Introduction

Antimicrobial photodynamic therapy (APDT), which makes use of photoexcitation of a photosensitive drug to inactivate bacteria is receiving a lot of attention for the management of antibiotic-resistant bacteria infection in wounds [1–3]. PDT, in addition to the antibacterial effect, has also been shown to have influence on several other factors like bacterial toxins [4], proinflammatory cytokines [5], matrix metalloproteases, collagen synthesis, and remodeling [6] in wounds. Therefore, APDT appears promising for the treatment of diabetic wounds where healing is impaired and the treatment is further complicated by bacterial infection [7–9]. An important cause for impaired healing in diabetes is reduced angiogenesis [10, 11] which is primarily caused by reduced arteriogenic response of monocyte/macrophages in the wound [12]. Although, the enhancement in angiogenesis in tumor by PDT has been reported [13], so far, the effect of APDT on angiogenesis process of diabetic wound has not received attention.

In this study, we investigated the influence of poly-l-lysine-conjugated chlorin p6 (pl-cp6) mediated APDT on angiogenesis in wounds of diabetic mice. Chlorin p6, a cholorophyll derivative, is a promising photosensitizer as it has strong absorbance in the red (660 nm) region and good triplet yield [14]. Since cp6 is anionic, it was conjugated with cationic peptide poly-l-lysine to enhance the targeting to both gram-positive and gram-negative bacterial cells. Our previous studies on pl-cp6-mediated PDT have shown it to be a very efficient antimicrobial photodynamic agent [5, 6]. Effect of PDT on angiogenesis was evaluated by monitoring the levels of nitric oxide (NO) and vascular endothelial growth factor-A (VEGF-A). Also, effect of PDT on level of inflammatory stress markers, phosphorylated nuclear factor kaapa B (phospho-NF-kB-p65) and p38 mitogen-activated protein kinase (phospho-p38 MAPK) were monitored, since expression of these factors are influenced by diabetes [15, 16], bacteria infection [17], and ROS [18, 19]. Further, the efficacy of APDT has been compared with that of aminoguanidine (AG), an inhibitor of advanced glycated end products which play an important role in the pathogenesis of diabetic complications [20, 21] and silver nitrate (AgNO3), an antibacterial agent [22].

Material and methods

Bacterial strain and growth conditions

The methicilin resistant Staphylococcus aureus (MRSA) strain used in this study (ATCC 43300) was maintained routinely by sub culturing in tryptone soya broth (TSB, Himedia, Mumbai, India). For experiments, a colony of the bacteria was inoculated in TSB and was grown aerobically for 18 h at 37 °C using a shaker incubator. For studies on wounds, optical density of the overnight culture of bacteria was measured at 600 nm using a spectrophotometer (Cintra 20, GBC) and diluted appropriately to achieve ~108 CFU/ml.

Induction of diabetes and establishment of infection in wounds

A total of 48 Swiss albino mice (male and female, 12 weeks) were used for all experiments. Diabetes was induced by injecting multiple doses of Streptozotocin (STZ; Sigma-Aldrich Chemicals, USA) intraperitoneally according to the protocol‚ described in [23]. The blood glucose concentration was measured by a commercial glucometer (One-Touch Horizon, Johnson and Johnson) for 3 months to ensure sustained hyperglycemia. Mice with blood glucose level of >250 mg/dl were considered as diabetic. The diabetic mice at 4 weeks post STZ treatment and the corresponding gender, age-matched nondiabetic mice were used for wound creation. For this, in mice anesthetized by intraperitoneal injection of Ketamine (80 mg/kg) and Xylazine (10 mg/kg; Sigma-Aldrich Chemicals, USA) cocktail, the dorsal skin was shaved, treated with a depilatory cream, and then cleaned with povidone-iodine solution followed by 70 % alcohol wipe. Single wound (~1.5 × 1 cm) was created at the back of each mouse by excising the skin down to panniculus carnosus, using sterile surgical scissors and forceps [5, 6].

For the development of infection in wounds, ~107 CFU of exponential phase MRSA were applied topically onto each wound of diabetic mice and bacteria were allowed to grow for 48 h (day 2 post-wounding (p.w.). Signs of infection like purulent discharge, redness, and swelling were observed on day 2 p.w., suggesting development of infection in wounds.

Photodynamic therapy and other treatment of wounds

Cp6 was prepared in house following the procedure described in [14], and the drug was conjugated with cp6 by carbodiimide coupling method described in [5]. All the mice with wounds were divided into following groups: (1) uninfected, (2) uninfected, PDT given once on day 2 p.w., (3) infected untreated control, (4) infected, PDT given at 24-h intervals during days 1 to 3 p.w. (multiple PDT), (5) infected AgNO3 and (6) infected AG. For carrying out PDT, 20 μl of 200 μM pl-cp6 was applied topically onto wounds. After 30 min of pl-cp6 application, wounds were exposed to red light (660 ± 25 nm) using light source LC-122A (Citek, USA) at a power density of ~100 mW/cm2, for 10 and 20 min, to achieve light fluence of ~60 and ~120 J/cm2, respectively. For the infected PDT (multiple) group, light fluence of ~60 J/cm2 was used.

AG was injected intraperitoneally at a dose of 100 mg/kg 1 day prior to wound creation and during days 1–3 p.w. [20]. In the AgNO3 group, 10 μl of 0.5 % AgNO3 [22] was applied onto wounds topically at 24-h intervals during days 1–3 p.w.

On day 3 p.w., animals (treated and untreated) were euthanized and wound tissues containing the scab, granulation tissue were harvested using scissors and forceps. The harvested wound tissues were either used for bacterial load analysis immediately or snap frozen in liquid nitrogen and stored at −80 °C for experiments requiring measurements on other biomarkers.

To determine bacterial load, wound tissues were homogenized (3500 rpm, 5 min) in phosphate-buffered saline (five times, w/v). The supernatant collected after centrifugation (10,000×g, 10 min) was plated on TSA agar plates. After incubation (24 h, 37 °C in dark), the number of colonies in each plate was counted. Bacterial load was determined by multiplying the number of colonies with dilution factor and volume of supernatant obtained during the tissue homogenization and was expressed as log CFU/wound [24].

Nitrite measurement in tissue lysate

For measurements on nitrite level, the tissues were first pulverized in liquid nitrogen with the help of a mortar/pestle. The tissue fragments were mixed (1:5, w/v) with 0.1 M chilled potassium phosphate buffer ( pH 7.4), homogenized (3500 rpm, 5 min), and then sonicated at ~20 kHz, 10 s for four times, with gap of 10 s (Model CPX 130, Cole Parmer, USA). The supernatant collected after centrifugation (10,000 rpm, 10 min) was used for measuring nitrite level using Griess reagent according to the protocol described in [25] with some modifications. Briefly, the tissue supernatant (50 μl) was mixed (1:1, v/v) with freshly prepared Griess reagent (0.1 % N-(1-naphthyl)ethylenediamine dihydrochloride, 1 % sulphanilamide, and 5 % phosphoric acid in a 1:1:1 ratio) and incubated for 30 min at 37 °C, in a 96-well microwell plate. Absorbance of the reaction product was measured at 545 nm using a microplate reader (Power Wave X340, Bio-Tek Instruments Inc, USA). The total protein content of the wound lysate was estimated by bicinchoninic acid assay [26], using bovine serum albumin as standard. The NO levels of samples were determined by comparing with a calibration curve prepared using known concentration of sodium nitrite and NO level was expressed as micrograms of nitrite per milligram of protein.

