Abstract
The skin microbiome undergoes dynamic changes during different phases of wound healing, however the role of bacteria in the wound healing process remains poorly described. In this study, we aimed to determine how wound bacteria develop over time in epidermal wounds, and how they interact with inflammatory processes during wound healing. To this end, we analyzed wound fluid and swab samples collected from epidermal suction blister wounds in healthy volunteers. We found that bacterial numbers, measured in swabs and dressing fluid, increased rapidly after wounding and stabilized by day 8. The composition of bacterial species identified by MALDI-TOF mass spectrometry differed between wounds, but generally consisted primarily of commensal bacteria and remained largely stable over time. Inflammation and neutrophil activity, measured by quantification of cytokines and neutrophil proteins in dressing fluid, peaked on day 5. Exudation, measured by quantification of protein content in dressings, also peaked at this time and strongly correlated with cytokine and neutrophil protein levels. Inflammation, neutrophil activity, and exudation were not correlated with bacterial counts at any time, indicating that in normally healing wounds, these processes are primarily driven by the host and are independent of colonizing bacteria. Our analysis provides a comprehensive understanding of epidermal wound healing dynamics in the host and the role of the microbiome in healthy wound healing.
Introduction
Wound healing is a meticulously regulated biological process characterized by a series of well-orchestrated events involving inflammation, proliferation, and remodeling, all aimed at restoring tissue integrity and functionality(1, 2). Our understanding of normal wound healing is complicated by the presence of microorganisms as they have the potential to interact with and influence these intricate processes. Notably, the emerging concept of the skin microbiome has begun to reshape our understanding of skin health(3).
The skin microbiome comprises a complex community of microorganisms, including bacteria, fungi, viruses, and other microorganisms(3). This dynamic ecosystem interacts with the host’s skin and plays a pivotal role in various aspects of skin physiology. It contributes to immune modulation, nutrient metabolism, and protection against pathogens(4). Perturbations in the balance of this microbiota, such as biofilm formation and imbalanced microbial communities, have been implicated in delayed wound healing, the induction of chronic inflammation, and an increased risk of infection(3). Despite these insights, our understanding of the precise role of the skin microbiome in normal epidermal wound healing is still evolving. Unraveling the intricate relationship between the skin microbiome and the host during wound healing can advance our comprehension of skin health and the development of novel therapeutic strategies for dermatological conditions.
Epidermal wounds, which specifically affect the outermost layer of the skin, the epidermis, trigger a complex healing process aimed at restoring the integrity and functionality of this vital barrier(1, 2). To study such wounds and the dynamics of their bacteria, researchers have turned to the suction blister model—a tool that offers standardized and controlled wound conditions(5-9). This model involves the complete removal of the epidermal layer to expose the underlying dermis, ensuring uniform wound depth, size, and closure time. The suction blister model can therefore provide valuable insights into the dynamics of bacterial colonization and inflammation in presence of an intact dermal component while avoiding the scarring caused by alternative methods of experimental wound induction, such as biopsy procedures.
We present a focused secondary analysis of control samples derived from a randomized, double-blinded, placebo-controlled, first-in-human clinical trial that investigated the safety and tolerability of a novel peptide-based wound gel (NCT05378997)(10). In this study, 24 healthy volunteers received four suction blister wounds each, with two of these wounds treated with active product and the other two with a placebo gel. Over a 15-day period, we meticulously monitored the healing process and collected an extensive biobank of samples. Using untreated control samples from this biobank, the primary objective of this study was to describe the dynamics of bacterial development and host inflammatory processes during normal epidermal wound healing.
Materials and methods
Study design
In this study we analyzed samples collected as part of a randomized, controlled trial designed to test the safety and pharmacokinetics of thrombin-derived C-terminal peptide (TCP)-25 in epidermal suction blister wounds (clinicaltrials.gov identifier NCT05378997). Twenty-four subjects were included in the study with four suction blister wounds each. We only include samples from control wounds in this study. The study design is described in detail in the published study protocol(10).
The trial was approved by the Swedish ethical review authority (etikprövningsmyndigheten application number 2022-00527-01). Written consent was received from all subjects prior to participation.
Formation of epidermal suction blister wounds
Two blister wounds were created on the medial aspect of each thigh with a 6 cm distance between them. Before wounding, the wound sites were shaved and wiped with ethanol.
