Abstract
Diabetes mellitus is a frequent cause for chronic, difficult-to-treat wounds. New therapies for diabetic wounds are urgently needed and in-vitro or ex-vivo test systems are essential for the initial identification of new active molecules. The aim of this study is to compare in-vitro and ex-vivo test systems for their usability for early drug screening and to investigate the efficacy of a birch bark triterpene extract (TE) that has been proven ex-vivo and clinically to accelerate non-diabetic wound healing (WH), in a diabetic context. We investigated in-vitro models for diabetic WH, i.e. scratch assays with human keratinocytes from diabetic donors or cultured under hyperglycaemic conditions and a newly developed porcine ex-vivo hyperglycaemic WH model for their potential to mimic delayed diabetic WH and for the influence of TE in these test systems. We show that keratinocytes from diabetic donors often fail to exhibit significantly delayed WH. For cells under hyperglycaemic conditions significant decrease is observed but is influenced by choice of medium and presence of supplements. Also, donor age plays a role. Interestingly, hyperglycaemic effects are mainly hyperosmolaric effects in scratch assays. Ex-vivo models under hyperglycaemic conditions show a clear and substantial decrease of WH, and here both glucose and hyperosmolarity effects are involved. Finally, we provide evidence that TE is also beneficial for ex-vivo hyperglycaemic WH, resulting in significantly increased length of regenerated epidermis to 188±16% and 183±11% (SEM; p<0.05) compared to controls when using two different TE formulations. In conclusion, our results suggest that microenvironmental influences are important in WH test systems and that therefore the more complex hyperglycaemic ex-vivo model is more suitable for early drug screening. Limitations of the in-vitro and ex-vivo models are discussed. Furthermore our data recommend TE as a promising candidate for in-vivo testings in diabetic wounds.
Introduction
Diabetes mellitus type 2 is an important cause for chronic wounds. Approximately 25% of the patients develop diabetic foot ulcer which is a severe burden for the patients and the society[1]. Therefore, test systems reflecting the diabetic phenotype are needed to investigate the underlying mechanisms and to develop new active pharmaceutical ingredients and formulations for the treatment of these chronic wounds. Animal models, particularly mouse and rat, are often used to investigate diabetic wound healing (WH) and delivered important knowledge [e.g.[2–4]. However, their usability is limited especially in early drug discovery due to ethical and economic aspects. Furthermore, there are physiological differences between rodent and human WH[5]. Pigs are better suited as model system due to physiological similarities to humans [5] but they are even more expensive. To use a human model system and to reduce costs, human in-vitro cell culture based systems are often utilized for the investigation of WH. For example, for the investigation of the regeneration phase, in-vitro WH assays have been developed using monocultures of human keratinocytes or fibroblasts[6, 7]. These assays allow investigating to which extent changes of migration (in some cases combined with proliferation) influence WH. These more simplified models have the advantage and disadvantage to involve fewer cellular interdependencies which, on one hand, might enable the dissection of the various mechanisms in WH but, on the other hand, makes them less comparable to the in-vivo situation. Thus, they are useful for the elucidation of specific aspects of the mode of action, but may be less useful for a general drug screening.
Several variations of these assays exist[6, 7]. When regarding the similarity to the WH process in-vivo, the scratch assay is the most comparable, because the removal of cells with a pipette tip wounds the monolayer and results in destroyed and activated cells. The removal of a space holder, which is used in exclusion zone assays, mainly abrogates a growth limitation, and migration in Boyden chambers measures migration of single cells through a pore. Therefore, we focus here on scratch wound assays.
In general, cells of diabetic origin as well as cells treated with high glucose conditions are often used to mimic diabetic conditions [e.g.[8–13]. The first aim of our study was to explore the usability of keratinocytes of diabetic origin or keratinocytes from non-diabetic donors grown under high glucose conditions in scratch wound assays as model systems for the investigation of diabetic WH with respect to early drug screening (end point: scratch wound healing at different time points). High glucose conditions were used in scratch assays before, but contradicting data were published. While some authors [13–15][11–13] described delayed wound closure in keratinocytes under hyperglycaemic conditions, this was not the case in experiments performed by others[14]. To elucidate factors which might lead to these contradicting observations, we performed experiments investigating the influence of the history of cells (juvenile (<5 years) versus adult donors, non-diabetic versus diabetic donors) and microenvironment (different media, presence and absence of supplements) in our in-vitro WH assays.
As we observed a dependency on microenvironment, our second aim was to develop a more complex model representing a more physiological context in terms of cells types and subcellular matrix, but still capable of moderate throughput and free of ethical considerations. Thus we developed a porcine ex-vivo WH model under hyperglycaemic conditions and tested whether reepithelialization in these models was delayed similar to diabetic skin and depicted the influence of glucose and hyperosmolarity. Although human and porcine ex-vivo skin WH models have successfully been used for the investigation of non-diabetic WH before [e.g. [15–21], the data presented in this study show, to our knowledge, the first successful established diabetic/hyperglycaemic WH model.
Finally, to apply our different test systems and therefore to further investigate their usability, we tested the efficacy of a birch bark extract (triterpene extract, TE) which is known to improve WH under non-diabetic conditions in-vivo and ex-vivo [22–24] and which is approved by the European Medicines Agency for mid-dermal wounds. We show that the in-vitro assays under euglycaemic conditions were not able to reproduce the beneficial effect of TE on WH seen in-vivo. Also under hyperglycaemic conditions TE did not improve scratch WH in-vitro. However, there was a stimulating effect on reepithelialization in ex-vivo WH models under hyperglycaemic conditions, similar to what was shown in ex-vivo models and in-vivo under euglycaemic conditions[22, 24]
Materials and Methods
Tissues
Normal skin tissue samples for keratinocyte cultures from adult donors were derived anonymously during routine clinical removal of epidermal cysts and tumors. The samples were localized at least 2 cm from the respective lesions. Skin samples for juvenile non-diabetic keratinocyte and fibroblast cultures were derived from circumcisions. The utilization of adult and juvenile skin was approved by the ethics committee of the Ärztekammer Hamburg: vote No 060900 and WF-61/12. Juvenile: n = 4, male, age < 5 years; adult: n = 7; 5 male, 1 female, 1 unknown; age 66–93 years, mean 75 years; Tissue samples from diabetic donors were derived from the upper thigh of patients with diabetes type 2 (approved by the ethics committee of the Ärztekammer Hamburg: vote No. PV4198). n = 7; 6 male, 1 female, age 59–86, mean 71 years, HbA1c: 5.5%– 8.8% (37 mmol/mol– 73 mmol/mol); mean 7.1% (54 mmol/mol). All patients (or their next of kin) gave their written informed consent. All investigations were conducted according to the principles expressed in the Declaration of Helsinki. Note that not all donors were included in all experiments (see figure legends for n-numbers).
