Paeonol was purchased from Sigma-Aldrich (St. Louis, MO, United States), and MTX was obtained from Minapharm Pharmaceuticals (Cairo, Egypt). Ready-to-use kidney injury molecule-1 (KIM-1; PA5-79345), nuclear factor-κB p65 subunit (NF-κB/p65; PA5-17264), and cleaved caspase 3 (PA5-23921) rabbit polyclonal antibodies were procured from Thermo Fisher Scientific (Waltham, MA, United States), while P-gp mouse monoclonal antibodies (sc-390883) were brought from Santa Cruz Biotechnology (Dallas, TX, United States). Other chemicals used were of analytical grade and were brought from their commercial sources.

Male Wistar rats weighing 180–220 g were acquired from the National Research Center (Giza, Egypt) and were caged under standard housing conditions (25°C ± 1 and 12-h light/dark cycle). Rats were offered ordinary chow and water ad libitum. The study protocol was performed in accordance with research ethical standards of the Faculty of Medicine-Research Ethics Committee, Minia University, Egypt (ethical approval No. 630/6/2020), which is consistent with EU directive 2010/63/EU. After a week of acclimatization, 24 rats were divided into four groups (n = 6). The first group served as the untreated control. Paeonol, suspended in 0.5% carboxymethyl cellulose solution, was given to the second group as a single daily oral dose of 100 mg/kg/day for 10 days (Al-Taher et al., 2020). The third group received MTX alone as a single i.p. dose of 20 mg/kg MTX (Morsy et al., 2020a) on the fifth day of the experiment, while the fourth group received a combination of MTX/paeonol regimen.

At the end of the 10th day, rat blood samples were collected and centrifuged at 5,000 rpm for 15 min for sera collection. Rats were euthanized, and their kidneys were excised. A longitudinal slice of one of each rat’s kidneys was fixed in 10% neutral-buffered formalin and embedded in paraffin, from which sections of 5 μm thickness were cut out and mounted on glass slides for histopathological and immunohistochemical staining. The rest of the kidneys were homogenized in 10% w/v ice-cold, 0.01 M, pH 7.4 phosphate buffer. After centrifugation of the homogenate at 4,000 rpm for 15 min, the supernatant of kidney homogenate was kept at −80°C until used for renal tissue biochemical measurements and real-time polymerase chain reaction (PCR).

Serum urea and creatinine as well as renal tissue reduced glutathione (GSH) were determined using commercially available kits (Biodiagnostic, Giza, Egypt). Renal superoxide dismutase (SOD) activity was evaluated via a method based on SOD inhibition of pyrogallol autoxidation (Marklund and Marklund, 1974). Briefly, kidney homogenates were mixed with Tris–HCl (pH 8.2) and pyrogallol (15 mM), and the absorbance of the sample was measured against blank at 420 nm over 3 min. The findings were presented as U/g tissue. The major lipid peroxidation product malondialdehyde (MDA) was measured as MDA–thiobarbituric acid adduct formed under high temperature and acidic conditions using 1,1,3,3-tetramethoxypropane as a standard. The findings were presented as nmol/g tissue (Buege and Aust, 1978). For the determination of nitric oxide (NO), the Griess method was used for estimation of the stable oxidation end products of NO (nitrites and nitrates), where the latter is reduced to the former by copperized cadmium granules, and the color change is revealed by the Griess reagent in an acidic medium. This color is then measured spectrophotometrically at 540 nm, and the findings are presented in nmol/g tissue (Sastry et al., 2002).

The renal tissue sections were stained with the hematoxylin and eosin (H&E) stain and examined under a light microscope (Olympus CX23LEDRFS1, Olympus, Tokyo, Japan). Morphometric estimation was performed using Leica QWin 500 image analysis software (Leica Microsystems, Wetzlar, Germany) for evaluation of the mean diameter of Malpighian renal corpuscles per section (Beshay et al., 2020). Other sections were stained with the periodic acid–Schiff (PAS) stain for examining the integrity of the glomerular membrane. For an immunohistochemical analysis, slides were stained with ready-to-use rabbit polyclonal primary antibodies of KIM-1, NF-κB/p65, and cleaved caspase 3 as well as P-gp, according to the manufacturer’s protocol. The mean area fraction of PAS staining, as well as that of immunohistochemical stains, was calculated via applying 10 non-overlapping fields for each group using ImageJ (freeware; rsbweb.nih.gov/ij).