A Model NP-4 (Electronic Diversities, Finksburg, MD) suctioning device(11) operating in a negative pressure range of 200–400□mm Hg with heating was used to create blister wounds with a 10 mm diameter. The roof of each blister was excised using sterile forceps and scissors to expose the wound.
On each thigh, one wound was randomized to receive active drug and the other was randomized to receive control gel without the drug substance. Since drug treated wounds were not included in this study, we included 48 control wounds from 24 subjects.
Swab procedure
The wound was swabbed on day 1 (immediately after blister formation) and on days 2, 3, 5, 8, and 11 using a sterile cotton swab (Selefa, Stockholm, Sweden). The swab was pre-wetted with sterile phosphate-buffered saline (PBS). Using a twisting motion, the swab was rotated 10 times using light pressure, covering the entire wound and 2 mm outside of the initial wound edge on each side, and then placed into a microfuge tube with 0.5 mL sterile PBS and kept on ice. The tubes were vortexed to dislodge bacteria from the swab, and bacterial levels were quantified as described below.
Dressing extraction
Wounds were covered with Mepilex dressing (Mölnlycke healthcare, Gothenburg, Sweden) in between study visits. Dressings were changed on days 2, 3, 5, 8, and 11. At each dressing change, the dressing was removed from the wound, placed into an empty 5mL tube, and kept on ice. Each dressing was weighed in the tube, and then the dressing was removed and placed into a 5 mL syringe fixed with a stopper. Two milliliters of cold, sterile 10 mM Tris buffer at pH 7.4 was added to the syringe, and the syringe was vortexed for 5 minutes. The fluid was extracted from the dressing by depressing the plunger, emptying the contents into the original 5mL collection tube in which the dressing had originally been placed. The extracted fluid was kept on ice. Bacterial levels were quantified immediately after extraction in fresh (never frozen) dressing fluid as described below. We also added 100x Halt Protease Inhibitor Cocktail (Thermo Fisher Scientific, USA) to half of the dressing fluid per the manufacturer’s directions to a final strength of 1x, and samples were then aliquoted and frozen at -80°C. Protein content, neutrophil proteins, and cytokines in the dressing fluid were quantified in the stored dressing fluid samples that contained protease inhibitor, as described below.
Calculation of dressing fluid extract volume
To calculate the average weight of the dressing and the tube, six unused dressings and six unused 5mL collection tubes were weighed, and together were found to have an average weight of 3285.4 mg. This weight was subtracted from the recorded weight of each dressing sample (which had been weighed in the collection tube as described above). We assumed that wound fluid has a density of 1g/mL and used this to calculate the volume of wound fluid in each dressing sample in milliliters. To calculate the total volume of dressing fluid extract, we added 2mL to this calculated volume to account for the volume of Tris buffer used during the extraction. This calculated total volume of dressing fluid extract was used to calculate total bacteria and total protein levels as described below.
Quantitative bacterial counts
The swabs and dressing fluid samples were diluted with sterile PBS to generate 7 10-fold serial dilutions (from 10× to 107×). Six separate 10-μL drops of the undiluted sample and each of the dilutions were deposited on a Todd-Hewitt agar plate. The plates were incubated at 37°C in 5% CO2 overnight. The next morning, the number of colonies was counted and recorded. The number of colony forming units (CFU)/mL was calculated, and then multiplied by the total volume of fluid in the sample (0.5 mL for swabs or the volume calculated from the dressing weight described above for dressing fluid extract) to obtain the total CFU per swab and CFU per dressing.
Identification of bacteria
Swab and dressing fluid samples were streaked on a blood agar plate and incubated at 37°C overnight. Six colonies were selected from each plate. If different types of colonies were present on the plate upon visual inspection, we selected at least one of each of the different types and attempted to choose a representative number of each type based on their relative distribution on the plate. Colonies were prepared using the extended direct transfer sample preparation procedure on stainless steel MALDI target plates as described by the manufacturer (Bruker Daltronik GmbH). A Microflex LT/SH SMART MALDI-TOF mass spectrometry (MS) instrument with flexControl v. 3.4 (Bruker Daltronik GmbH) was used to analyze the target plate and collect mass spectra in linear mode over a mass range of 2 to 20 kDa. A spectrum of 240 summed laser shots was acquired for each sample spot. The spectra were analyzed using a MALDI Biotyper (MBT) Compass v. 4.1 with the MBT Compass Library Revision L (DB-9607, 2020) (Bruker Daltronik GmbH). A MALDI Biotyper score of 2.0 or more was required for species level determination, except for Corynebacterium where an identification with a score of 2.0 or more was limited to the genus level, and for species belonging to the Bacillus cereus complex where the identification was limited to the complex level.