Media, extracts and formulations
K-SFM (Gibco/Life Technologies, Darmstadt, Germany), EpiLife (Cascade Biologics; Invitrogen, Darmstadt, Germany), and DermaLife (Lifeline Cell Technology/Cellsystems, Troisdorf, Germany) were used for keratinocyte cultures, RPMI with 10% FCS (both Biochrom, Berlin, Germany) for fibroblast culture. TE (Table 1) and betulin were obtained from Birken AG (Niefern-Öschlbronn, Germany). For in-vitro cell culture experiments TE and betulin were dissolved in DMSO and further diluted in medium. For controls, DMSO was diluted in medium with the same dilution factor. Oleogels containing 10% (w/w) TE were prepared by dispersing the TE in sunflower oil using an Ultra-Turrax T25 (IKA, Staufen, Germany) at 8000 rpm for 3 min. After a storage period of 24 h, the predetermined amount of water was added to the oleogels using a syringe-to-syringe mixing technique yielding homogenous w/o-emulsions.
Table 1. Chemical composition and physical characteristics of the TE used.
| Chemical composition | Betulin 86.85%, lupeol 3.94%, betulinic acid 3.54%, erythrodiol 0.77%, oleanolic acid 0.62%, unidentified substances 4.28% |
| Specific surface area | 42 ± 0.4 m2/g |
| Particle size D50% | 5.8 μm |
Cell culture
For cultivation of keratinocytes and fibroblasts from human foreskins and adult tissues see[20].
For euglycaemic conditions, keratinocytes were cultured in the respective keratinocyte medium containing 6 mM glucose, fibroblasts in normal RPMI medium (containing 11 mM glucose), for hyperglycaemic conditions, in medium containing 25 mM glucose (keratinocytes and fibroblasts) as well as 50 mM glucose (fibroblasts) and for hyperosmolaric conditions in medium containing 6 mM glucose and 19 mM mannitol for 48 h prior to the start of the experiment.
Scratch assays
Conventional and semi-automated (high throughput) (Essen BioScience, Ltd., Hertfordshire, United Kingdom) scratch assays were used. For details, comparison and discussion of the two systems see S1 File, S1 Table and S1 Fig.
Evaluation of scratch WH progress
The images documenting the WH progress were evaluated by using the software TScratch[25]. Non-wounded areas within a visual field were marked, while the scratch wound was left out. Based on this classification, TScratch calculated the scratch wound area as a percentage of the total visual field. To depict the scratch wound area in mm2 a picture of a length scale was taken with the same settings and the measurement tool of ImageJ software (NIH, Bethesda, MD; http://imagej.nih.gov/ij) was used.
Diabetic porcine ex-vivo WH model
Non-diabetic ex-vivo WH models were produced as described previously [[26]; patent No DE10317400]. For diabetic conditions, the 6 mm biopsies were incubated with 50 mM glucose for 48 h prior to wounding and also after wounding. For hyperosmolaric conditions, 6 mM glucose and 44 mM mannitol were used. WH was stopped 48 h after wounding by shock freezing and the WH progress was evaluated by measuring the length of the regenerated epidermis in H+E stained sections of the central part of the models with an Axiophot II microscope (Zeiss, Jena, Germany) and the measurement tool of the Openlab software 2.0.9 (Improvision, Coventry, UK).
Cytokine expression
To investigate the influence of TE on cytokines at different glucose levels in fibroblasts we seeded 50,000 cells per 6-well, incubated for 3 days in RPMI, adjusted then glucose levels to 25 mM and 50 mM, incubated for 24 h in medium including 10% FCS and then another 24 h in medium without 10% FCS. Subsequently we added 1 μg/ml or 10 μg/ml TE respectively, and incubated for another 6 h. DMSO in the concentration as used in the higher concentrated TE preparation was used as control. Cells were then harvested in RTL buffer and mRNA isolation was performed with the Micro RNA kit from Qiagen (Hilden, Germany) according to the manufacturer´s instructions. 1 μg of total RNA was used for first-strand cDNA synthesis with the cDNA synthesis kit from Thermo Fisher Scientific (Waltham, MA, USA) as suggested by the manufacturer. 4.5 μl of a 1:25 dilution of cDNA was used as a template in real-time quantitative PCR analysis with FAM dye-labeled TaqMan probes (GAPDH: Hs03929097_g1; IL-1 ß: Hs01555410_m1; IL-8 Hs00174103_m1; IL-6: Hs00985639_m1; CCL2: Hs002334140_m1; FGF2: Hs00266645_m1; FGF7: Hs00940253_m1; all Invitrogen Darmstadt, Germany) in a Light Cycler 96 (Roche Diagnostics, Mannheim, Germany) under conditions recommended by Roche. All real-time quantitative PCR analyses were performed with n = 4 in technical triplicates. Relative transcriptional levels comparing different treatments were calculated using the 2- ΔΔ Ct method[27].
Statistical analyses
Data are presented as mean values +/- standard error of mean.
For all settings, except cytokine expression, multilevel mixed-effects linear regressions with setting-dependent random intercepts were performed. The distribution of the data was checked and if necessary log transformed to reach normal distribution.
Using the Likelihood-ratio test the significance of the predictors as well as their interaction was tested. Two-tailed values of p< 0.05 were considered statistically significant. In the case of insignificant effects the predictor/interaction was excluded from the model (backward elimination). For the resulting model polynomial-contrast-tests with Bonferroni adjustment for multiplicity were performed to compare subgroups. The appropriate means with corresponding 95% confidence intervals and p-values were reported, all log transformed values were back-transformed.
For ex-vivo models, “wound” was chosen as random intercept to account for the repeated measurement structure within each wound. Interaction between glucose/mannitol concentration and the influencing factor (preincubation and treatment, respectively) were included as fixed effects.
For all in-vitro models, the cluster structure of the data was modeled with three random effects. First, the different technical replicates were nested within the appropriate cell donors, which were modeled by two nested random effects. Furthermore all technical replicates were measured at several regions of interests (ROI), therefore additionally a crossed random effect for ROI was included. Across the in-vitro settings only the fixed effects/predictors varied and were appropriately included as main effect and interaction term (S2 Table). As outcome the change of the scratch wound size as difference from the initial value was modeled to depict for time effects. To control potential bias the baseline size was included as confounder.
For comparison of cytokine expression levels ANOVA with Bonferoni post hoc test was used.
The analyses were conducted with StataCorp. 2015. Stata Statistical Software: Release 14. College Station, TX: StataCorp LP, or SPSS version 22.
Results
Diabetic versus non-diabetic cells
At first we tested adult diabetic versus non-diabetic keratinocytes under euglycaemic conditions to investigate the usability of these cells in early drug screening for diabetic wound healing. We observed a trend of decreased scratch wound closure rate for cells of diabetic origin versus non-diabetic origin in conventional scratch assay (Fig 1A); but this was not the case in a high-throughput, semi-automated system (S1A and S1B Fig).
Fig 1. Comparison of scratch wound healing of human primary keratinocytes of various origins and under various culture conditions.
Closed wound area per visual field (mm2) of (A) diabetic versus non-diabetic keratinocytes at 4, 8, 12, 24 and 36 h after wounding using K-SFM medium. n = 4 in duplicates (non-diabetic keratinocytes) and n = 3 in duplicates (diabetic keratinocytes) (B) Keratinocytes from juvenile donors under euglycaemic (6 mM glucose), hyperglycaemic (25 mM glucose) and hyperosmolar (6 mM glucose + 19 mM mannitol) conditions at 12 and 24 h after wounding in K-SFM, Dermalife and Epilife medium. n = 4 in duplicates. (C) Keratinocytes from adult, non-diabetic donors under euglycaemic (6 mM) and hyperglycaemic (25 mM) conditions 4, 8, 12, 24 and 36 h after wounding in K-SFM medium. n = 4 in duplicates. (D) Keratinocytes from juvenile donors under euglycaemic, hyperglycaemic and hyperosmolar conditions at 12 and 24 h after wounding in K-SFM without supplements. n = 4 in duplicates; mean ± SEM; *: statistically significant with p < 0.05.
Non-diabetic cells under hyperglycaemic versus euglycaemic conditions
It was shown by Lan et al. [12, 13] that high glucose conditions result in decreased scratch wound closure in non-diabetic juvenile keratinocytes. We therefore investigated juvenile (foreskin) keratinocytes, as a well understood and frequently applied cellular system, and adult keratinocytes, as a system closer to the real scenario of elderly patients with diabetes mellitus type 2, whether we can recapitulate this finding. We observed that hyperglycaemic conditions significantly (p<0.05) decreased scratch WH in juvenile keratinocytes (Fig 1B; 12 h: closed wound area under euglycaemic conditions: 0.47 ± 0.06 mm2, closed wound area under hyperglycaemic conditions: 0.31 ± 0.07 mm2, i.e. 66 ± 14% of control; 24 h: closed wound area under euglycaemic conditions: 0.63 ± 0.06 mm2, closed wound area under hyperglycaemic conditions: 0.48 ± 0.11 mm2, i.e. 68 ± 14% of control) and adult, non-diabetic keratinocytes (Fig 1C; 8 h: euglycaemic conditions: 0.37 ± 0.07 mm2, hyperglycaemic conditions: 0.30 ± 0.02 mm2, i.e. 85 ± 7% of control; 12 h: euglycaemic conditions: 0.48 ± 0.07 mm2, hyperglycaemic conditions: 0.42 ± 0.03 mm2, i.e. 87 ± 7% of control). The decrease was also seen with the semi-automated system (S1D Fig). However, of interest, a decrease of wound closure rate was only seen in K-SFM, but not in Dermalife or Epilife medium (Fig 1B). Thus, the choice of medium is of importance. Due to these results, K-SFM medium was chosen for the further experiments. In addition, decrease was seen in juvenile cells at 12 h and 24 h after scratch, but in adult cells only after 12 h, and it was less pronounced (Fig 1B and 1C). Thus, also donor age plays a role.
As the main differences between the media are the supplements added to the basal media, we tested the influence of the supplements present in K-SFM on the hyperglycaemic effect. We observed that while hyperglycaemic conditions resulted in significantly reduced WH in medium containing supplements (Fig 1B), this was not the case in medium without supplements (Fig 1D).
In addition, we investigated whether the high glucose effect can be mimicked by high osmolarity. Indeed, the use of mannitol instead of glucose resulted in similarly delayed wound closure rates in non-diabetic keratinocytes (Fig 1B; 12 h: euglycaemic conditions: 0.47 ± 0.06 mm2, hyperosmolaric conditions: 0.31 ± 0.08 mm2, i.e. 66 ± 17% of control; 24 h: euglycaemic conditions: 0.63 ± 0.06 mm2, hyperosmotic conditions: 0.39 ± 0.08 mm2, i.e. 63 ± 15% of control). Again, this effect was only seen in cells cultured with supplements (Fig 1B and 1D).
Ex-vivo WH model under hyperglycaemic versus euglycaemic conditions
As shown above, the choice of medium and the presence of supplements in the medium had a remarkable influence on the effect of hyperglycaemic conditions on the rate of scratch wound closure. In an in-vivo wound, these factors may be provided by soluble substances produced by other cell types. Therefore, we established a more complex high glucose porcine ex-vivo WH model that contains all essential cells residing in the skin. This model is based on our previously described euglycaemic porcine ex-vivo WH model[16, 18, 26]. A significant (p<0.05) decrease in length of regenerated epidermis from 399 ± 15 μm under euglycaemic conditions to 263 ± 15 μm under hyperglycaemic conditions, i.e. to 66 ± 4% of control was observed when using 50 mM glucose conditions and preincubation before wounding the tissue (Fig 2A). This was not the case without preincubation (Fig 2A) or at a 25 mM glucose concentration with or without preincubation (S2 Fig). Mimicking the high osmotic conditions by substituting 44 mM glucose by mannitol resulted also in a decrease of WH in the ex-vivo model but to a lesser extent than in the models under hyperglycaemic conditions (to 322 ± 38 μm, i.e. 81 ± 9% of control) and in a non-significant manner (Fig 2A). Of note, high glucose conditions also resulted in a yellowish color of the skin which was not the case in the hyperosmolar models (Fig 2B).
