Biochemical Analysis and Human Aldose Reductase Inhibition Activity of Selected Medicinal Plants of Nepal

Aldose reductase has received extensive research as a key enzyme in the development of long-term problems linked to diabetes mellitus. Overexpression of this enzyme or with exceeded glucose concentration in the blood increases sorbitol on the retina leading to retinopathy, which is the adverse efect of type II diabetes. Approximately 100 million people are sufering from diabetic retinopathy globally. Tis research is focused on studying the total phenolic content (TPC), total favonoid content (TFC), antioxidant potential, and aldose reductase inhibiting properties of selected medicinal plants such as Anacyclus pyrethrum, Bergenia ciliata, Rhododendron arboreum, and Swertia chirayita. In addition, ADMET analysis and molecular docking of seven previously identifed compounds from the chosen medicinal plants were carried out against human aldose reductase (PDB ID: 4JIR). Te ethanol extract of S. chirayita exhibited the highest TPC (4.63 ± 0.16mg GAE/g) and TFC (0.90 ± 0.06mg QE/g). Analysis of 2,2-diphenyl-1-picrylhydrazyl (DPPH)-based antioxidant assay showed that IC 50 of the ethanolic extract of B. cilata and R. arboreum showed a signifcant antioxidant activity with IC 50 value of 0.05mg/mL. Te percentage inhibition of AR by extract of B. ciliata (94.74 ± 0.01%) was higher than other plant extracts. A molecular docking study showed that morin isolated from B. ciliata showed a good binding interaction with AR. Tis study showed that the extracts of A. pyrethrum, B. ciliata, and R. arboreum could be potential sources of inhibitors against AR to treat retinopathy.


Introduction
Aldose reductase (AR) is an NADPH-dependent oxidoreductase that metabolizes the transformation of glucose to sorbitol in the polyol pathway [1]. In hyperglycemic states, an increased fux of glucose through the polyol pathway has been thought to harm tissues via diferent processes, including sorbitol accumulation leading to an osmotic imbalance [2] and pyridine nucleotide redox state dysregulation decreasing cellular antioxidant capacity [3], as well as a rise in advanced glycated end products [4,5]. In diabetes mellitus, the polyol pathway produces more sorbitol than normal, which does not easily difuse across the cell membranes, and the intracellular sorbitol accumulation has also been linked to the development of chronic complications of diabetes, including cataracts, neuropathy, and retinopathy ( Figure 1) [6][7][8]. A complication of diabetes called diabetic retinopathy (DR) is brought on by high blood sugar levels harming the retina. If not identifed and treated early, it might result in blindness. In adults aged 20 to 74, DR is the most common factor in new instances of blindness [9] and is a leading cause of preventable blindness among the working population of adults. Approximately 100 million adults worldwide sufer from DR and that number is expected to rise to 160.5 million by 2045 [10]. According to the Global Burden of Disease Study, DR is the ffth leading cause of blindness and vision impairment among adults aged 50 and older. Between 1990 and 2020, the agestandardized global prevalence of blindness caused by diabetic eye disease increased from 14.9% to 18.5% [9][10][11].
Inhibitors of AR, including epalrestat, alrestatin, zenarestat, ponalrestat, lidorestat, tolrestat, sorbinil, minalrestat, and fdarestat, were synthesized [12,13]. Te AR inhibitors come from two distinct chemical subgroups; sorbinil, Dilantin, and minalrestat are examples of hydantoin derivatives. Similarly, epalrestat, alrestatin, and tolrestat are examples of carboxylic acid derivatives [14]. Te only commercially accessible synthetic aldose reductase inhibitor is epalrestat [15]. Te adverse efects of epalrestat have caused it to be taken of the market in some countries, and all other inhibitors have failed in clinical studies due to poor pharmacokinetic qualities [16]. Terefore, it is necessary to investigate the natural sources to fnd safer therapeutic chemicals. Naturally occurring compounds and medicinal plant extracts exhibit AR inhibitory efects, and their preclinical ability to treat diabetes problems is promising [12]. For instance, fowers of the Rhododendron arboreum are of very high medicinal value. It is well known for its versatility and efcacy in treating various disorders, including eczema, diarrhea, menstrual problems, choleretic, diuretic, antispasmodic, and anti-infammatory disorder, and as it acts as an antioxidant [17]. Diferent pharmacological activities of Swertia chirayita include anthelmintic activities [18], hypoglycemic and antipyretic properties [19], antiviral activities [20], anticancer activities [21], and hypoglycemic and anti-infammatory activities [22]. Te medicinal plant Bergenia ciliata treats a variety of disorders, including diabetes, cancer, respiratory problems, diarrhea, fever, cough, vomiting, and is also used for wound healing [23]. Te roots of Anacyclus pyrethrum are advocated for use in folk medicine to treat a variety of ailments, including angina, digestive issues, female infertility, lethargy, and even paralysis of the tongue and limbs [24]. A. pyrethrum roots also exhibit sialagogue [25], aphrodisiac [26], immunostimulant [27], anti-infammatory [28], anticonvulsant, antioxidant [29], antidiabetic, and memory-improving efects [30]. Some computational analyses have been developed to support the in vitro assays investigating the potential binding mechanisms of compounds. Molecular docking helps in the feld of in silico drug design by identifying the molecules that can bind to a protein's active site. It illustrates how a promising therapeutic candidate (ligand) interacts with the target receptor's binding site and inhibits the target receptor's biological and catalytic activities [31]. Likewise, before chemical synthesis and biological testing, the prediction of biological activity for substances (PASS) can be used to estimate the biological activity profles of compounds [32]. In this study, we focused on some medicinal plants, which contain important bioactive compounds that could inhibit the catalytic activity of an enzyme aldose reductase (AR) thereby preventing diabetic complications through in vitro and in silico studies.

