Early effect of laser irradiation in signaling pathways of diabetic rat submandibular salivary glands

Low-power laser irradiation (LPLI) is clinically used to modulate inflammation, proliferation and apoptosis. However, its molecular mechanisms are still not fully understood. This study aimed to describe the effects of LPLI upon inflammatory, apoptotic and proliferation markers in submandibular salivary glands (SMGs) in an experimental model of chronic disorder, 24h after one time irradiation. Diabetes was induced in rats by the injection of streptozotocin. After 29 days, these animals were treated with LPLI in the SMG area, and euthanized 24h after this irradiation. Treatment with LPLI significantly decreased diabetes-induced high mobility group box 1 (HMGB1) and tumor necrosis factor alpha (TNF-α) expression, while enhancing the activation of the transcriptional factor cAMP response element binding (CREB) protein. LPLI also reduced the expression of bax, a mitochondrial apoptotic marker, favoring the cell survival. These findings suggest that LPLI can hamper the state of chronic inflammation and favor homeostasis in diabetic rats SMGs.


Introduction
Photobiomodulation therapy (PBMT) using low-power laser irradiation (LPLI) is a promising treatment for inflammatory disorders and biomodulation processes. Clinically it shows great results in Sjögren syndrome, oral mucositis and rheumatoid arthritis treatment by its effects upon the biomodulation of the inflammation and tissue repair processes [1][2][3].
Molecular studies indicate that LPLI can decrease the expression of many inflammatory markers, as the high mobility group box 1 (HMGB1) and the tumor necrosis factor alpha (TNF-α) [4][5][6]. In vitro and in vivo studies also show the laser effects upon proliferation and apoptosis [5,7], increasing the expression of many growth factors [4].
Diabetes is a disease characterized by chronic hyperglycemia that results in damage in many organs [8]. It increases the formation of advanced glycation end-products (AGEs) [9,10], activating the receptor for advanced glycation end-products (RAGE) and self-sustaining experimental day, all animals were anesthetized with ketamine/xylazine solution (4 ml/kg IM), but only group D20 received the low-level laser irradiation, according to Simões et al. [23]. Twenty-four hours after the irradiation, all the animals were anesthetized, the 12h-fasting glycemia was measured, the SMG were removed and the animals were decapitated (S1 Fig). This animal experiment was performed in the same way as reported by Fukuoka et al. [5] All experiments were performed in accordance with the ARRIVE guidelines and were carried out in accordance with the U.K. Animals (Scientific Procedures) Act, 1986 and associated guidelines, EU Directive 2010/63/EU for animal experiments, and the National Institutes of Health guide for the care and use of Laboratory animals (NIH Publications No. 8023, revised 1978), approved by the Ethics Committee of Animal Experimentation of the University of São Paulo School of Dentistry, Brazil (Protocol 002/2019-CEUA).

Low-power laser irradiation procedure
For this procedure, first we anesthetized and made the trichotomy in the rats SMG area. After that, we drawn a circle with 1.13 cm 2 (above the area of both SMG together). Inside this area, LPLI was applied in 40 points, transcutaneous and perpendicular to the skin surface, in contact with it, but without pressure (S1 Fig).
The wavelength used was 660-nm continuous (70 mW; laser spot area of 0.028 cm 2 ; Photon Lase III, DMC IND. CO., Ltd., São Paulo, Brazil). Forty points of irradiation with 0.56 J, 8 seconds per point was performed at energy densities of 20 J/cm 2 , covering the SMG area, 1.13 cm 2 , being the total energy of 22.4 J and laser spot area of 0.028 cm 2 , as described by Simões et al. [23] modified.