Analysis of wound exudation level
Dressing fluid was analyzed for protein content by bicinchoninic acid (BCA) assay (Thermo Fisher, Waltham, MA, USA) per the manufacturer’s directions using bovine serum albumin as the standard. The concentration of protein in mg/mL was calculated from the standard curve. To determine the total protein content in the dressing, this concentration was multiplied by the total volume of dressing fluid extract, calculated as described above.
Wound imaging
A disposable centimeter-scale ruler was placed on the skin near the wounds. Two images were taken of each set of wounds on each leg at a distance of 35cm with a Canfield Twin Flash camera (Canfield Scientific, Parsippany-Troy Hills, NJ, USA) with a modified version of Canfields close up scale(12).
Quantification of cytokines
Cytokines (interferon [IFN]-γ, interleukin [IL]-1β, IL-2, IL-4, IL-6, IL-8, IL-10, IL-12p70, IL-13, tumor necrosis factor [TNF]-α) were measured using a V-PLEX Proinflammatory Panel 1 (human) kit (Meso Scale Diagnostics, Rockville, MD, USA, art.no: K15049D) per the manufacturer’s instructions at a 5× dilution. Samples for which IL-8 was above the upper limit of detection were analyzed again using the U-PLEX Biomarker Group 1 (human) kit (Meso Scale Diagnostics, art.no: K151TYK-2) at a 100× dilution. The concentration of each cytokine was calculated using the MSD discovery Workbench analysis software. Samples with a concentration below the detection limit were not able to be quantified by the analysis software and were assigned a value of zero for all analyses.
In some analyses, to account for different levels of wound fluid in the dressing extract samples, cytokine levels were normalized for the protein concentration measured by BCA assay, as described above, to obtain the amount of each cytokine per mg of protein in the dressing fluid extract sample.
Quantification of neutrophil proteins
Human neutrophil elastase (NE) and human myeloperoxidase (MPO) were measured using ELISA kits from R&D Systems (Minneapolis, MN, USA); heparin-binding protein (HBP) was determined using an ELISA kit from Aviva Systems Biology (San Diego, CA, USA). Samples were analyzed according to the manufacturer’s instructions in duplicate. ELISA signals were read at 450 nm on a microplate reader (BioRad, Hercules, CA, USA). A standard curve was established with reference standards to calculate the concentration in each sample.
In some analyses, to account for different levels of wound fluid in the dressing extract samples, neutrophil protein levels were normalized for the protein concentration measured by BCA assay, as described above, to obtain the amount of each neutrophil protein per mg of protein in the dressing fluid extract sample.
Statistics
Summary data were plotted as median and interquartile range. For correlation between two variables, Spearman correlation coefficients were determined. Though each participant had two control wounds that were included in this study, individual wounds were considered independent biological replicates in all analyses (n=48 wounds). Due to the non-normal distribution of the data (determined by visual inspection of the data), non-parametric analyses were used throughout. All statistical tests used were two-tailed. Due to the exploratory nature of the study, p-values were not adjusted for multiple comparisons and should be considered hypothesis-generating only. Data were analyzed and plotted using GraphPad Prism 10 software. When a logarithmic axis was used, zero values were replaced with 1 to enable visualization of all data points on the graph, but were not replaced in any statistical analyses.
Ethical considerations
The trial and collection of specimens to the biobank was approved by the Swedish ethical review authority (etikprövningsmyndigheten application number 2022-00527-01). Written informed consent was received from all subjects prior to participation.
Results
Bacterial levels
To explore how bacterial levels in normally healing epidermal wounds change over time, we quantified the number of bacteria in swab and dressing fluid extract samples. Immediately after formation of the wound (Day 1), bacterial levels were zero in most of the swab samples (Figure 1 A, supplemental table 1). Subsequent samples revealed the rapid reconstitution of the microbiome over time already one day after the wounds were formed (Day 2), corresponding to a median of 2.9×103 CFU per swab and 1.9×104 CFU per dressing.