Determination of the Total Phenolic Content and Total
Flavonoid Content. Te naturally occurring phenolic and favonoid components have an antioxidant ability, which in turn prevents the chain reaction of reactive oxygen species. Reactive oxygen species have the potential to induce oxidative stress [40]. Te total phenolic content of the extracts was determined by using the Folin−Ciocalteu colorimetric method [41,42]. Initially, 0.5 mL of 95% ethanol extract was mixed with 5 mL of 10% Folin−Ciocalteau reagent, and 4 mL of 1M Na 2 CO 3 solution was added. Te mixture was subsequently left to stand for 15 minutes at room temperature. Te absorbance of the reaction mixture was measured at 765 nm by using a spectrophotometer.
Similarly, the total favonoid content was determined according to the colorimetric method [43,44]. For this, 0.5 mL of the extract (5 mg/mL) was mixed with 1.5 mL of 95% ethanol and 0.1 mL of aluminum trichloride (AlCl 3 , 10%). Subsequently, 0.1 mL of 1 M potassium acetate and 2.8 mL of distilled water were added to each bottle, and the reaction mixture was allowed to stand for 30 minutes. Te UV-visible spectrophotometer was used to measure the absorbance at 415 nm. Te standard curve for quercetin (10-50 µg/mL) was utilized for TFC and standard gallic acid (10-80 µg/mL) was used for TPC. Te amount of all polyphenolic and favonoid components in the extracts was represented as milligrams of gallic acid and quercetin equivalent per gram of dry weight, respectively.

Determination of Antioxidant Activity.
Antioxidants interact with free radicals and prevent the oxidative stress induced by excess free radicals. One of the stable free radicals is 2,2-diphenyl-1-picrylhydrazyl (DPPH), which shows a strong absorption band at 517 nm, and the absorption decreases when reduced by an oxidant [45]. Based on the radical scavenging properties of DPPH, the antioxidant activity of the extracts and the standard (ascorbic acid) were evaluated following the protocol of Alabri et al. [46]. Various concentrations of the plant extract (sample solutions) and ascorbic acid (reference samples) were prepared in ethanol (1000, 500, 100, and 50 µg/mL). To the various concentrations of the sample plant extracts and ascorbic acid solutions, 4 mL of a 0.1 mM DPPH solution in ethanol was mixed. Te mixture was left in the dark for 30 minutes. Similarly, 1 mL of ethanol (solvent) was added to 4 mL of 0.1 mM DPPH as a control, and the mixture was left in the dark for 30 minutes. Te absorbance was measured at 517 nm after 30 minutes. Te following equation is used to determine the ability to scavenge the DPPH radical [47]: where A o � absorbance of DPPH solution (control, without samples) and A t � absorbance of solution mixture of the test or reference sample and DPPH. Te percentage scavenging was then plotted against concentration and a regression equation was obtained from which IC 50 values were calculated for each plant extract by the formula given in Microsoft Excel 2007 software.