SMG qRT-PCR analysis
After removal, SMG samples were kept in RNA Later (Life Technologies). Total RNA was extracted using RNeasy Mini Kit (Qiagen, Hilden, Germany) and the RNA concentrations were determined with Nanodrop (Thermo Fisher Scientific Inc., Paisley, PA, RU). Reverse transcription for cDNA synthesis was performed with 1 μg of each total RNA using Super Script VILO Master Mix (Life Technologies, CA, USA). One μl of each cDNA was subjected to real-time quantitative polymerase chain reaction (qPCR), using TaqMan Master Mix together with TaqMan Gene Expression Assays for HMGB1 (Rn02377062_g1), RAGE (Rn01525753_ g1), CREB5 (Rn01451883_m1) and β-actin (Rn01412977_g1) (Applied Biosystems, Framingham, MA, USA), performed on Step One Plus equipment (Thermo Fisher Scientific, MA, USA). Standard thermocycling conditions were used in a total volume of 10 μl. β-actin, was used as an endogenous control gene. Fold change in intensity of the target gene was calculated using Step One Software version 2.3 (https://www.thermofisher.com/br/en/home/technicalresources/software-downloads/StepOne-and-StepOnePlus-Real-Time-PCR-System.html) based on the control group.

Double immunofluorescence for phospho ERK and phospho CREB in SMG
Paraffin sections of the SMG (4-μm thick) were deparaffinized in xylene and rehydrated in graded alcohol series. Sections were immersed in citrate buffer pH 6.0 (Epitomics) at 97˚C for 20 min, washed in PBS and immersed in peroxidase blocking (DAKO) for 10 min. The samples were then washed in PBS and incubated in protein block (DAKO) for 1 h. The first antibody phosphorylated ERK (1:150, Cell Signaling Technology, Cat# 4370, RRID:AB_2315112) was prepared with Zenon Alexa Fluor 594 (Molecular Probes, Inc., Eugene, OR) according to manufacturer's instructions, and applied to the samples and incubated in a humidified dark chamber for 2 h. Then the sections were washed in PBS for 25 min and subsequently incubated with the second antibody phosphorylated CREB (1:150, Abcam Cat# ab32096, RRID: AB_731734) prepared with Zenon Alexa Fluor 488 for 2 h. DAPI was used for nuclear counterstaining.

SMG apoptosis analysis with terminal deoxynucleotidyl transferasemediated dUTP nick end-labelling (TUNEL) assay
The paraffin-embedded SMG tissues were sectioned and mounted on glass slides. The slides were de-paraffinized and dehydrated, treated with proteinase K 1:100 for 15 min at room temperature. Slides were then immersed in peroxidase blocking (DAKO) for 5 min, followed by washing with PBS. Next, slides were incubated in 1XTdT Labeling Buffer for 5 minutes, and covered with 50 μl of Labeling Reaction Mix at 37˚C for 1 h. Sections were then immersed in 1XTdT Stop Buffer for 5 min and washed in PBS. To visualize TUNEL-positive cells, slides were staining with 3, 3 0 -diaminobenzidine (DAB). The samples were later washed, counterstained with 1% methyl green for 5 min before washing. All these process were performed according to the manufacturer's instructions (Trevigen, Inc., Gaithersburg, MD, USA).
After the image acquisition, we quantified the positive staining by image colour deconvolution in Image J. First, the image was opened into the program and set the command: Analyze/ Set scale to calibrate all the images with the scale bar. Then we measure the whole area by selecting it and add the command: Analyze/Measure to obtain the total area. After that the image was deconvoluted by color: Plugins/Colour deconvolution/HDAB. Only the brown color image was selected and set to 8-bit (Image/Type/8-bit), the threshould was stablished (Image/ Adjust/Threshould) and the area with positive staining was calculated by Analyze/Analyze particles. We obtained the whole image area and the percentage of the area with positive staining, as described by Castro [35].

Enzyme-Linked Immunosorbent Assay (ELISA) for HMGB1 and TNF-α in SMG
The concentrations of HMGB1 and TNF-α in the protein samples extracted from SMGs, were evaluated using the HMGB1 Rat Elisa kit (MyBiosource, Inc., San Diego, CA) and Legend Max Rat TNF-α ELISA kit (Biolegend, Inc., San Diego, CA) respectively, according to the manufacturer's instructions. The total protein concentrations in SMGs samples were determined with BCA protein assay kit (Pierce Biotechnology, Rockford, Il, USA) following the protocol provided with the kit.

Statistical analysis
Data was presented as mean ± SEM. Statistical analysis was performed using ANOVA with Tukey. Differences were considered to be significant at P<0.05. Minitab 19 software (https:// www.minitab.com/pt-br/products/minitab/) was used for statistical calculations.

Metabolic parameters
At the end of the experiment, D0 animals have lost about 22% of their initial body weight and D20 animals lost 26% of their weight (P < 0.05) (S1 Table). No changes were observed in C0 group.
The initial fasting blood glucose was about 4 times higher in the diabetic groups (P < 0.05), in relation to the C0 group. The LPLI significantly reduced 22% of the diabetic rats glycemia (S1 Table).
Our findings showed increased HMGB1 protein levels in the SMG of D0 group in relation to C0 group, means of 15.06 ± 13.70 ng/mg of protein and 0.09 ± 0.02 ng/mg of protein for D0 and C0, respectively (P = 0.008). LPLI reduced these levels of HMGB1 in the diabetic rats in 96.60%, to the mean value of 0.50 ± 0.24 ng/mg of total protein in D20, similar to the control group (C0) (Fig 1B).
Although SMG morphology does not differ much between groups, diabetic animals SMG glands seems to have increased connective tissue areas with more space between acinus ( Fig  1C). HMGB1 was localized mainly in the nucleus of endothelial cells in the diabetic animals SMG, as well as, in the surrounding perivascular connective tissue ( Fig 1C-section 1E and 1F). The presence of HMGB1 staining was also observed in the nucleus of ducts cells, striated and excretory ducts, in all groups ( Fig 1C).

TNF-α expression in SMG
The levels of TNF-α in the diabetic animals (D0 group) showed an increase of 2.50 times in relation to control group (C0), means of 1.28 ± 0.60 pg/mg and 3.23 ± 1.20 pg/mg of total protein for C0 and D0, respectively (P = 0.038). After LPLI, these levels reduced 36.84% when compared to D0 group, mean of 2.04 ± 0.70 pg/mg in D20, similar to the values founded in the control group (Fig 2A).
In the immunohistochemistry analysis, staining for TNF-α was not detected in the C0 group (Fig 2B-section 2A). However, the presence of TNF-α was observed in some nucleus and cytoplasm of ducts cells, striated and excretory, as well as in the connective tissue in D0 group (Fig 2B-section 2B). A slight staining was founded in the cytoplasm of ducts cells in D20 group (Fig 2B-section 2C).

RAGE and phosphorylated NFκB protein expression in SMG
Diabetes significantly increased RAGE expression in SMG when compared to control group, mean of 0.11 ± 0.05 for C0, 0.66 ± 0.30 for D0 and 0.60 ± 0.26 for D20 (P = 0.001). There were no significant changes after LPLI (Fig 3A).
The localization of RAGE in SMG was observed in acinar and ducts cells, striated, excretory and convoluted granular tubules, as well as, in the perivascular connective tissue in all groups (Fig 3B). The activated p65 was founded in the nucleus of ducts and in the perivascular connective tissue, mainly stained in the groups D0 and D20 (Fig 3B-section 3E and 3F).

cAMP and ERK expression in SMG
Our results indicate that LPLI might induce cAMP expression, in acinus, ducts, endothelium and myofibroblasts cells, stained in D20 group (Fig 4B-section 4C). To further analyze expression of components of the cAMP signaling cascade, we checked ERK expression. ERK activation seems to be hampered in the diabetic submandibular glands (Fig 4A). In contrast, laser irradiation might enhance ERK phosphorylation, predominantly in ducts cells (Figs 4B-section 2F).