A) Quantitative bacterial counts in swab samples (left) and dressing fluid samples (right) over time. Each data point is the bacterial level measured in a single wound, while the blue line is the median of all wounds that day (n=48) and whiskers are the interquartile range. Zero values were replaced with 1 in the graph to enable visualization on the logarithmic axis. B) Heat map of quantitative bacterial counts in swab samples (left) and dressing fluid samples (right) from each wound. Data from the left and right wound are each presented in separate cells at the top and bottom of each row, respectively. C) Scatterplot showing quantitative bacterial counts in swabs vs quantitative bacterial counts in dressing fluid in each wound. Zero values were replaced with 1 in the graph to enable visualization on the logarithmic axis. Each day is indicated in a different colour.
Thereafter bacterial levels in the wound increased steadily until day 8 when they reached a peak, corresponding to a median of 1.7×106 CFU per swab and 1.6×107 CFU per dressing. On day 11, bacterial levels decreased slightly in swabs to a median of 1.3×106 CFU per swab. They decreased to a much larger extent in dressings on day 11, to a median of 1.4 x 106 CFU per dressing.
Heat maps of bacterial levels in each wound showed considerable individual variation (Figure 1B). Generally, wounds with low bacterial levels in swabs also appeared to have low bacterial levels in dressing fluids, and vice versa. To confirm this, we plotted bacterial levels in swabs and dressings as a scatterplot (Figure 1C) and determined the Spearman correlation coefficient on each day and overall (Table 1). The data indicated a strong correlation between bacterial levels in swab and dressing fluid samples on each day (R = 0.607 – 0.705) and overall (R = 0.794).
Bacterial levels plotted on the heatmap appeared to be similar in left and right wounds in most participants, but with some variation. To confirm this, we determined the Spearman correlation coefficient on each day and overall (Table 1). The data indicated a weak to moderate correlation between left and right wounds on each day in swabs (R = 0.305 – 0.511) and dressings (R = 0.230 – 0.653), but a strong correlation overall in both swabs (R = 0.723) and dressings (R = 0.816) when all days were included, indicating that there is moderate variation between left and right wounds in each individual as the wound heals, but that overall bacterial levels do not differ greatly between wounds on the left and right leg, corroborating the data presented in Figure 1C and Table 1, which showed no difference in median bacterial levels between wounds on the left and right leg.
Identity of bacteria
To determine which bacteria colonize wounds during the healing process and whether they differ over time and between individuals, we analyzed 6 colonies that grew from streaks of each swab and dressing sample on day 3 and 8. The data indicate that each individual developed a unique bacterial growth pattern consisting mainly of commensal bacteria, with some wounds containing potentially pathogenic bacteria such as S. aureus (Figure 2A). The overall proportion of each bacterial species was similar between day 3 and 8 and between swabs and dressings (Figure 2B) indicating that the bacterial composition does not change greatly over time in wounds during the healing process. A notable exception was the increase in the number of wounds containing the potential pathogen S. aureus and the commensal Corynebacterium sp. On examination of images of the wounds over time we noted that, in spite of differences in the composition of bacteria, all wounds healed at roughly the same pace and had a similar appearance at each stage(12). Images of representative wounds that were dominated by commensals or by potential pathogens (Bacillus sp. and S. aureus) are shown in Supplemental Figure 1.
A) Bacterial species identified by MALDI-TOF MS in six colonies picked from swab and dressing fluid cultures from each wound on days 3 and 8. Bacteria from the left wound are in the top cells and bacteria from the right wound are in the bottom cells of each row. B) Pie charts showing the proportion of each identified bacterial species in swab and dressing fluid samples on day 3 and 8. The colours assigned to the species are the same as in the legend in panel A.
Wound exudation
To determine how the level of wound exudation changes over time, we measured protein content in dressing fluid as a proxy. Because all dressing fluids were extracted with the same volume of extraction buffer, the amount of protein indicates the accumulated amount of wound fluid in the dressing without being affected by possible evaporation through the Op-Site film covering the dressing. The results indicated that the level of exudate was moderate on days 2 and 3 and then peaked on day 5 before dropping on day 8 (Figure 3A). On day 11, dressings were very dry and had near-zero levels of protein content, indicating very low exudation on this day.
