020 and P = 0.00040). S100A9 levels were also significantly higher in the ciprofloxacin-treated group compared to the group that received only S100A8/A9 subcutaneously (P = 0.045). This difference was not identified for S100A8.

Lastly, higher levels of both S100A8 and S100A9 were identified in the monotherapy ciprofloxacin group versus the biofilm controls (P = 0.050 and P = 0.013) (Fig. 2, Table 1).

Figure 2.

Level of four PMN-related cytokines (rg) in wound (g) homogenate supernatant 4 days post-infection: S100A8, S100A9, KC and G-CSF. The levels of both S100A8 (A) and S100A9 (B) were different in all three intervention groups in the two multivariate analyses (P = 0.033 and P = 0.0023). The S100A8 and S100A9 levels were highest in the double-treated group compared individually to the S100 monotherapy group (P = 0.011 and P = 0.001). S100A8 and S100A9 were also higher in the ciprofloxacin-treated group compared to the comparator arm (P = 0.05 and P = 0.013). This difference in S100A8 level was not significantly different between the S100A8/A9 and ciprofloxacin monotherapy groups. The double-treated group yielded significantly higher levels of KC (C) compared to the group receiving only A100A8/A9 and saline (P = 0.046) and the control group (P = 0.037). When all three intervention groups were analyzed together in the multivariate analysis, a difference in KC level was detected. G-CSF levels (D) were raised in all biofilm-inflicted groups compared to the no-biofilm baseline group (respectively P < 0.0001, P < 0.0001, P = 0.0009 and P = 0.0069). No further statistically significant differences were detected between the different groups.

Cytokines

The levels of IL-1b were different when all three intervention groups were analyzed with an ANOVA analysis (P = 0.0009). We found the level of IL-1b to be significantly higher in the group receiving S100A8/A9 as compared to the ciprofloxacin group (P = 0.008) and the double-treated group (P = 0.0015) (Fig. 3).

Figure 3.

Level of pro-inflammatory signaling proteins (rg) in wound (g) homogenate supernatant 4 days post-infection. The levels of IL-1b (E) were significantly different in all three interventions in an ANOVA (P = 0.0009). The level of IL-1b was highest in the group receiving S100A8/A9 compared to the ciprofloxacin group (P = 0.008) and the double-treated group (P = 0.0015). The level of VEGF (F) was lower in the ciprofloxacin group compared to the biofilm controls (P = 0.033) and the no-biofilm control group (P = 0.030). No other comparison was found to be significant.

The levels of IL-10 were not significantly different in the three intervention groups (P = 0.055). IL-10 was found to be at a higher concentration in the group treated with only ciprofloxacin as compared to the comparator arm and the S100A8/A9 group (P = 0.0005 and P = 0.02). No difference was detected when the combination treatment was compared to the S100A8/A9 monotherapy group (Fig. 4). We found that the group of ciprofloxacin and S100A8/A9 yielded significantly higher levels of KC in wound supernatants compared to the group receiving only S100A8/A9 and saline (P = 0.05) and the biofilm control group (P = 0.037). When all three intervention groups were analyzed together, a borderline difference in KC level was detected between the groups (P = 0.05) (Fig. 2, Table 1).

Figure 4.

Level of anti-inflammatory signaling proteins (rg) in wound (g) homogenate supernatant 4 days post-infection. The level of TIMP-1 (G) in the group receiving both ciprofloxacin and S100A8/A9 was higher than in the group treated with S100A8/A9 exclusively (P = 0.05). Both levels were higher than that of the control group (P = 0.0037 and P = 0.018). No differences in TIMP-1 or in the IL-10 levels were seen in the multivariate analyses of the three intervention groups (P = 0.078 and P = 0.055). IL-10 was found to be at a higher concentration in the group treated with only ciprofloxacin as compared to the comparator arm (P = 0.0005) and the S100A8/A9 group (P = 0.02). No further differences were detected.

Glycoproteins

G-CSF levels were elevated in all P. aeruginosa biofilm inflicted groups compared to the sham control group, P < 0.0001, P < 0.0001, P = 0.0009 and P = 0.0069, respectively. No significant differences were seen between intervention groups (Fig. 2).

The level of VEGF was found to be lower in the ciprofloxacin group compared to the control group without a P. aeruginosa biofilm (P = 0.033) (Fig. 3).

A significant association between the level of TIMP-1 and treatment with both ciprofloxacin and S100A8/A9 compared to the group treated with S100A8/A9 exclusively (P = 0.05) was observed. Both the double-treated group and the group receiving S100A8/A9 had higher levels of TIMP-1 than the biofilm control group (P = 0.0037 and P = 0.018). No differences were seen in the multivariate analysis (P = 0.078) (Fig. 4, Table 1).