CREB expression in SMG
Our results suggest that diabetes does not interfere in the transcription of the cAMP responsive element binding protein 5, CREB5, in SMG, and LPLI did not change its mRNA levels expression, means of 1.20 ± 0.34, 2.44 ± 1.00 and 2.01 ± 0.79 for C0, D0 and D20, respectively (Fig 5A).
In spite of this finding, diabetes reduced the expression of phosphorylated CREB (Ser 133) to 10.95% of the values found in control, means of 2.19 ± 0.58 for C0 and 0.24 ± 0.11 for D0. In turn, LPLI significantly (P = 0.000) enhanced the activation of this transcriptional factor in SMG of the treated diabetic rats, mean of 0.99 ± 0.83 in D20 (Fig 5B).
Phosphorylated CREB could be observed in all cell types, acinus, ducts, endothelium, in control animals, but mainly in the nucleus of striated and excretory ducts cells, strongly stained in C0 and D20 groups (Fig 5C-section 5A and 5C).

CREB and ERK co-localization in SMG
Since phosphorylated CREB and ERK can interact with each other, we performed double immunofluorescence and found double positives cells in laser-irradiated diabetic submandibular glands, endothelium and ducts cells (S2G Fig and S2I Fig)
TUNEL assay shows positive-marked cells mainly in striated ducts, mainly in D0 samples (Fig 7B-section 7B). The quantification of apoptotic cells by image deconvolution indicates an increase in apoptotic cells in diabetic rats SMG, and a significant reduction after LPLI (P = 0.022), to levels similar to control group (Fig 7A).

Discussion
The current study investigated the early mechanism of action of LPLI in inflammation, apoptosis and proliferation pathways of salivary glands. As expected, diabetes significantly increased inflammatory markers such as, HMGB1 and TNF-α in SMG, as well as the receptor RAGE with a tendency of increase in the activation of the transcriptional factor NFκB. LPLI reduced HMGB1 and TNF-α expression to similar levels to the ones found in control group. These findings are in accordance to our previous study, where we suggested these effects by western blotting analysis for HMGB1 and immunohistochemistry for HMGB1 and TNF-α in diabetic rats SMG [5]. Interestingly, LPLI did not affect RAGE expression in SMG, and although it tended to decreased NFκB phosphorylation, this was not statistically significant. Other studies showed that laser therapy can modulate the inflammatory reaction by decreasing inflammatory infiltration and proteins as, TNF-α and INF-γ in mesenteric and periaortic lymph node cells culture [36], also it helped in the reduction of IL-6 and TNF-α levels in the serum of aging animals [37]. This is the first report to quantify its effects upon all these inflammatory markers in diabetic SMG glands.
https://doi.org/10.1371/journal.pone.0236727.g005  has been proposed to inhibit NF-κB activity through competition for limiting amounts of CBP/p300 p300 [39], since this interaction is necessary for NF-κB optimal activity at some target genes [40,41]. For the first time, we showed that the activation of CREB can be induced by LPLI in diabetic SMGs. CREB binds to a DNA target sequence named cAMP-responsive element (CRE) as a dimer, regulating diverse cellular responses, including cell proliferation, survival and differentiation [42]. Accumulation of cAMP can lead to CREB phosphorylation at Ser 133 via PKA activation [39]. Activation of PKA also mediates ERK1/2 phosphorylation [43,44]. Intriguingly, double immunofluorescence analysis reveals the same cellular localization for LPLI-induced expression of phosphorylated CREB and ERK in diabetic salivary glands, mainly in ducts and endothelium cells. It has been reported that activated ERK can translocate into nucleus and indirectly phosphorylate CREB at Ser133 [45]. Kwon et all, showed that both CREB and ERK were necessary for fibroblasts cells proliferation in vitro [46]). Phosphorylated ERK binding to the DNA target sequence SRE and phosphorylated CREB binding to CRE site were required for maximal c-fos promoter activation and proliferation [46]. Although our results did not analyze cell proliferation markers, LPLI can increase cell survival by reducing apoptosis in diabetic SMG, by reduction of bax expression. Since, the cell survival is mediated by the balance between the pro-apoptotic and anti-apoptotic markers, LPLI mechanism of action against inflammation induced apoptosis might not only involve the regulation of mitochondrial apoptotic markers but also the regulation of inflammatory markers, as HMGB1 and TNF-α that trigger this process.
Diabetes increased apoptosis in SMG ducts cells. In this study, we have demonstrated that diabetes increased the cleavage of caspase-3, probably by via up-regulation of TNF-α expression, since the levels of bax, an apoptotic marker of the intrinsic pathway, was similar to the ones founded in control animals. Increased expression of apoptotic cell death proteins was also reported in our previous study [5]. The pro-apoptotic markers, TNF-α and bax, were strongly repressed by LPLI in diabetic tissues, suggesting that the laser protective effects might be mediated by the regulation of mitochondrial apoptotic proteins. In accordance to our results, LPLI also suppressed TNF-α secretion in bone marrow stem cells under inflammatory conditions, and promote proliferation and osteogenesis in vitro [47]. LPLI also prevented apoptosis induced by TNF-α in endothelial cells, down-regulating caspase 3, 7, 8 and 9 expression and enhancing ERK phosphorylation [48].
This laser protective effect could be related to the ERK activation observed in the present study. Yan et al. showed that low-level laser irradiation induction of CREB phosphorylation is dependent on ERK, and ERK/CREB pathway were important for laser beneficial effects in neuronal survival [34].
ERK was also reported to inhibit caspase-9 via phosphorylation of T125 which inhibits activation of cleaved-caspase-3 and apoptosis, it is also involved in the modulation of the activity of many proteins involved in apoptosis [49].
Yun et al. also reported that laser therapy could not only up-regulate CREB gene expression to normal levels in rats hippocampus after injury, but also increase bcl-2, an anti-apoptotic protein, and decrease bax expression, preventing neurodegeneration [33]. LPLI also protected macrophages and muscle precursors cells from apoptosis after injury in gilts [50]. Its biomodulation effects were also reported in an animal experimental model of osteoarthritis treated with laser, where it significantly reduced caspase-3 expression [51]. In our study, a single time LPLI could not significantly reduce cleaved-caspase-3 expression in diabetic rats SMG, but did reduce apoptosis in this tissue. We hypothesized that LPLI might have an effect in non-caspase apoptotic pathways regulation.
In agreement with our previous findings [5,6], LPLI significantly reduced the hyperglycemia in diabetic rats. Our results suggest a mechanism of action of LPLI (Fig 8), where it could efficiently down-regulate diabetes-induced HMGB1 and TNF-α expression. Reducing the expression of bax as well, and preventing apoptosis in diabetic SMG. At the same time, LPLI up-regulated the activation of CREB pathway, partly through ERK, favoring cell survival. These findings indicate a promising therapeutic horizon for LPLI and deserves further studies to enlighten LPLI mechanisms of action and its effects in a long-term experiment.