A) Protein concentration quantified in dressing fluid from each wound at each time point. Each data point is protein concentration in a single wound, while the blue line is the median of all wounds that day (n=48), and whiskers are the interquartile range. B) Heat map of protein concentration quantified in dressing fluid samples from each wound. The left and right wound are each presented as a single cell on the top and bottom of each row, respectively. C) Scatterplot comparing protein concentration in dressing fluids from wounds on the left leg vs the right leg in each participant. D) Scatterplot comparing protein concentration in dressing fluids vs quantitative bacterial counts measured in swabs from each wound. Zero values were replaced with 1 in the graph to enable visualization on the logarithmic axis.
When plotted as a heatmap, protein content showed considerable individual variation, and appeared to be similar between left and right wounds (Figure 3B). To confirm this, we plotted protein levels in the left and right wound as a scatterplot (Figure 3C) and determined the Spearman correlation coefficient on each day and overall (Table 2). The data indicated a moderate to high correlation between left and right wounds on all days (R = 0.356 – 0.803), except for day 11 which had many zero values. The data also indicated a strong correlation overall when all days were included (R=0.791), indicating that there is some variation in exudation between left and right wounds in each individual as the wound heals, but that overall exudation levels do not differ greatly between wounds on the left and right leg.
To explore whether the amount of bacteria affects the level of exudation, we plotted bacterial levels measured in swabs against protein content as a scatterplot (Figure 4D) and determined the Spearman correlation coefficient on each day and overall (Table 2). Surprisingly, the data indicated that there is weak to no correlation between bacterial levels and protein content on each day (R=0.049 – 0.221) and no correlation overall (R=-0.119), indicating that the amount of bacteria in normally healing epidermal wounds does not affect their level of exudation.
A) Levels of ten different cytokines measured by MesoScale in dressing fluid over time. Each data point is the wound fluid cytokine concentration in a single wound, while the blue line is the median of all wounds that day (n=48), and whiskers are the interquartile range. B) Spearman correlation coefficients comparing levels of each cytokine with quantitative bacterial counts measured in swabs (left) and comparing levels of each cytokine with protein content in that wound (right) (n=48).
Inflammation
Inflammation plays a key role in the wound healing process. Therefore, to determine how inflammation develops over time in normally healing epidermal wounds, we measured the levels of 10 different cytokines in the dressing fluid on days 2, 3, 5, and 8. Because there was very little exudation on day 11, we did not measure cytokine levels on this day as they were likely to be undetectable. Although the individual cytokines had vastly different levels, ranging from about 1ng/mL (IL-4) to over 10000 ng/mL (IL-8) at their peak, all cytokines followed similar dynamics over time (Figure 4A). Most measured cytokines remained at approximately the same level from day 2 to 3 (IL-1β, IL-2, IL-8, IL-12p70, IL-13, TNF-α), while some declined slightly on day 3 relative to day 2 (IL-4, IL-6, IL-10). IFN-γ, in contrast, increased steadily from day 2 to 5. In spite of differences in initial dynamics, all measured cytokines reached their peak level on day 5 and then declined on day 8.
To determine whether bacterial levels affect the level of inflammation, we determined the correlation coefficient between the level of each cytokine and the bacterial level detected in swabs in that wound and plotted the resulting correlation coefficients as a heatmap (Figure 4B). For most cytokines, there was little to no correlation between bacterial counts and cytokine levels on days 2, 3, and 8, and a weak correlation on day 5. The cytokines most strongly correlated with bacterial counts were TNF-α on day 3 and 5 (R = 0.475 and 0.488 respectively), IL-10 on day 5 (R = 0.531), and IL-1β on days 5 and 8 (R = 0.497 and 0.480 respectively).
Inflammation is known to induce capillary leakage and subsequent exudation. We determined the correlation coefficient between the level of each cytokine and protein content of the dressing fluid (a proxy for exudation). We found that there was a moderate to strong correlation between cytokine levels and protein content on all days (R = 0.257 to 0.860), particularly on day 5 (R=0.669 to 0.860).
To confirm that measured cytokine levels are not dependent on the amount of protein content in the dressing fluid, we normalized the levels of each cytokine for the protein content of that same sample (Supplemental Figure 2A). The overall dynamics of the cytokines did not change greatly when normalized for protein content, reaching their peak at day 5. The normalized level of some cytokines (IL-1 β, IL-4, IL-6, IL-12p70, TNF-α) did not decrease on day 8, unlike un-normalized levels. The correlation with bacterial levels became slightly stronger for most cytokines when they were normalized for protein content (Supplemental Figure 2B).