VEGF-A measurement in tissue lysate

VEGF-A in the tissue lysate was measured using ELISA kit (Sigma-Aldrich, USA). Briefly, the powdered tissues were mixed with lysis buffer provided along with the ELISA kit and kept on ice for 30 min. The lysates were sonicated (“Nitrite measurement in tissue lysate”) by keeping on ice and then centrifuged (5000 rpm, 10 min, 4 °C). The sample supernatant of each group was mixed (1:1, v/v) with a sample diluent buffer and incubated in the capture antibody-coated microwells provided with the kit, for 18 h at 4 °C. The assay was carried out according to the protocol described in the kit, and absorbance was recorded at 450 nm using a microplate reader. The tissue VEGF-A concentration was calculated from a standard curve prepared using the absorbance values of known concentrations of VEGF-A (10 pg–10 ng/ml), normalized with the protein content, and expressed as picograms per milligram of protein.

Estimation of stress signaling response markers in tissue lysate

The level of phospho-IKB-α, NF-kB p65, phospho-NF-kB p65, and phospho-p38 MAPK in the tissue lysate were estimated using multi-analyte ELISA kit (PATH SCAN ELISA Kit, Cell Signaling Technology, USA) following the protocol described by the manufacturer. The wound tissue lysate was prepared as described above (“Nitrite measurement in tissue lysate” and “VEGF-A measurement in tissue lysate”). The absorbance values of samples were normalized with protein content [26] and expressed as OD per milligram of tissue protein [5].

Determination of DNA content of wounds

The color reaction of deoxyribose with diphenylamine DNA content of the wound tissues were measured following the protocol described in ref. [25] with some modifications. Briefly, wound tissues were homogenized in 2 N HCl (1: 5, w/v), and the homogenate was incubated at 75 °C for 30 min in a water bath and then centrifuged (845 g, 20 min). The supernatant (0.5 ml) was mixed with 1 ml of diphenylamine reagent (1.5 g diphenylamine dissolved in 100 ml acetic acid and 1.5 ml sulfuric acid) and boiled in a water bath (95 °C) for 10 min. Absorbance of the colored product was measured at 600 nm in a spectrophotometer (Cintra 20, GBC Corp). The amount of DNA was determined by comparing calibration curve prepared using the absorbance of DPA and known concentrations (0.05–0.5 mg/ml) of calf thymus DNA (Bangalore Genei, India). The DNA concentration of the tissue lysates (mg/ml) obtained from the standard curve was normalized with the dry weight of tissue and expressed as milligrams of DNA per gram of tissue weight.

Wound closure

Mean wound closure time was determined as the time point (day) when the wound was devoid of any visible crust on the surface [5].

Statistical analysis

Data were analyzed and expressed as mean ± standard deviation. Statistical comparison between means was carried out using one-way ANOVA. To quantify the correlation between the different parameters, the regression coefficient R from the linear regression analysis was used. p < 0.05 were considered significant.

Results

Effect of APDT on total nitrite content of wounds

In Fig. 1a, we show the results on NO level of uninfected wounds of diabetic mice exposed to single APDT at fluence ~60 and 120 J/cm2. As shown in Fig. 1a, the level of NO in wounds treated with PDT at a light fluence of ~60 J/cm2 increased significantly (by ~1.5 fold), whereas in wounds exposed to higher light fluence (~120 J/cm2), their level decreased by ~1.5-fold. It should be noted that for the infected wounds, the multiple PDT protocol at lower light fluence (~60 J/cm2) was used to maximize the bacterial inactivation and to reduce bacterial regrowth and damage to the inflammatory and other cells of the wound tissue. The results in Fig. 1b show the comparison of NO levels in wounds of the APDT, AgNO3, and AG treatment groups. In MRSA-infected wounds, multiple APDT resulted in a significant increase in NO level. The level of NO in AgNO3 treated wounds did not show significant change as compared with the wounds of untreated control group, although there was a significant reduction in bacteria (Fig. 1b, inset).