Fig 2. Ex-vivo porcine wound healing models cultured under euglycaemic, hyperglycaemic and hyperosmolar conditions.
(A) Length of the regenerated epidermis for the experimental settings with and without preincubation for 48 h prior to wounding under euglycaemic (6 mM glucose), hyperglycaemic (50 mM glucose) and hyperosmolar (6 mM glucose + 44 mM mannitol) conditions 48 h after wounding. n = 8 in triplicates; mean ± SEM; *: statistically significant with p < 0.05; (B) Yellow coloring of the skin biopsies cultured under hyperglycaemic conditions for a total of 96 h (48 h preincubation and 48 h after wounding).
Influence of TE on in-vitro and ex-vivo wound healing
Finally, we tested the effect of TE and its main constituent, betulin, on its WH capacity in the eu- and hyperglycaemic and diabetic in-vitro and the hyperglycaemic ex-vivo assay systems. It was shown before, that this extract, approved by the European Medicines Agency, exhibited a positive effect on non-diabetic in-vivo WH [23, 28] as well as on ex-vivo WH models under euglycaemic conditions[22, 24].
However, TE did not show a beneficial effect in any of the in-vitro tests (juvenile keratinocytes, adult keratinocytes, euglycaemic conditions, hyperglycaemic conditions; see S2 File and S3, S4 and S5 Figs). This was also true for betulin except for a slight but significant scratch WH accelerating effect 24 h after scratch wounding (from 0.50 ± 0.02 mm2 to 0.53 ± 0.02 mm2, 106 ± 4.4% of control) (see S2 File and S3, S4 and S5 Figs). There was also no beneficial influence of TE on scratch wound healing in fibroblasts (S6 Fig).
The non-existing or even negative effect of TE and betulin on in-vitro WH under euglycaemic conditions was surprising because the extract positively influences WH in-vivo and ex-vivo[22–24]. Therefore, we concluded that this beneficial effect is dependent on the interplay with other cells. To evaluate whether this might also be the case for diabetic WH, we tested TE in two different formulations in our hyperglycaemic ex-vivo model. We used an oleogel identical to the one used in-vivo / ex-vivo for non-diabetic wounds [22–24] and a water-in-oil emulsion. The latter was used because it is the basis of a more advanced formulation concept which will finally lead to TE-containing foams, which can be applied to wounds almost touchless. However, the addition of a water phase can influence the availability of the TE and has thus to be studied in addition to the oleogel. Indeed, we observed a significantly accelerating influence of both formulations on epidermal regeneration 48 h after wounding in ex-vivo wounds under hyperglycaemic conditions (oleogel: 326 ± 28 μm compared to 173 ± 23 in PBS treated controls, that means 188 ± 16% of control, w/o emulsion: 317 ± 20 μm, that means 183 ± 11% of control) (Fig 3). The positive control, PDGF, which was shown before to be effective in in-vivo diabetic wounds [29] also showed a significantly positive effect (358 ± 45 μm, that means 206 ± 26% of control) (Fig 3). There was also a significantly accelerating effect of the TE W/O emulsion (284 ± 19 μm compared to 195 ± 18 μm, that means 146 ± 10% of control) and PDGF (312 ± 39 μm, that means 160 ± 20% of control) on models under hyperosmolaric conditions, but this was less pronounced than for the models under hyperglycaemic conditions (Fig 3). For TE oleogel, no significant effect on models under hyperosmolaric conditions was observed (Fig 3).
Fig 3. Influence of TE on reepithelialization of hyperglycaemic and hyperosmolaric ex-vivo models.
Length of the regenerated epidermis of porcine ex-vivo wound healing models that were treated with TE oleogel, W/O emulsion (60%) containing 10% TE oleogel, PDGF (120 ng/ml) or PBS directly after wounding for 48 h under hyperglycaemic or hyperosmolaric conditions (including preincubation); n = 16 in duplicates; mean ± SEM; *: statistically significant with p < 0.05.
Influence of TE on cytokine expression in fibroblasts
One important difference between the in-vitro and the ex-vivo test system is the presence of additional cell types which might also influence WH, e.g. via cytokines. As the major second type of cells in the skin is fibroblasts, we tested whether TE might induce the expression of cytokines in fibroblasts under different glucose conditions. We investigated cytokines which have previously been shown to be induced by TE in fibroblasts under normal culture conditions [30] and which are important for wound healing[31]. We observed a significant increase of IL-6, IL-8, IL-1 ß, and CCL-2 expression in fibroblasts under various glucose concentrations (Table 2).
Table 2. Influence of TE on cytokine expression under various glucose concentrations (n = 4).
| DMSO control | TE 1 μg/ml | TE 10 μg/ml | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Glucose Conc. | normal | 25 mM | 50 mM | normal | 25 mM | 50 mM | normal | 25 mM | 50 mM |
| IL-6 | 0.7 | 1.3 | 0.9 | 13.7 | 7.8 | 14.1 | 143.4 | 58.4 | 184.5 |
| *** | *** | ** | *** | *** | *** | ||||
| IL-8 | 2.7 | 1.8 | 0.7 | 152.4 | 23.6 | 17.6 | 1013.3 | 115.5 | 75.4 |
| P = 0.055 | *** | * | ** | *** | ** | ||||
| IL-1beta | 2.0 | 1.4 | 0.8 | 2.2 | 5.5 | 3.5 | 6.2 | 5.0 | 7.0 |
| * | * | * | ** | ||||||
| CCL2 | 2.5 | 1.9 | 1.0 | 16.9 | 4.8 | 4.7 | 28.1 | 5.2 | 7.7* |
| * | ** | * | |||||||
| FGF2 | 0.9 | 1.0 | 0.7 | 0.9 | 1.1 | 1.0 | 1.7 | 1.3 | 1.3 |
| *** | * | ||||||||
| FGF7 | 0.6 | 0.7 | 0.9 | 0.7 | 1.1 | 1.0 | 1.6 | 0.8 | 1.5 |
| *** | |||||||||
Values denote fold increase/decrease compared to untreated control cells with the same glucose conditions.