Neutrophils
Neutrophils play a key role in wound healing and bacterial control. Therefore, we measured the levels of three different neutrophil proteins (MPO, NE and HBP) on days 2, 3, 5, and 8 (Figure 5A). Due to the small amount of exudation on day 11, proteins were likely to be undetectable so we did not measure neutrophil protein levels on this day. MPO and NE remained at approximately the same level from day 2 to 3, peaking on day 5, and then dropping sharply on day 8. In contrast. HBP increased steadily until it peaked on day 5 and then dropped on day 8.
A) Levels of three different neutrophil proteins measured in dressing fluid over time. Each data point is the wound fluid concentration in a single wound, while the blue line is the median of all wounds that day (n=48), and whiskers are the interquartile range. B) Spearman correlation coefficients comparing levels of each neutrophil protein with: IL-8 (top); quantitative bacterial counts measured in swabs (middle;) and protein content (bottom) (n=48).
The levels of each neutrophil protein were moderately to strongly correlated with levels of IL-8, a potent chemoattractant for neutrophils, on each day (R=0.495 to 0.812) (Figure 5B). There was little to no correlation between bacterial levels in swabs and neutrophil proteins on each day (R = 0.011 to 0.353) suggesting that neutrophil protein levels are driven more strongly by host factors (IL-8) rather than bacteria in these normally healing and uninfected wounds (Figure 5B).
Neutrophil recruitment and activation are associated with increased capillary leakage, and thus increased extravasation. We determined the correlation coefficient between the level of each neutrophil protein and protein content of the dressing fluid (a proxy for exudation) (Figure 5B). We found that there was a moderate to strong correlation between neutrophil protein levels and protein content on all days (R = 0.419 to 0.869), particularly on day 8 (R=0.813 to 0.867). Notably, this correlation was stronger than it was for most cytokines and occurred later (Figure 4B vs Figure 5B), suggesting that inflammation and neutrophil recruitment/activation may drive exudation in different stages of wound healing. To confirm that measured neutrophil protein levels are not dependent on the amount of protein content in the dressing fluid, we normalized the levels of each neutrophil protein for the protein content of that same sample (Supplemental Figure 2C). The overall dynamics changed slightly, with normalized neutrophil proteins reaching their peak level on day 3, instead of day 5.
Discussion
In this study, we comprehensively describe the dynamics of bacteria, inflammation, and exudation in epidermal suction blister wounds and their interrelatedness during the wound healing process (Figure 6). Our findings provide valuable insights into the intricate interplay between these factors and their implications for normal wound healing.
Overview of the dynamics of bacterial levels, exudation, neutrophil proteins, and inflammation in normally healing wounds over time, and the relationship between them. Dashed lines are a reasonable extrapolation to timepoints not measured in this study.
One of the central aspects of our investigation was the examination of bacterial colonization in epidermal wounds. We observed that newly formed wounds had nearly no bacteria due to removal of the epidermal layer and cleaning of the skin with ethanol prior to wounding. Shortly after wound formation bacteria were detected, with a steady increase in bacterial levels until day 8, followed by a slight decrease on day 11. This pattern suggests a dynamic relationship between bacterial colonization and the healing process. The wet, nutrient-rich environment of the dressing-covered wound likely results in increased bacterial growth in the wound compared with intact skin which generally has lower levels of bacteria(13, 14). The importance of moisture for bacterial growth is well known. Intact skin covered by an occlusive dressing that retains moisture displays an increase in the number of bacteria after only one day(15). As the wound becomes re-epithelialized over time, less of the wound environment is exposed and exudation decreases, which could explain the observed decrease in bacterial levels on day 11. We did not collect swabs after day 11 so it is unclear when exactly bacterial levels return to those found in unwounded skin and how removal of the occlusive dressing affects their dynamics.