For determination of renal toll-like receptor 4 (TLR4) and interleukin (IL)-1β mRNA, total RNA was extracted from kidney tissue homogenate by the RiboZol Reagent (AMRESCO, Solon, OH, United States), as instructed by the manufacturer. The primer pairs used for TLR4 were 5′-AAT​CCC​TGC​ATA​GAG​GTA​CTT​CCT​AAT-3' as the forward primer and 5′-CTC​AGA​TCT​AGG​TTC​TTG​GTT​GAA​TAA​G-3′ as the reverse primer (Al-Taher et al., 2020). For IL-1β, the forward primer was 5'-GTC​GTT​GCT​TGT​CTC​TCC​TTG​TA-3′ and the reverse primer was 5'-CAC​CTT​CTT​TTC​CTT​CAT​CTT​TG-3' (Leger et al., 2019). Applying 500 pg of RNA template per reaction, real-time PCR (Applied Biosystems 7500 Fast Real-Time PCR System, Foster City, CA, United States) was performed using a SensiFAST™ SYBR® Hi-ROX One-Step Kit (Bioline/Meridian Bioscience, Cincinnati, OH, United States) utilizing 70 nM of specific primers in 25 μL reaction volume. The data were quantified relative to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as a control gene (the forward primer was 5'-GTC​GGT​GTG​AAC​GGA​TTT​G-3′ and the reverse primer was 5'-CTT​GCC​GTG​GGT​AGA​GTC​AT-3′). The relative expression level of the studied genes was calculated using the comparative cycle threshold method (VanGuilder et al., 2008).

The P-gp specificity module (version 2016) of Percepta software (ACD/Labs, Toronto, Ontario, Canada) was used to predict P-gp specificity (P-gp substrates and/or inhibitors) from the chemical structure using a huge screening library. The Paeonol structure was drawn using ChemDraw software version 15 (PerkinElmer, Waltham, MA, United States).

Cellular cytotoxicity was performed by applying 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) on HCT-116 colon cancer cells in vitro, as previously described (Morsy et al., 2020b). Briefly, HCT-116 colon cancer cells were seeded in 96-well plates for 24-h incubation in DMEM at a concentration of 10⁴ cells/well. Cells were then incubated for 24 h with paeonol at concentrations of 0.01, 0.1, 1, 10, or 100 µM and with MTX at concentrations of 0.1, 1, or 10 µM. After which, 15 μL MTT reagent (5 mg/mL phosphate-buffered saline) was supplemented, followed by incubation for 4 h at 37°C in the dark. After that, 100 μL of dimethyl sulfoxide at 37°C was added to each well to dissolve formazan crystals. The color change was then estimated at 540 nm using a microplate reader.

The data were analyzed using GraphPad Prism software, version 6.01 for Windows (San Diego, CA, United States) and were presented as mean ± SEM. Statistical analysis was performed by one one-way analysis of variance (ANOVA) followed by Tukey’s post hoc analysis test for multiple comparisons. For MTT cytotoxicity studies, a non-linear regression analysis was performed. The findings were considered significant if p was less than 0.05.

At the end of the study, the levels of urea and creatinine were evaluated, and it showed that paeonol alone did not affect these kidney functional parameters, whereas MTX alone caused a significant increase in their levels compared to the control. Combined treatment with paeonol and MTX significantly prevented the effect compared to MTX alone. Similarly, paeonol alone had no significant effect on all tested oxidative stress markers, whereas MTX significantly decreased GSH and SOD while increasing MDA and NO levels. The combined paeonol/MTX-treated group however showed a significant increase in GSH and SOD as well as a significant decrease in MDA and NO compared to the group treated with MTX alone (Table 1).