*: p<0.05;
** p<0.01;
*** p<0.001 compared to DMSO controls with the same glucose conditions.
Discussion
Diabetic keratinocytes as a putative screening system
We did not find significantly decreased in-vitro wound closure in cells from diabetic origin, even though in-vivo delayed WH is a hallmark of diabetes type 2[32]. This might be due to the high variability of wound closure rates of the diabetic cells and might reflect different adjustment of blood glucose levels and HbA1c values. But also the non-diabetic controls showed a certain level of variability. Thus, cells of diabetic skin, even though valuable because they comprise—at least in early passages [20]—the diabetic history of the donors, may only give reliable results for drug screening when using a high number of donors which is in practice often difficult to achieve due to limited access to donor tissues.
High glucose conditions decrease in-vitro WH but only in a particular medium
While some authors [11–13] observed delayed scratch WH in juvenile keratinocytes cultured under hyperglycaemic conditions, this was not the case in experiments from others [14] even though using the same amount of glucose. However, when looking into detail, the authors used different media. Thus we compared different media and indeed found that high glucose only resulted in delayed WH in K-SFM—which was also used in [11–13]—and that this effect is dependent on the presence of supplements. Unfortunately, the companies are quite restrictive in elucidating the composition of supplements, so we can only speculate about the pivotal substances. For example, the concentration of epidermal growth factor (EGF) is 50 times higher in K-SFM than in Epilife (there is no information for Dermalife). EGF plays an important role during WH because it is necessary for the activation of keratinocyte migration[33]. Furthermore, it was shown that EGF causes a mild hyperglycaemia in mice[34]. In general, supplements reflect substances normally produced by other cell types present in the natural microenvironment of cells. Thus, this dependency on supplements underlines the importance of the interplay of several cell types and their products to mimic the in-vivo effect of delayed WH.
Decreased in-vitro WH due to hyperglycaemic conditions can be mimicked by hyperosmolaric conditions
The wound closure delaying effect observed under hyperglycaemic conditions could be a glucose effect, e.g. by the formation of advanced glycation end products (AGEs) [35] or by enhancing glucose metabolism in the cells including increased formation of reactive oxygen species[3, 36]. However, it could also be the result of an osmotic effect. It is known that hyperosmolar conditions can enhance the inflammatory response by the secretion of cytokines[37]. Additionally, hyperosmotic stress might be associated with insulin resistance in diabetes type 2 by inhibiting insulin-mediated signal transduction in adipocytes[38]. To determine the contributing factors, we substituted the elevated amount of glucose by mannitol. Indeed, we could show that the osmotic effect plays a major role. Thus, when using hyperglycaemic keratinocytes in scratch wound assays to mimic diabetic WH for drug screening, one has to keep in mind that mainly the hyperosmolaric aspect of hyperglycaemia plays a role and that only substances that influence the effect of hyperosmolarity will be detected.
Decreased WH in the ex-vivo WH model under hyperglycaemic conditions
The effect of high glucose on 2D keratinocyte cultures was significantly influenced by the microenvironment. The influence of the microenvironment in in-vivo wounds is given by other cell types and their interplay with keratinocytes. To better mimic this in-vivo situation, we established a porcine ex-vivo WH model that provides already intrinsically additional cell types. For this so-called “diabetic” WH model, an elevation to 25 mM glucose—which is used in 2D culture—is not sufficient. This might be due to the fact that glucose reaches the wound closing keratinocytes only via diffusion and thus the final concentration in the epidermis is lower. Of note, fibroblasts, which are the first cells in contact with glucose, are known to grow better with higher concentrations of glucose than 6 mM. Typical culture media for fibroblasts contain either 11.1 mM glucose (RPMI 1640 medium) or even 25 mM glucose (DMEM w/o L-glutamin). Thus, a negative effect on fibroblasts can only be achieved with higher concentrations. It is also necessary to preincubate the model for 48 h, which reflects that glucose needs some time to execute its effect. Of note, decreased WH in the ex-vivo model is not completely mimicked by hyperosmolaric conditions, suggesting that also typical glucose effects play a role. This is also supported by the fact that the hyperglycaemic model, but not the hyperosmolaric one, shows a yellowish color which may be due to AGEs as reported by Yokota and Tokudome[39]. Thus, with the ex-vivo model a hyperglycaemia-induced delay in WH comprising glucose-chemistry as well as osmolarity effects can be generated reproducibly. This model is based on porcine skin, thus tissue accessibility is not a problem and the model is not regarded as an animal experiment, as tissue from slaughtered animals is used. However, in comparison to in-vivo WH this model also has important limitations. There is no blood supply, no new invasion of immune cells and no long-term diabetic effects (e.g. neuropathy) which can only be mimicked by in-vivo models.
Beneficial effect of TE under hyperglycaemic conditions only in ex-vivo, but not in in-vitro models
As we knew about the beneficial effect of TE in-vivo and ex-vivo under euglycaemic conditions [22–24] and its effect on cytokine production and cytoskeleton in diabetic keratinocytes and fibroblasts [30] and because triterpenes are known to have beneficial effects in diabetes[40], we used TE as a test substance in our various in-vitro and ex-vivo models. Interestingly, it only showed a positive effect in the ex-vivo model. An explanation for the discrepancy between the in-vitro and the ex-vivo effects is probably the fact that ex-vivo many more cell types are involved and that their interplay may be important. Indeed we could show that fibroblasts under hyperglycaemic conditions release cytokines in response to TE treatment. These cytokines, in turn, may influence other dermal cells important for wound healing, including keratinocytes. But of course the influence of TE is likely not restricted to fibroblasts and keratinocytes, but can include further dermal cells like dermal immune cells or—at least in-vivo–immigrating extra-dermic cells[41].
Of note, we observed also under euglycaemic conditions no effect or even a long term negative effect in-vitro, in contrast to the positive effect observed ex-vivo and in-vivo[22–24]. This clearly supports that the ex-vivo model reflects better the in-vivo situation and is therefore more suitable for early drug screening. Thus, the positive effect of TE in the hyperglycaemic ex-vivo system may indeed hint for a beneficial effect of the birch bark extract in in-vivo diabetic WH.