Anti-inflammatory/pro-inflammatory balance

The TIMP-1/IL-1b ratio was proven to be significantly different when comparing all three intervention groups in a multivariate analysis (P = 0.015). The ratio of the double-treated group was different compared to that of the S100A8/A9 monotherapy group and of the control group (P = 0.0012 and P = 0.013). In contrast, the ciprofloxacin monotherapy group was not significantly different from the double intervention group or the control group (Fig. 5).

Figure 5.

A difference in TIMP-1/IL-1b ratio was identified between the three groups receiving treatment (P = 0.0015). The group receiving ciprofloxacin and S100A8/A9 was the only group resulting in an increased anti-/pro-inflammatory ratio compared to the biofilm control group (P = 0.013).

Digital planimetry

Digital planimetry analysis did not show any macroscopic differences in healing response when looking at the necrotic area and wound size between the groups.

In vitro checkerboard analysis

The in vitro checkerboard analysis indicated a tendency towards a ciprofloxacin potentiating effect of S100A8/A9. However, FIC values below 0.5 were not obtained, and, therefore, in vitro synergism was not revealed (FIC values from 0.55 to 2.05).

DISCUSSION

The interplay between host and pathogen in chronic wounds is highly complex, with numerous disposing factors and etiologies (Gurtner et al. 2008). Acknowledgment of the impact of bacterial biofilm and host response interactions on chronic wounds is important. Before successful wound healing and tissue remodeling can take place, resolution of this interplay is crucial.

Chronic wounds are arrested in the inflammatory phase of normal healing, possibly due to chronic biofilm infections (Bjarnsholt et al. 2008). In order to push the chronic wounds out of the inflammatory phase into the proliferative and remodeling phases, there is a need for compounds that can transform the chronic wound into an active healing wound. Since antibiotic therapy alone is generally accepted as an insufficient therapy and induces antibiotic resistance, adjunctive therapies improving effects on the wound biofilms and on wound healing are needed.

The present study demonstrates such a potential candidate due to the observed synergistic effect between a fluoroquinolone, in the form of ciprofloxacin, and immunomodulation therapy with S100A8/A9 against biofilm-growing P. aeruginosa in a chronic murine wound model. The experiment demonstrated that the application of the immunomodulatory protein S100A8/A9 directly in wounds combined with antibiotic therapy lowered the bacterial load in the chronic wounds significantly. The non-significant CFU reduction observed in the monotherapy with the topical S100A8/A9 group was in accordance with previous observations on Day 3 post-infection and is probably due to the short observation period or even an insufficient dosage of S100A8/A9 for this bacterial biofilm infection.

Our findings confirm that the P. aeruginosa biofilm suppresses the level of the neutrophil marker S100A8/A9, but the mechanism is not known. This effect is observed despite the high protease resistance (Nacken and Kerkhoff 2007), the inhibitory effect of S100A8/A9 (Trøstrup et al. 2017b) and the possible dissolving effects of S100A8/A9 on the P. aeruginosa biofilm (Akerström and Björck 2009).

When administering only ciprofloxacin, the levels of S100A8/A9 in the wound homogenates were higher than when treating with S100A8/A9 alone. This can be explained by ciprofloxacin penetrating and killing a fraction of the biofilm-growing P. aeruginosa, thus limiting the biofilm-mediated suppression of neutrophils and their secreted S100A8/A9. When double ciprofloxacin and S100A8/A9 therapy was administered, the levels of the individual S100A8 and S100A9 were further increased, as a surrogate marker of activated phagocytizing cells and neutrophil chemotaxis promoting the healing process and moving forward in the healing cascade and thus stimulating keratinocyte proliferation. This result further advocates for the beneficial effect of supplementing the antibiotic therapy with factors potentially improving healing, like the otherwise suppressed S100A8/A9. Whether the additional antibiotic effect of S100A8/A9 lowers the risk of antibiotic resistance development was not evaluated in the present study but will be the topic of further studies.

The lack of demonstrating a convincing in vitro synergistic effect between S100A8/A9 and ciprofloxacin in the checkerboard analysis is supported by the above-stated observations, and strongly indicates that the synergistic ciprofloxacin potentiating effect of S100A8/A9 is highly dependent on host cells and tissues. Planktonic growing bacteria were used for this analysis. If in vitro synergy was observed, we would have included biofilm growing bacteria for this analysis.

In addition, these findings support the multifaceted roles of S100A8/A9. The direct anti-Pseudomonas effect of S100A8/A9 demonstrated by our group was obtained by high S100A8/A9 concentrations and in another experimental set-up (Trøstrup et al. 2017a).