Histological examination of the kidney, stained by H and E, showed a normal renal structure in both control and paeonol-treated groups; the Malpighian renal corpuscles contained glomeruli surrounded with Bowman’s space, which is clear of any cell debris. The parietal layer of Bowman’s capsule is lined by flat squamous cells. Proximal convoluted tubules appeared with narrow lumina, rounded basal vesicular nuclei, and apical clear brush borders, while distal convoluted tubules appeared with wider lumina, central rounded nuclei, and unclear brush borders (Figures 1A,B, respectively). On the other hand, the MTX-treated group displayed shrunken renal corpuscles leaving wide Bowman’s spaces. The renal tubules appeared dilated with intraluminal cellular debris. The proximal tubules were seen with an almost lost brush border and showed cytoplasmic vacuolations. The dilated congested blood vessels and the focal areas of inflammatory cell infiltration surrounding thick arterioles were also detected (Figure 1C). Animals treated with both paeonol and MTX showed normal Malpighian renal corpuscles containing apparently normal glomerulus and Bowman’s spaces. Most proximal convoluted tubules had regained their brush borders. Still, vacuolations were seen in a few tubular cells (Figure 1D). The structural changes in the renal histological picture were reflected morphometrically in the mean glomerular diameter (Figure 1E), where paeonol treatment had no effect, while MTX treatment caused nearly a 2-fold increase in the mean glomerular diameter compared to control. Co-administration of paeonol and MTX significantly decreased the mean glomerular diameter compared to MTX alone.

Using the PAS stain to highlight renal basement membranes showed that the control group had well-defined cell membranes (Figure 2A). Treatment with paeonol alone had no effect (Figure 2B), while MTX treatment caused significantly severe deterioration of the PAS-stained basement membrane (Figure 2C). Paeonol/MTX combined treatments significantly restored the integrity of the basement membrane (Figure 2D), as indicated with the analysis of the mean area fraction of PAS staining (Figure 2E).

To estimate the level of kidney injury, KIM-1 immunohistochemical expression was used, where both control and paeonol-treated groups had minimal KIM-1 expression (Figure 3A,B, respectively). MTX treatment, on the other hand, caused an apparent increase in KIM-1 nuclear expression (Figure 3C), which was decreased by combined treatment of paeonol with MTX compared to MTX alone (Figure 3D). This was statistically relevant as shown in the analysis of the percentage of KIM-1–positive cells (Figure 3E).

To estimate renal inflammatory status, TLR4 and IL-1β mRNA expressions (Figures 4A,B, respectively) as well as immunohistochemical expression of NF-κB (Figure 5) were examined. Paeonol treatment alone did not significantly affect any of the inflammatory markers tested. On the other hand, treatment with MTX alone significantly increased TLR4, NF-κB, and IL-1β compared to control. Co-administration of paeonol and MTX caused a significant decrease in the expression of all three markers compared to MTX alone. Using caspase 3 immunochemical staining as a marker for apoptosis, both control and paeonol-treated groups showed minimal expression of caspase 3 (Figures 6A,B, respectively). MTX treatment however significantly increased the expression of the apoptotic marker (Figure 6C) compared to control. Combining paeonol/MTX treatments caused a significant decrease in caspase 3 expression (Figure 6D) compared to the group receiving MTX alone. The significance of these results was reflected in the analysis of the percent of caspase 3–positive cells (Figure 6E).

Basal expression of P-gp was shown in the control group (Figure 7A). Interestingly, both paeonol and MTX drastically increased renal expression of P-gp (Figures 7B,C, respectively) compared to control, and their combined administration caused further increase in expression of the efflux protein (Figure 7D). Setting control at 100%, it was shown that the percent of positive cells in paeonol- or MTX-treated groups was nearly 20 folds higher than control (Figure 7E). The effect of the paeonol/MTX combination was additive as it caused an increment of nearly 40 folds higher P-gp expression compared to control.

As shown in Table 2, paeonol shows a low probability for being a P-gp substrate and/or inhibitor with moderate reliability.

To test the effect of administration of paeonol on MTX cytotoxicity, paeonol at concentrations ranging from 0.01 to 100 μM was administered to colon cancer cells together with MTX in a concentration either 0.1 (Figure 8A), 1 (Figure 8B), or 10 μM (Figure 8C). Interestingly, all tested paeonol concentrations when given in combination with MTX caused a dose-dependent progressive cytotoxic effect.