Summarizing our data the “hyperglycaemic” ex-vivo model turns out to be the more reliable system for a first test of new active ingredients compared to the in-vitro scratch assay. Nonetheless, the in-vitro systems may also be useful to depict specific influences on isolated cell types and mechanisms.
In addition, one has to keep in mind that hyperglycemic conditions in vitro alone cannot replicate the complex damage that diabetes causes over time in-vivo. In-vitro assays as well as our porcine ex-vivo assay lack "chronicity" and do not take into account the broad effects of hyperglycemia on skin vascularity, immune response etc. Thus, in-vitro and ex-vivo diabetic wound models show clear limitations but can provide a first hint for the putative effectivity of an active ingredient which has to be further investigated in-vivo. A further advancement may be the use of ex-vivo diabetic skin, if accessibility is not a problem. This should be tested in a future study.
Supporting Information
Closed scratch wound area per visual field in classical scratch assay (A, C, E, G) and semi-automated scratch assay (B, D, F, H) of diabetic/non-diabetic keratinocytes (A, B), adult non-diabetic keratinocytes under eu- and hyperglycaemic conditions (C, D) and adult non-diabetic keratinocytes under euglycaemic (E, F) and hyperglycaemic (G, H) conditions treated with DMSO (1:1000 in medium), TE (1 μg/ml) or betulin (0.87 μg/ml) (E-H) Mean ± SEM; *: statistically significant with p < 0.05.
(DOCX)
Length of the regenerated epidermis of porcine ex-vivo models cultured in medium containing 6 mM or 25 mM glucose for 48 hours after wounding with and without preincubation for 48 h (n = 10 in triplicates; mean ± SEM).
(DOCX)
Closed scratch wound area per visual field of human primary keratinocytes from juvenile (A, B) as well as adult non-diabetic (C, D) and adult diabetic (E, F) donors, treated with DMSO (1:1000 in medium), TE (1 μg/ml) or betulin (0.87 μg/ml), under (A, C, E) euglycaemic (6 mM glucose) and (B, D, F) hyperglycaemic (25 mM glucose) conditions at the indicated time points after wounding (n = 4 in duplicates); mean ± SEM; *: statistically significant with p < 0.05.
(DOCX)
Closed scratch wound area per visual field of human primary keratinocytes from adult, non-diabetic donors, that were treated with DMSO (1:10000), TE (100 ng/ml) or betulin (87 ng/ml) under (A, C) euglycaemic (6 mM) and (B, D) hyperglycaemic (25 mM) conditions at 4, 8, 12, 24 and 36 hours after wounding. (A, B) conventional scratch assay and (C, D) semi-automatic system. (n = 4 in duplicates for the conventional scratch assay and n = 3 at least in duplicates for the semi-automatic system; mean ± SEM).
(DOCX)
Closed scratch wound area per visual field of human primary keratinocytes from adult, diabetic donors, that were treated with DMSO (1:10000), TE (100 ng/ml) or betulin (87 ng/ml), under (A) euglycaemic (6 mM) and (B) hyperglycaemic (25 mM) conditions at 4, 8, 12, 24 and 36 hours after wounding. (n = 4 at least in duplicates; mean ± SEM).
(DOCX)
Closed scratch wound area as % of control 12 hours after wounding. (n = 4 mean ± SD).
(DOCX)
(DOCX)
(DOCX)
(DOCX)
(DOCX)
Acknowledgments
We would like to thank Armin Scheffler, Tobias Zahn, Melanie Laszczyk (all Birken AG,Niefern-Oeschlbronn, Germany) for helpful discussions and Germar Schuering for image editing.
Data Availability
All relevant data are within the paper and its Supporting Information files.
Funding Statement
J.M.B., I. Merfort, and R.D. were supported by a grant of the German Ministry of Economics and Technology (KF239060SK1). Birken AG was also supported by this grant. In addition, J.M.B., I. Merfort, and R.D. got additional funding by Birken AG for this study. The funders were involved in study design, but not in data collection, analysis, decision to publish or preparation of the manuscript.
References
- 1.Singh N, Armstrong DG, Lipsky BA. Preventing foot ulcers in patients with diabetes. JAMA. 2005;293(2):217–28. 10.1001/jama.293.2.217 [DOI] [PubMed] [Google Scholar]
- 2.Broadley KN, Aquino AM, Hicks B, Ditesheim JA, McGee GS, Demetriou AA, et al. The diabetic rat as an impaired wound healing model: stimulatory effects of transforming growth factor-beta and basic fibroblast growth factor. Biotechnol Ther. 1989;1(1):55–68. [PubMed] [Google Scholar]
- 3.Lan CC, Wu CS, Huang SM, Wu IH, Chen GS. High-glucose environment enhanced oxidative stress and increased interleukin-8 secretion from keratinocytes: new insights into impaired diabetic wound healing. Diabetes. 2013;62(7):2530–8. 10.2337/db12-1714 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Sullivan SR, Underwood RA, Gibran NS, Sigle RO, Usui ML, Carter WG, et al. Validation of a model for the study of multiple wounds in the diabetic mouse (db/db). Plast Reconstr Surg. 2004;113(3):953–60. [DOI] [PubMed] [Google Scholar]
- 5.Nuutila K, Katayama S, Vuola J, Kankuri E. Human Wound-Healing Research: Issues and Perspectives for Studies Using Wide-Scale Analytic Platforms. Adv Wound Care (New Rochelle). 2014;3(3):264–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Justus CR, Leffler N, Ruiz-Echevarria M, Yang LV. In vitro cell migration and invasion assays. J Vis Exp. 2014;(88). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Kramer N, Walzl A, Unger C, Rosner M, Krupitza G, Hengstschlager M, et al. In vitro cell migration and invasion assays. Mutat Res. 2013;752(1):10–24. 10.1016/j.mrrev.2012.08.001 [DOI] [PubMed] [Google Scholar]