Conclusion
LPLI can reduce inflammation and apoptosis in diabetic SMG, possibly by activating cAMP/ CREB signaling pathway, partially mediated by ERK. Diabetes significantly increased the inflammatory markers, HMGB1 and its ligand RAGE, and promoted apoptosis by upregulation of TNF-α expression and caspase-3 cleavage. Interestingly, LPLI strongly decreased HMGB1 and TNF-α accumulation in the tissue, also promoted the activation of the transcription factor CREB, by its phosphorylation on Ser 133, and reduced the expression of bax, therefore, modulating apoptosis and favoring the cell survival in diabetic SMG.

S2 Fig. Co-localization of phosphorylated CREB and phosphorylated ERK assessed by double immunofluorescence analysis in SMG.
In I, arrows points to double positive-stained striated and excretory ducts cells (n = 1/group). Scale bar, 50μM. ED, excretory duct; SD, striated duct. (TIF) S1 Table. LPLI decreased diabetic animals blood glucose levels. Body weight on the 1 st experimental day (initial) and in the 30 th day (final). Blood glucose levels were measured in the 3 rd experimental day (initial) and in the day of the sacrifice (final), after 12h-fasting (n = 8/group). Data are presented as mean ± SD; � P < 0.05 vs control group (C0); # P < 0.05 vs respective diabetic group (D0); ǂ P < 0.05 vs initial time in the same group. (DOCX) S1 Raw file. (ZIP)