The composition of bacteria found in the wounds remained relatively stable over time. Although there was some variation in microbiome composition between individuals, most wounds had bacteria that are typically classified as skin commensals. Some wounds harbored potential pathogens such as S. aureus, highlighting the importance of monitoring bacterial profiles during wound healing. It is well established that S. aureus underlies postoperative infections in various types of wounds, such as postoperative wounds(16). However it is also worth noting that presence of S. aureus does not define infection per se, as the bacterium can also occur in normally healing wounds(17), indicating that complex bacteria-host interactions determine S. aureus infection risk. Overall, in spite of the difference in microbial composition between wounds and the presence of potential pathogens, we did not note any major differences in the rate or quality of wound healing in the epidermal wounds studied here. While other studies have determined the microbial composition of intact skin at various sites of the body(3), and of epidermal(8) and various chronic wounds(3, 18-20), as far as we are aware, this is the first study that in a systematic way explores major cultivable bacterial species (i.e. those that can be cultured using standard clinical microbiology techniques) present in normally healing epidermal wounds and their relationship to inflammation biomarkers.
Our study also revealed a strong correlation between bacterial levels in swab and dressing fluid samples on each day and overall. This finding underscores the reliability of swab samples in representing the bacterial composition of wound dressing fluid. Our results suggest that, given this strong correlation, analysis of dressing fluid bacteria could be an alternative way to collect information about bacteria in the wound environment when swabs can’t be used, as the swab procedure can be painful for patients. Furthermore, while intra-individual variation existed between left and right wounds in each participant, shown by the poor correlation between bacterial levels in wounds on the left and right leg, overall bacterial dynamics followed the same trend on both legs. Microbial composition between left and right wounds also exhibited some intra-individual variation, with some patients having a dissimilar bacterial composition between the two wounds.
Inflammation is a fundamental aspect of the wound healing process(21-23). Our analysis of cytokines in dressing fluid provided insights into the dynamics of inflammation in normally healing epidermal wounds. Most measured cytokines followed similar patterns, remaining low on days 1 and 3 and reaching peak levels on day 5. Interestingly, there was little to no correlation between bacterial counts and cytokine levels, indicating that bacterial presence may not be the primary driver of inflammation in normally healing wounds. Inflammation was correlated with protein content in dressing fluid, particularly on day 5 when inflammation was highest, corroborating a well-known link between inflammation and capillary leakage, which in turn leads to exudation(24).
Neutrophils play a crucial role in wound healing and bacterial control and both neutrophil infiltration and activation are known to increase in the inflammatory stage of wound healing(21-23). We observed that neutrophil protein levels, which could be attributed to increased neutrophil recruitment and activation, followed similar dynamics to inflammation. MPO and NE increased on day 5, while HBP began to increase already on day 3. These dynamics are consistent with the fact that HBP is stored in the secretory vesicles which have a high propensity for release, while MPO and NE are stored exclusively in the azurophilic granules which have a low propensity for release(25). Neutrophil protein levels were strongly correlated with IL-8, a potent chemoattractant for neutrophils. Again, we found little correlation between observed bacterial levels and neutrophil proteins, reinforcing the notion that host factors, such as inflammation and chemotactic signals, may be more influential in determining neutrophil activity than are bacterial levels. We found that neutrophil protein levels were strongly correlated with wound exudation. Neutrophils are potent inducers of capillary permeability and so a strong correlation between neutrophil protein levels and wound exudation is expected.
We found that exudation, as indicated by protein content in dressing fluid, was moderately high already on day 2, then peaked on day 5, before declining on day 8. Wound exudation, the process of fluid discharge from wounds, is a critical aspect of the wound healing process(24). Our findings suggest that the factors influencing exudation in wounds are multifaceted and extend beyond the mere presence and abundance of commensal bacteria, as no significant correlation with commensal bacterial levels was observed in our study. Inflammation, a hallmark of early wound healing, can lead to capillary leakage due to the release of signaling molecules and cytokines by immune cells, contributing to exudation(24). Certain microbial species and their proinflammatory products can, irrespective of overall bacterial levels, indirectly affect exudation by perpetuating inflammation and may in some cases hinder wound healing if they are present at excessive levels(26). We did not find a correlation between exudation and bacterial levels, suggesting that this mechanism is not at play in normally healing epidermal wounds. Finally, exudation could be related to mediators not yet studied in wounds. Understanding these relationships is critical for effective wound care strategies, and further research is needed to elucidate their specific roles in wound healing and exudation regulation.
These findings collectively highlight the dynamic and intricate nature of the wound healing process. Bacterial colonization, inflammation, neutrophil activity, and exudation appear to be interrelated but not solely dependent on one another (Figure 6). Instead, they each contribute to various stages of wound healing, and their coordination is essential for successful tissue repair.