The levels of G-CSF and IL-1β were both higher in the comparator arm inflicted with P. aeruginosa and in the S100A8/A9 monotherapy group compared to the ciprofloxacin/S100A8/A9 and the ciprofloxacin groups. For the G-CSF, the result indicates a lack of infection control and continuous need for influx of PMNs. The high-level IL-1β response suggests a pro-inflammation process. It has been suggested that IL-1β is part of a positive feedback loop that contributes to impaired wound healing (Mirza et al. 2013). However, in our study, the role rather seems to reflect a relative lack of biofilm control and an increased bacterial load. Another pro-inflammatory marker VEGF, an angiogenesis contributor regulated by HIF-1a/hypoxia, produced by endothelial cells, keratinocytes, platelets, macrophages and fibroblast was decreased in the group treated with only ciprofloxacin, but no other associations were found. The lack of a more distinct VEGF response is likely due to the relatively short observation period.

The level of IL-10 was analyzed to evaluate the anti-inflammatory response from the early T-cells, macrophages and dendritic cells. We found an inverse pattern in the IL-10 level compared to the IL-1β and the G-CSF response, fortifying the assessment made before. Together the G-CSF, IL-1β and IL-10 levels insinuate a basis for better healing in the double-treated group, and less obviously, but still significantly in both monotherapy groups.

The measured KC is a homolog to the human cytokine-induced neutrophil chemoattractant 1 (CXCL-1). The cytokine has chemotactic features for neutrophils and monocytes/macrophages and stimulates keratinocyte proliferation/migration and angiogenesis in the event of a cutaneous injury. Chronic wounds have lower expression levels of several CXC chemokine genes compared to an undamaged skin barrier (Wall et al. 2008). Our results indicate a higher level/activity of neutrophils and monocytes/macrophages by proxy of KC levels in the group receiving both ciprofloxacin and S100A8/A9. The levels of KC were quite diverse in the biofilm control group. It can be explained by the fact that five mice gained a little infection control (high level of inflammation in tissue), whereas two mice had a relative infection control at the point of termination. This is also evident in Fig. 1, where the same five mice have a high CFU count, compared to the two remaining mice moving towards the proliferative healing phase.

The TIMP-1 inhibit matrix metalloproteinases (MMPs) that degrade the extracellular matrix. The MMP/TIMP ratio influences the extracellular matrix protein degradation and remodeling. In chronic diabetic wounds, the MMP levels are high, whereas the TIMP levels are low (Muller et al. 2008). Interestingly, we found significantly higher levels of TIMP-1 in the double-treated group, demonstrating an ongoing active healing process. This observation paralleled the anti-inflammatory IL-10 levels, indicating a role as markers for an improved outcome of these two cytokines.

The level of the pro-inflammatory cytokine IL-17A was also investigated. It has been suggested that the inflammation caused by IL-17A could be associated with impaired healing in chronic wounds (Takagi et al. 2017) and that an inhibition accelerates healing by changing the polarization of the macrophage (Lee et al. 2018).

The level of IL-17A was not significantly different in the five groups of the present trial. This can be explained by the fact that the T-helper cells producing these interleukins are not activated in the innate phase of the immune defense. The PDGF analysis did not yield any significantly different levels in the compared groups, despite the therapeutic use of a compound containing this growth factor. This may be explained by the murine set-up or, again, our short observation period. Ratios between angiogenic/angiostatic CXC cytokines and MMP/TIMPs have previously been proposed to play an important role in moving the chronic wound from the chronic inflammatory state to a healing state (Fivenson et al. 1997; Shah et al. 2012; McCarty and Percival 2013). Despite the graphical insinuation of a significant difference between the TIMP-1/IL-1b ratio in the double-intervention group and that in the ciprofloxacin monotherapy group, no such difference was identified statistically (Fig. 5). However, the combination therapy was the only group resulting in an increased anti-/pro-inflammatory ratio compared to the biofilm controls, again supporting an improved and synergistic effect of combining the topical S100A8/A9 with systemic ciprofloxacin treatment.

Lastly, the wound size did not reveal a beneficial effect of any of the therapeutic interventions. However, this was not surprising due to the observation period of only 4 days. Although significant differences in cytokine response were observed supportive of our hypothesis and previous findings, we speculate whether the time period between last treatment and euthanasia of 24 hours might have impacted the level of the cytokines/growth factors in comparison to the bacteriology. That is, from the quantitative bacteriology, larger differences in host responses may have been expected.

In conclusion, the present study shows a synergistic or additive effect of combining traditional antibiotic therapy by means of ciprofloxacin with S100A8/A9, otherwise suppressed in non-healing wounds. The effect was observed in improved killing of biofilm-growing P. aeruginosa and indications of lowered pro-inflammatory and induced anti-inflammatory host factors. In addition, our results demonstrate that the ciprofloxacin potentiating effect of S100A8/A9 is highly dependent on the presence of host cells and tissues. Whether improved healing and reduced antibiotic resistance can be obtained remains to be investigated.

Conflicts of interest

None declared.