- 8.Faries PL, Rohan DI, Takahara H, Wyers MC, Contreras MA, Quist WC, et al. Human vascular smooth muscle cells of diabetic origin exhibit increased proliferation, adhesion, and migration. J Vasc Surg. 2001;33(3):601–7. 10.1067/mva.2001.111806 [DOI] [PubMed] [Google Scholar]
- 9.Grant MB, Guay C. Plasminogen activator production by human retinal endothelial cells of nondiabetic and diabetic origin. Invest Ophthalmol Vis Sci. 1991;32(1):53–64. [PubMed] [Google Scholar]
- 10.Roberts AC, Gohil J, Hudson L, Connolly K, Warburton P, Suman R, et al. Aberrant phenotype in human endothelial cells of diabetic origin: implications for saphenous vein graft failure? J Diabetes Res. 2015;2015:409432 10.1155/2015/409432 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Pan F, Guo R, Cheng W, Chai L, Wang W, Cao C, et al. High glucose inhibits ClC-2 chloride channels and attenuates cell migration of rat keratinocytes. Drug Des Devel Ther. 2015;9:4779–91. 10.2147/DDDT.S84628 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Lan CC, Liu IH, Fang AH, Wen CH, Wu CS. Hyperglycaemic conditions decrease cultured keratinocyte mobility: implications for impaired wound healing in patients with diabetes. Br J Dermatol. 2008;159(5):1103–15. 10.1111/j.1365-2133.2008.08789.x [DOI] [PubMed] [Google Scholar]
- 13.Lan CC, Wu CS, Kuo HY, Huang SM, Chen GS. Hyperglycaemic conditions hamper keratinocyte locomotion via sequential inhibition of distinct pathways: new insights on poor wound closure in patients with diabetes. Br J Dermatol. 2009;160(6):1206–14. 10.1111/j.1365-2133.2009.09089.x [DOI] [PubMed] [Google Scholar]
- 14.Wright CS, Berends RF, Flint DJ, Martin PE. Cell motility in models of wounded human skin is improved by Gap27 despite raised glucose, insulin and IGFBP-5. Exp Cell Res. 2013;319(4):390–401. 10.1016/j.yexcr.2012.12.013 [DOI] [PubMed] [Google Scholar]
- 15.Balaji S, LeSaint M, Bhattacharya SS, Moles C, Dhamija Y, Kidd M, et al. Adenoviral-mediated gene transfer of insulin-like growth factor 1 enhances wound healing and induces angiogenesis. J Surg Res. 2014;190(1):367–77. 10.1016/j.jss.2014.02.051 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Brandner JM, Houdek P, Quitschau T, Siemann-Harms U, Ohnemus U, Willhardt I, et al. An Ex-vivo model to evaluate dressings & drugs for wound healing. Example: Influence of lucilia sericata extracts on wound healing progress. EWMA J. 2006;6:11–5. [Google Scholar]
- 17.Brandner JM, Kief S, Grund C, Rendl M, Houdek P, Kuhn C, et al. Organization and formation of the tight junction system in human epidermis and cultured keratinocytes. Eur J Cell Biol. 2002;81(5):253–63. 10.1078/0171-9335-00244 [DOI] [PubMed] [Google Scholar]
- 18.Brandner JM, Zacheja S, Houdek P, Moll I, Lobmann R. Expression of matrix metalloproteinases, cytokines, and connexins in diabetic and nondiabetic human keratinocytes before and after transplantation into an ex vivo wound-healing model. Diabetes Care. 2008;31(1):114–20. 10.2337/dc07-1304 [DOI] [PubMed] [Google Scholar]
- 19.Moll I, Houdek P, Schmidt H, Moll R. Characterization of epidermal wound healing in a human skin organ culture model: acceleration by transplanted keratinocytes. J Invest Dermatol. 1998;111(2):251–8. 10.1046/j.1523-1747.1998.00265.x [DOI] [PubMed] [Google Scholar]
- 20.Pollok S, Pfeiffer AC, Lobmann R, Wright CS, Moll I, Martin PE, et al. Connexin 43 mimetic peptide Gap27 reveals potential differences in the role of Cx43 in wound repair between diabetic and non-diabetic cells. J Cell Mol Med. 2011;15(4):861–73. 10.1111/j.1582-4934.2010.01057.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Stojadinovic O, Tomic-Canic M. Human ex vivo wound healing model. Methods Mol Biol. 2013;1037:255–64. 10.1007/978-1-62703-505-7_14 [DOI] [PubMed] [Google Scholar]
- 22.Ebeling S, Naumann K, Pollok S, Wardecki T, Vidal YSS, Nascimento JM, et al. From a traditional medicinal plant to a rational drug: understanding the clinically proven wound healing efficacy of birch bark extract. PLoS One. 2014;9(1):e86147 10.1371/journal.pone.0086147 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Metelmann HR, Brandner JM, Schumann H, Bross F, Fimmers R, Bottger K, et al. Accelerated reepithelialization by triterpenes: proof of concept in the healing of surgical skin lesions. Skin Pharmacol Physiol. 2015;28(1):1–11. 10.1159/000357501 [DOI] [PubMed] [Google Scholar]
- 24.Steinbrenner I, Houdek P, Pollok S, Brandner JM, Daniels R. Influence of the Oil Phase and Topical Formulation on the Wound Healing Ability of a Birch Bark Dry Extract. PLoS One. 2016;11(5):e0155582 10.1371/journal.pone.0155582 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Geback T, Schulz MM, Koumoutsakos P, Detmar M. TScratch: a novel and simple software tool for automated analysis of monolayer wound healing assays. Biotechniques. 2009;46(4):265–74. 10.2144/000113083 [DOI] [PubMed] [Google Scholar]
- 26.Vockel M, Pollok S, Breitenbach U, Ridderbusch I, Kreienkamp HJ, Brandner JM. Somatostatin inhibits cell migration and reduces cell counts of human keratinocytes and delays epidermal wound healing in an ex vivo wound model. PLoS One. 2011;6(5):e19740 10.1371/journal.pone.0019740 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods. 2001;25(4):402–8. 10.1006/meth.2001.1262 [DOI] [PubMed] [Google Scholar]
- 28.Metelmann HR, Brandner J, Schumann H, Bross F, Hoffmann M, Podmelle F. Accelerating the aesthetic benefit of wound healing by triterpene. J Craniomaxillofac Surg. 2012;40(5):e150–4. 10.1016/j.jcms.2011.07.020 [DOI] [PubMed] [Google Scholar]