Inhibition of Aldose
Reductase. Te aldose reductase inhibition activity of the selected plant extract was accessed spectrophotometrically, by using glyceraldehyde as a substrate and NADPH as a cofactor following the protocol of Nakano and Petrash [48]. Initially, 100 μL of potassium phosphate bufer (pH 7.0), 755 μL of double distilled water, 20 μL of 1 mM DL-glyceraldehyde, 100 μL of 1.5 mM NADPH, and 20 μL of plant extract were added to the cuvette. Similarly, the blank solution contains 20 μL of double distilled water instead of plant extract. After that, 5 μL of RHAR was added and sorbinil was used as a positive Journal of Chemistry control for the inhibitor study. Te optical density (OD) of the reaction was monitored at 340 nm for 5 minutes at the interval of 30 seconds and the percentage inhibition was calculated. A decrease of OD/min represents the inhibition activity.

Percentage inhibition �
OD of blank solution -OD of plant extract OD of blank solution × 100 %.

Pharmacokinetic and ADMET Profle.
To reduce side efects, ADMET and drug-likeness of potential hit compounds are essential for the pharmaceutical industry [49]. Te pharmacokinetic parameters were predicted using the web-based program SwissADME (https://www.swissadme. ch) [50]. Te rule of fve, commonly known as Lipinski's rule, is used to determine drug-likeness [51] . Another web server, ProTox-II, was used for the toxicity analysis [52].

Estimation of Biological Activity.
Te selected phytochemical's biological activity was predicted by using the main Way2Drug server. It predicts a wide range of biological activity based on the structure of molecules instantaneously. Te activity is determined using the variables Pa (probable activity) and Pi (probable inactivity). For a specifc pharmacological activity, only substances with a Pa greater than Pi were examined [32]. As a result, it is possible to forecast whether ligands are potentially active or inactive based on their Pa and Pi values. Te detailed experimental procedure of this research work is shown in Figure 3.

Phytochemical Analysis.
Te extract prepared from dissolving 60 g of dry weight in ethanol showed variation in the percentage yield from 18% to 42%. Out of the four extracts, R. arboreum showed the highest percentage yield with 42% followed by A. pyrethrum (37%), B. ciliata (30%), and S. chirayita (18%) as shown in Table 2. Similarly, the phytochemical analysis of plant crude extracts displayed the presence of favonoids, glycosides, tannins, and terpenoids in all plants. However, proteins, amino acids, and saponins

DPPH Free Radical Scavenging
Activity. Free radical scavenging was performed to evaluate the antioxidant activity of plant extracts, and the result was expressed as IC 50 (half inhibitory concentration). Stable 1, 1-diphenyl-2-picrylhydrazyl (DPPH) was used to measure the free radical scavenging activity. Te results were compared with the standard ascorbic acid with an IC 50 value of 0.015 mg/mL. Te ethanolic extract of B. ciliata and R. arboreum showed signifcant antioxidant activity with an IC 50 value of 0.05 mg/ mL, while A. pyrethrum and S. chirayita extract exhibited a poor antioxidant activity with IC 50 of 0.28 mg/mL and 0.54 mg/ml, respectively.