A major strength of the study is the use of a suction blister model to create standardized wounds in human subjects and the rigorous follow-up that included collection of an extensive biobank at several timepoints. This allowed us to conduct a rigorous multifaceted study of bacteria and host factors throughout the healing process in normal epidermal wounds. However, the study has some limitations which can be addressed in future studies. First, the bacterial levels found in this study may not be representative of bacterial levels in all wounds. Occlusive dressings, like Mepilex which was used to cover wounds in this study, may affect bacterial growth, so it is possible that wounds covered by other dressings could have different levels during healing or reach a lower peak level(15, 27, 28). On the other hand, culture-based bacterial quantification methods can possibly underestimate the bacterial burden in wounds so it is possible that bacterial levels in these wounds could be even higher than what we report(29). Although microbiological analysis of wound biopsies provides no additional prognostic information when compared with analysis of the surface microflora obtained by swabs(30), swab and tissue biopsy results may show differences in certain species, including anaerobes. Another limitation is the fact that the MALDI-TOF MS method was only used to identify six colonies per wound and therefore obviously does not provide a full picture of the composition of the microbiome.
Conclusions
In conclusion, our study provides a comprehensive analysis of bacteria, inflammation, and exudation in epidermal suction blister wounds during normal wound healing. We have unveiled complex relationships between these factors, shedding light on their roles and interactions throughout the healing process. This knowledge contributes to our understanding of wound healing dynamics and may have implications for the development of future wound care strategies and interventions, especially in the context of dermatological conditions and skin health. Further research in this area could yield valuable insights into optimizing wound healing outcomes and preventing complications.
Sources of funding
The exploratory study presented here was supported by grants from the Swedish Research Council (project 2017-02341, 2020-02016), Edvard Welanders Stiftelse and Finsenstiftelsen (Hudfonden), the Royal Physiographic Society, the Crafoord and Österlund Foundations, and the Swedish Government Funds for Clinical Research (ALF).
Authors’ contributions
The following authors contributed to each of the following roles (defined according to the CRediT taxonomy): Conceptualization, SL, GP, KW, MP, KS, AS; design of methodology, SL, KW, MP; formal analysis (statistics), GP, JFPC, CL, JF; investigation (performing experiments and data collection) and validation, SL, GP, KW, JFPC, ACS, FF, CL, BN; resources (provision of patients and samples), SL, KW, KS, AS; data curation, JFPC, JF; writing (original draft preparation), SL, KW, JF, AS; writing (review and editing), SL, GP, KW, JFPC, ACS, FF, CL, BN, EH, JF, MP, KS, AS.; visualization, GP, JFPC, CL, EH, JF; supervision, GP, KS, AS; project administration, SL, GP, KW, JF, AS; funding acquisition, AS. All authors have read and agreed to the published version of the manuscript.
Data sharing and availability
The data that support the findings of this study are openly available in Zenodo at http://doi.org/10.5281/zenodo.10283373, reference number 10283373.
Conflict of Interest disclosure statement
A.S. is a founder of in2cure AB, a parent company of Xinnate AB which was the sponsor of the clinical trial from which the biobank samples used in this study are derived. G.P. is employed part-time (20%) by Xinnate AB. The other authors have declared that no conflict of interest exists.
Acknowledgments
Xinnate AB provided the funding, project management resources, and expertise for the regulatory development, enabling the safety study that generated the control samples used in this work. We would like to thank Dr. Andreas Sonesson for valuable discussions and Susanne Erdmann and Anne Nielsen and other personnel at the Department of Dermatology Lund. Lastly, we would also like to thank Åsa Håkansson and Kerstin Weber and other personnel at the Clinical Trial Unit at Skåne University Hospital Lund for support and use of their facilities during the clinical study.
Abbreviations
- BCA
- bicinchoninic acid
- CFU
- colony forming units
- ELISA
- enzyme-linked immunosorbent assay
- HBP
- heparin-binding protein
- IFN
- interferon
- IL
- interleukin
- MALDI-TOF
- Matrix Assisted Laser Desorption lonization -Time Of Flight
- MBT
- MALDI Biotyper
- MPO
- myeloperoxidase
- MS
- mass spectrometry
- NE
- neutrophil elastase
- PBS
- phosphate-buffered saline
- TCP
- thrombin-derived C-terminal peptide
- TNF
- tumor necrosis factor