- 29.Steed DL. Clinical evaluation of recombinant human platelet-derived growth factor for the treatment of lower extremity ulcers. Plast Reconstr Surg. 2006;117(7 Suppl):143S–9S; discussion 50S-51S. 10.1097/01.prs.0000222526.21512.4c [DOI] [PubMed] [Google Scholar]
- 30.Wardecki T, Werner P, Thomas M, Templin MF, Schmidt G, Brandner JM, et al. Influence of Birch Bark Triterpenes on Keratinocytes and Fibroblasts from Diabetic and Nondiabetic Donors. J Nat Prod. 2016;79(4):1112–23. 10.1021/acs.jnatprod.6b00027 [DOI] [PubMed] [Google Scholar]
- 31.Barrientos S, Stojadinovic O, Golinko MS, Brem H, Tomic-Canic M. Growth factors and cytokines in wound healing. Wound Repair Regen. 2008;16(5):585–601. 10.1111/j.1524-475X.2008.00410.x [DOI] [PubMed] [Google Scholar]
- 32.Baltzis D, Eleftheriadou I, Veves A. Pathogenesis and treatment of impaired wound healing in diabetes mellitus: new insights. Adv Ther. 2014;31(8):817–36. 10.1007/s12325-014-0140-x [DOI] [PubMed] [Google Scholar]
- 33.Werner S, Krieg T, Smola H. Keratinocyte-Fibroblast Interactions in Wound Healing. J Invest Dermatol. 2007;127(5):998–1008. 10.1038/sj.jid.5700786 [DOI] [PubMed] [Google Scholar]
- 34.Grau M, Tebar F, Ramirez I, Soley M. Epidermal growth factor administration decreases liver glycogen and causes mild hyperglycaemia in mice. Biochem J. 1996;315 (Pt 1):289–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Goh SY, Cooper ME. Clinical review: The role of advanced glycation end products in progression and complications of diabetes. J Clin Endocrinol Metab. 2008;93(4):1143–52. 10.1210/jc.2007-1817 [DOI] [PubMed] [Google Scholar]
- 36.Portugal-Cohen M, Kohen R. Exposure of human keratinocytes to ischemia, hyperglycemia and their combination induces oxidative stress via the enzymes inducible nitric oxide synthase and xanthine oxidase. J Dermatol Sci. 2009;55(2):82–90. 10.1016/j.jdermsci.2009.05.006 [DOI] [PubMed] [Google Scholar]
- 37.Brocker C, Thompson DC, Vasiliou V. The role of hyperosmotic stress in inflammation and disease. Biomol Concepts. 2012;3(4):345–64. 10.1515/bmc-2012-0001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Gual P, Gonzalez T, Gremeaux T, Barres R, Le Marchand-Brustel Y, Tanti JF. Hyperosmotic stress inhibits insulin receptor substrate-1 function by distinct mechanisms in 3T3-L1 adipocytes. J Biol Chem. 2003;278(29):26550–7. 10.1074/jbc.M212273200 [DOI] [PubMed] [Google Scholar]
- 39.Yokota M, Tokudome Y. Permeation of Hydrophilic Molecules across Glycated Skin Is Differentially Regulated by the Stratum Corneum and Epidermis-Dermis. Biol Pharm Bull. 2015;38(9):1383–8. 10.1248/bpb.b15-00372 [DOI] [PubMed] [Google Scholar]
- 40.Nazaruk J, Borzym-Kluczyk M. The role of triterpenes in the management of diabetes mellitus and its complications. Phytochem Rev. 2015;14(4):675–90. 10.1007/s11101-014-9369-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Fiorina P, Pietramaggiori G, Scherer SS, Jurewicz M, Mathews JC, Vergani A, et al. The mobilization and effect of endogenous bone marrow progenitor cells in diabetic wound healing. Cell Transplant. 2010;19(11):1369–81. 10.3727/096368910X514288 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Closed scratch wound area per visual field in classical scratch assay (A, C, E, G) and semi-automated scratch assay (B, D, F, H) of diabetic/non-diabetic keratinocytes (A, B), adult non-diabetic keratinocytes under eu- and hyperglycaemic conditions (C, D) and adult non-diabetic keratinocytes under euglycaemic (E, F) and hyperglycaemic (G, H) conditions treated with DMSO (1:1000 in medium), TE (1 μg/ml) or betulin (0.87 μg/ml) (E-H) Mean ± SEM; *: statistically significant with p < 0.05.
(DOCX)
Length of the regenerated epidermis of porcine ex-vivo models cultured in medium containing 6 mM or 25 mM glucose for 48 hours after wounding with and without preincubation for 48 h (n = 10 in triplicates; mean ± SEM).
(DOCX)
Closed scratch wound area per visual field of human primary keratinocytes from juvenile (A, B) as well as adult non-diabetic (C, D) and adult diabetic (E, F) donors, treated with DMSO (1:1000 in medium), TE (1 μg/ml) or betulin (0.87 μg/ml), under (A, C, E) euglycaemic (6 mM glucose) and (B, D, F) hyperglycaemic (25 mM glucose) conditions at the indicated time points after wounding (n = 4 in duplicates); mean ± SEM; *: statistically significant with p < 0.05.
(DOCX)
Closed scratch wound area per visual field of human primary keratinocytes from adult, non-diabetic donors, that were treated with DMSO (1:10000), TE (100 ng/ml) or betulin (87 ng/ml) under (A, C) euglycaemic (6 mM) and (B, D) hyperglycaemic (25 mM) conditions at 4, 8, 12, 24 and 36 hours after wounding. (A, B) conventional scratch assay and (C, D) semi-automatic system. (n = 4 in duplicates for the conventional scratch assay and n = 3 at least in duplicates for the semi-automatic system; mean ± SEM).
(DOCX)
Closed scratch wound area per visual field of human primary keratinocytes from adult, diabetic donors, that were treated with DMSO (1:10000), TE (100 ng/ml) or betulin (87 ng/ml), under (A) euglycaemic (6 mM) and (B) hyperglycaemic (25 mM) conditions at 4, 8, 12, 24 and 36 hours after wounding. (n = 4 at least in duplicates; mean ± SEM).
(DOCX)
Closed scratch wound area as % of control 12 hours after wounding. (n = 4 mean ± SD).
(DOCX)
(DOCX)
(DOCX)
(DOCX)
(DOCX)
Data Availability Statement
All relevant data are within the paper and its Supporting Information files.