Correlation.
Te Shapiro−Wilk test has indicated that all data are normally distributed (having a p value greater than 0.05), as shown in Table 3.
Since data are normally distributed, the Karl−Pearson correlation coefcient was evaluated (Table 4).
Te Karl−Pearson coefcient demonstrates a strong positive correlation between TPC and IC 50 and a strong negative correlation between TPC and RHAR. However, TPC and TFC are weakly correlated and RHAR shows a strong negative correlation with all three components, i.e., TPC, TFC, and IC 50. Figure 4, and it shows that only two components have an eigenvalue greater or equal to 1. So, only principal components 1 and 2 were used for further analysis as they account for the majority of variance in data. Tables S3 and S4 show that all the variables except RHAR are negatively correlated to PC1. In addition, only TPC is negatively correlated to PC2. Table 5 shows the ligand's binding afnities (docking scores) and H-bonding catalytic residues for human aldose reductase (PDB ID: 4JIR). Te results of the docking analysis suggested that morin (−9.2 kcal/mol), ursolic acid (−8.9 kcal/mol), and kaempferol (−8.2 kcal/mol) were the potential candidates that could inhibit the target protein. Te active sites of the receptor are shown in bold in Table 5. Morin was observed to form fve conventional hydrogen bonds at LYS-21, TYR-48, GLN-183, SER-214, and CYS-298, whereas two pi-pi stacked bonds formed at TRP-20 and TYR-289. Similarly, the control drug epalrestat was stabilized by one conventional hydrogen bond at TRP-111. Along with H-bonding, additional interactions between the epalrestat and morin with the receptor included pi-pi, alkyl, and pi-alkyl, as shown in Figure 5. Similarly, the 2D and 3D interactions of other compounds are shown in Figure S1.

Pharmacokinetic and ADMET Properties.
Te pharmacokinetics and drug-like characteristics of the chosen compounds are shown in Table S5. Ursolic acid breaks one out of the fve rules, but all other compounds were found     with no violations, indicating greater drug-like characteristics. Te standard compound, epalrestat, also showed zero violation of the Lipinski's rule but fell under toxicity class two. Table S6 shows the results of ADMET and toxicity analysis.

PASS Activity Analysis.
Te prediction of activity spectra for substances (PASS) was analyzed with selected phytochemicals as shown in Table S7. Pa and Pi are the two parameters that played a role in PASS prediction, and their values ranged from 0 to 1. Table S3 shows that all our phytocompounds except ursolic acid showed activity of aldose reductase inhibitor. Among the phytocompounds, morin had the highest value of Pa 0.456.

Discussion
Many people worldwide are attempting to fnd efective drugs to treat diabetes and related illnesses as a result of the growing attention given to the global challenges of diabetes and related disorders over the past few decades. Te development of new drugs is greatly aided by medicinal plants.
Te use of medicinal plants in the treatment of diabetes and their efectiveness in reducing its secondary consequences are topics that have attracted a lot of research in the recent  [53]. For more than 40 years, extensive research has been conducted on the signifcant enzyme known as AR. It has repeatedly been linked to the etiology of diabetic problems and acts as a rate-limiting enzyme in the polyol pathway. Te reduction of glucose to sorbitol is catalyzed by this enzyme [54].   [55]. Tis investigation reported reliable amounts of phenols, favonoids, tannins, and alkaloids detected from the ethanolic extracts of plants. Te presence of secondary metabolites is usually associated with potential biological efects [56]. Te vast variety of physiologically active substances make up the main secondary metabolites known as phenolic compounds which function as reducing and antioxidant agents because of their redox characteristics [57]. In addition, favonoids are also recognized as signifcant biological compounds with a variety of biological efects, including antioxidant, anticancer, anti-infammatory, antiangiogenic, and antiallergic properties [58]. Moreover, in the plant system, antioxidants are responsible for detoxifying reactive oxygen intermediates. Commonly, DPPH is used to screen antioxidants. Te extract's ability to scavenge free radicals is evidenced by the DPPH solution's discoloration [59]. Table 6 compares the antioxidant, TPC, and TFC activity of the selected plants with data from earlier studies and current fndings.
On the other hand, docking analysis showed that out of seven selected phytocompounds, morin and ursolic acid showed signifcant interaction (ΔG < −8.0 kcal/mol) with the receptor. Morin had the best docking score (ΔG � −9.2 kcal/ mol), forming fve H-bonds, two with active residues (TYR-48 and CYS-298) and three with nonactive residues (LYS-21, GLN-183, and SER-214). Arbutin showed a good afnity (ΔG � −6.6 kcal/mol) for forming H-bonds with TRP-111, ALA-299, and LEU-300. Te coumaric acid and ursolic acid showed only one H-bond (HIS-110) with a receptor. In our study among the docked compounds, morin had the highest binding afnity (−9.2 kcal/mol), outperforming the clinically tested ARIs lidorestat (−8.5 kcal/ mol), sorbinil (−8.1 kcal/mol), and RG7774 (−7.6 kcal/mol), as well as the commercial drug epalrestat (−8.1 kcal/mol). Te catalytic residue contains H-bonding with TRP-20, TYR-48, HIS-110, TRP-111, CYS-298, ALA-299, LEU-300, and SER-302, which resembles the previously described active residues [63]. TYR-48 and HIS-110 are positioned adjacent to the C4 of the nicotinamide ring in structural models of human aldose reductase complexed with NADPH, suggesting that one of these residues may serve as the proton donor in the reaction process. A hydrogen-bonding network that comprises LYS-77 and ASP-43 also includes TYR-48. So, ASP-43, LYS-77, and HIS-110 are important active sites for aldose reductase inhibition [64]. While ADMET analysis showed that all compounds get absorbed readily in the intestine, Cytochrome CYP450 (1A2, 2C9, 2C19, 2D6, and 3A4), which is primarily in-charge of the biotransformation of more than 90% of the drugs in phase-1 metabolism, has a considerable impact on drug metabolism [65]. Morin showed CYP1A2, CYP2D6, and CYP3A4 inhibitions, and kaempferol showed CYP2D6 and CYP3A4 inhibitions. Coumaric acid readily crossed the BBB while other compounds were not found to cross the BBB. Morin and ursolic acid showed immune toxicity, while only ursolic acid showed hepatotoxicity. None of the compounds showed cytotoxicity. From the toxicity class and LD 50 analysis, it can be concluded that catechin (having LD 50 10,000) was safer to use. Besides this, from PASS analysis, all phytochemicals showed antidiabetic activity.
Furthermore, the correlation among TPC, TFC, antioxidant activity, and RHAR inhibition property was determined by principal component analysis (PCA). PCA was carried out to break the dataset for the reduction of dimensionality. Tough principal component analysis was carried out for four components, only two components showed an eigenvalue greater than 1. So, only two components were chosen for further analysis. Tese two components were responsible for 99.8% of the variance of data. From, the factor loading score, it can be concluded that PC1 is primarily a measure of RHAR, and PC2 is primarily a measure of TFC, as only these variables have a positive factor loading greater than 0.5.