Fig. 1
figure 1

a Effect of PDT fluence on NO level of wounds of diabetic mice on day 3 p.w. b Effect of topical APDT and other treatments on nitrite level of infected wounds of diabetic mice on day 3 p.w. Fig. 1b: inset; Bacterial load of wounds. Treatment frequencies for each group are denoted in parentheses, on x-axis. UI uninfected wounds, Inf infected wounds, UC untreated control, ND wound, nondiabetic mice. PDT (3) refers to multiple PDT fluence of ~60 J/cm2

Full size image

Effect of APDT on VEGF-A content of wounds

In Fig. 2a, we present the results on VEGF-A level in uninfected wounds of diabetic mice exposed to single APDT fluence of ~60 or 120 J/cm2. As seen in the figure, the level of VEGF-A in response to PDT was similar to that of the changes observed for NO level. In infected wounds subjected to multiple PDT, the level of VEGF-A was observed to increase by ~1.75-fold as compared with the untreated controls. However, in wounds treated with AgNO3 and AG (Fig. 2b), the VEGF-A level was significantly lower than for APDT. These results were also qualitatively similar to PDT response on NO level (Fig. 1).

Fig. 2
figure 2

a Effect of PDT fluence on VEGF-A level of wounds of diabetic mice on day 3 p.w. b Effect of topical APDT and other treatments on VEGF-A level of wounds of diabetic mice on day 3 p.w. Treatment frequencies for each group are denoted in parentheses, on x-axis. UI uninfected, Inf infected wounds, UC untreated control, ND wound, nondiabetic mice. PDT (3) refers to multiple PDT fluence of ~60 J/cm2

Full size image

Effect of PDT on NF-kB and p38 MAPK in wounds of diabetic mice

In Fig. 3, we show the results of our measurements on phospho-IKB-α, NF-kB p65, phospho-NF-kB p-65, and phospho-p38 MAPK levels in infected wounds of diabetic mice subjected to APDT, AG, and AgNO3 treatment. While the infected wounds showed significantly higher levels of NF-kB p65, phospho-NF-kB p-65, and phospho-p38 MAPK, exposure of these wounds to PDT led to their downregulation. Further, levels of these markers in wounds of the three treatment groups followed the order APDT < AG < AgNO3. In fact, it was observed that in the AgNO3 group, the level of inflammatory markers were higher than the levels found in the untreated control group.

Fig. 3
figure 3

Effect of APDT on biomarkers of stress signaling response of MRSA-infected wounds of diabetic mice on day 3 p.w. Treatment frequencies for each group are denoted in parentheses, on x-axis. UI uninfected wounds, Inf infected wounds, UC untreated control, ND wound, nondiabetic mice. PDT (3) refers to multiple PDT fluence of 60 J/cm2

Full size image

Effect of APDT on DNA and protein content

The results in Table 1 show that in MRSA-infected wounds of diabetic mice subjected to multiple APDT, the DNA and protein contents increased significantly (p < 0.05), compared with their untreated counterparts.

Table 1 Effect of PDT on DNA and protein contents of MRSA-infected wounds of diabetic mice on day 3 p.w.
Full size table

Correlation of wound NO, VEGF, and wound closure

Figure 4a shows a correlation between NO and VEGF-A levels in wound with wound closure. A good positive correlation (R = 0.8) between wound tissue NO and VEGF-A level (Fig. 4a) was observed. Also, the wound closure time decreased with increase in NO level (Fig. 4b).

Fig. 4
figure 4

a Correlation of wound NO versus VEGF-A in diabetic mice (n = 9). b Effect of APDT-induced increased nitrite level on wound closure of MRSA-infected wounds of nondiabetic and diabetic mice. UI uninfected wounds, Inf infected wounds, UC untreated control, ND wound, nondiabetic mice. PDT (3) refers to multiple PDT fluence of 60 J/cm2

Full size image

Discussion

It is known that impairment of wound healing in diabetes is caused by several systemic and local abnormalities. One important factor is reduced angiogenesis due to microangiopathy and vasculogenesis [10–12]. Since previous studies have demonstrated that ROS produced in response to hypoxia or ischemia, play important role in angiogenesis [27], it was of interest to investigate whether APDT-induced oxidative stress modulates angiogenesis in wounds of diabetic mice.

The results presented in this study (Figs. 1b and 2b) show that level of both the angiogenic markers (NO, VEGF-A) in the wounds of diabetic mice are significantly lowered compared with that of nondiabetic mice. This may be due to impairment of NO production and inducible nitric oxide synthase (iNOS) activity due to hyperglycemia-induced depletion of NADPH [28, 29]. The enhancement in angiogenic markers (NO, VEGF-A) observed in diabetic mice in response to PDT (Figs. 1a and 2a), could occur due to increase iNOS [28–30], release of protein-bound NO [31], and activation of hypoxia inducible factor VEGF signaling pathway [27]. However, these responses were reversed at the high APDT fluence, possibly due to the higher oxidative stress-induced damage to inflammatory cells [32]. Our results are consistent with previous studies wherein it has been shown that PDT-induced angiogenic response in tumors is light fluence dependent [13]. Increase in blood vessel formation has also been reported during PDT [33, 34].

It is known that S. aureus infection can delay angiogenesis [35] by inducing prolonged presence of neutrophils [36, 37], apoptosis in macrophages [38], downregulated VEGF signaling [39], and reduced VEGF stability in the wound environment [35]. Therefore, the observations of APDT-induced increase in VEGF-A and NO levels in infected wounds could be due to inactivation of S. aureus toxins such as the alpha toxins and other proteases which would result in higher macrophage recruitment to the wound site. This increased NO, produced by macrophages, in a feedback loop would lead to more VEGF-A as well as contribute to macrophage accumulation at the wound site. It is also important to note that in addition to increase in proangiogenic response, APDT also led to increase in DNA and protein contents (Table 1) of the infected wounds subjected to PDT, compared with their untreated controls. It is pertinent to note that this increase cannot be accounted for by the increased bacteria content. The protein content of a single bacterium is only 50–300 fg [40] and even bacterial load as high as 109 CFU/wound would contribute only ~0.3 mg increase in the total protein content of the wound. This is insignificant considering the magnitude of increase (~4–5 mg) in protein content observed for the infected wounds. Further, compared with untreated infected wounds, DNA and protein contents in the infected wounds subjected to PDT are significantly higher (Table 1), in spite of the ~1.5 log decrease in bacterial load (Fig. 1b, inset). Therefore, the increased DNA/protein contents of the wounds suggest increase in host cell proliferation. This is consistent with the faster closure of wounds in photodynamically treated group (Fig. 4b).

Further, the results show down regulation of NF-kB p65, phos-NF-kB p65 by PDT which is expected to be due to the destruction of bacteria and the moderate enhancement in NO [18]. NF-kB and p38 MAP kinase (MAPK) participate in signaling cascade controlling the cellular response to proinflammatory stress during hyperglycemia, bacteria infection, and overexpression of these proteins can lead to inhibition of wound closure [15] by various ways such as prolonged inflammation [10, 11, 15] by extending neutrophil half life [41], inhibition of angiogenesis [42, 43], upregulation of matrix metalloproteases (MMPs) [16], and bacterial invasion [41]. Therefore, downregulation of NF-kB and p38 MAPK levels of wounds, on day 3 p.w. can lead to beneficial response such as reduction in neutrophil number [41], increased angiogenesis [42, 43], inhibition of bacteria growth [41], and improved wound closure [15].

Our study also shows that among the three treatment groups, a reduction in inflammatory response is observed only in the APDT group; though both AG, AgNO3 treatments led to significant decrease in bacterial load. In fact, AgNO3 led to enhancement of inflammatory markers. This is in agreement with induction of acute inflammation by AgNO3 reported previously [44] and also because AgNO3 does not have any effect on bacterial toxins. Although AG is expected to reduce inflammation and promote wound healing by inhibiting AGE in diabetic wounds [20, 21], it does not have any bactericidal effect like APDT.

Conclusion

Results presented in this study suggest that APDT enhances angiogenic response in the bacteria-infected wounds. Further, the results presented suggest that antibacterial effect of PDT-induced abrogation of hyperinflammatory response and enhancement of cell proliferation results in faster wound closure.