Conclusion
Since antiquity, several diseases have been treated with traditional medicinal plants. About 90% of the Nepalese population residing in rural areas depends on traditional medicine as they lack governmental healthcare facilities. Tis study focused on the evaluation of phytochemicals and biological activity of four diferent plants collected from diferent places in Nepal. From in vitro analysis, we found that the ethanolic extract of the selected plant has signifcant aldose reductase inhibition. Tis was confrmed by a molecular docking analysis of the human aldose reductase protein. Te ethanolic extract of B. ciliata, A. pyrethrum, and R. arboreum exhibits a signifcant antioxidant activity with a high TPC and exhibited greater inhibition of RHAR. Te previously isolated morin from B. ciliata showed a high binding afnity with a greater number of H-bonding towards catalytic residues by docking analysis. Although in vitro results of this study may have limited implications, these fndings provide the direction for exploring medicinal plants taken under study to avert or delay the onset of diabetic complications. Further studies on animal models and the isolation of pure compounds are required to support these fndings [73].

Data Availability
Te data used to support the fndings of this study are available from the corresponding author upon request.

Conflicts of Interest
Te authors declare that they have no conficts of interest.

Authors' Contributions
N.P. designed the research; L.B.M. and K.K. performed the research; S.R.U. and R.T. wrote the manuscript and generated molecular docking and did statistical analysis; and S.J., B.P.M., and N.P. edited the manuscript. Siddha Raj Upadhyaya and Lila Bahadur Magar contributed equally to this work.

Supplementary Materials
Table S1: reported IC 50 or percentage inhibition and enzyme type of selected natural products; Table S2: phytochemical screening of crude extract of plants; Table S3: principal component (standard deviation, proportion of variance, and cumulative proportion) analysis; Table S4: factor loading data of variables with PC1 and PC2; Table S5: pharmacokinetic and drug-likeliness properties of selected compounds; Table S6: ADMET profles of selected compounds; Table S7: prediction of activity spectra for substances for selected compounds by the main Way2Drug server; Figure  S1: