For evaluation of the drug-likeness, bioactivity score, pharmacokinetic, and a brief toxicological profile in silico, the programs Molinspiration Cheminformatics (https://www.molinspiration.com/), SwissADME (http://www.swissadme.ch/index.php), and PreADME/Tox (https://preadmet.qsarhub.com/), were used. The structure of Hyp was analyzed in the modules for the prediction of molecular properties (Lipinski’s rule of 5), bioactivity and absorption, distribution, metabolism, excretion (ADME), and toxicity. For the evaluation of each parameter, the following descriptors were chosen:

• properties–octanol-water partition coefficient (LogP; given as miLogP), total polar surface area (TPSA), molecular weight, number of H- acceptors and donors, number of violations, number of rotatable bonds, and molecular volume;
• bioactivity–receptor and enzyme target prediction;
• absorption–absorption in Caco-2 cells and human intestinal absorption (HIA);
• distribution–plasma protein binding (PPB) and permeability in the blood-brain barrier (BBB);
• metabolism–interaction with enzymes from the Cytochrome P450 (CYP) complex;
• excretion–renal excretion through permeability in Madin-Darby canine kidney (MDCK) cells;
• toxicity–mutagenicity, carcinogenicity in rats and mice, and inhibition of the hERG gene.

For a more detailed evaluation of Hyp’s toxicity, an OECD QSAR Toolbox software was used. The profiler ‘In vivo rat metabolism simulator’ consists of 30–40 abiotic (non-enzymatic) and 630–640 enzyme-controlled reactions. The simulator also contains 520–530 enzymatic Phase I and 100–110 Phase II transformation reactions, but only metabolites (products) resulting from Phase I reactions are visualized. The ‘Rat liver S9 metabolism’ molecular transformations set comprise about 40–50 abiotic and a few enzyme-controlled reactions believed to occur at a very high rate. The simulator contains 450–460 enzymatic Phase I and 40–50 Phase II transformations. Only the generated in vitro metabolites (products) resulting from Phase I reactions are visualized. The principal applicability of these simulators is associated with the reproduction and prediction of the metabolic pathways of xenobiotics, which may elicit in vivo genotoxic effects (e.g., bacterial mutagenicity and chromosomal aberrations) (https://qsartoolbox.org/).

The profilers ‘DNA binding by OASIS,’ ‘Protein binding by OASIS,’ ‘Toxic hazard classification by Cramer,’ ‘Carcinogenicity,’ ‘In vitro mutagenicity,’ and ‘In vivo mutagenicity’ were used to elicit the toxicological profile of Hyp and its metabolites. The scope of ‘DNA and Protein binding by OASIS’ is to investigate the presence of alerts within target molecules that may interact with DNA and/or proteins (Mekenyan et al. 2004; Serafimova et al. 2007; https://qsartoolbox.org/). Categorization rules of chemicals into different levels (Class I (Low), II (Intermediate) and III (High)) of toxicological concern (when administered orally) are organized in a tree-like scheme in the ‘Toxic hazard classification by Cramer’ (https://qsartoolbox.org/). The ‘Carcinogenicity’ and ‘In vitro and in vivo mutagenicity’ work as a decision tree for estimating carcinogenicity and in vitro (Ames test) and in vivo mutagenicity, based on a list of 55, 30, and 35 structural alerts, respectively (https://qsartoolbox.org/).

Determination of the molecular properties of Hyp was done using Molinspiration software. The results of the analysis are presented in Table 1.

Lipinski‘s rule of 5 states, ‘The ideal drug molecule must have certain physicochemical properties.’ and predicts if a biologically active molecule can be taken orally. According to this rule, a compound must have a molecular mass of less than 500 Da; LogP of less than 5; less than five hydrogen bond donors and ten hydrogen bond acceptors; polar surface area of 140 Å; and less than ten rotatable bonds (Lipinski 2004).

If a molecule does not infract more than one rule, it should have good pharmacokinetic properties and bioavailability in the organism. According to Lipinski’s rule of 5, Hyp gives two violations, which make it less likely to be orally active. Nevertheless, one of the parameters, the molecular weight, is 7.2% higher than the limit–in this magnitude, it could not drastically affect the pharmacokinetics and bioavailability. In their scientific work, Biber et al. (1998) described the ranges of Hyp plasma levels after oral administration of St. John’s wort alcoholic extracts. They have clarified that after oral administration, therapeutic doses can be achieved.

A more detailed pharmacokinetic analysis was made with PreADME/Tox software giving us information about the absorption, distribution, metabolism and excretion, and toxicity. Results are shown in Table 2.

Two main predictive models were analyzed to assess the absorption profile of Hyp: permeability in colon adenocarcinoma (Caco-2)-derived cells and the rate of human intestinal absorption (HIA). The obtained data showed excellent intestinal absorption (96.700160%) and medium cellular permeability after oral administration of Hyp.

To predict the distribution profile of Hyp, its ability to bind to plasma proteins and pass through the BBB was evaluated. The phloroglucinol derivative was found to have a high affinity for plasma proteins (100%) and high central nervous system (CNS) absorption (9.52384).

The metabolism of Hyp was assessed by its ability to inhibit four isoenzymes of the CYP450 complex—a family of liver enzymes responsible for the metabolism of endogenous substances and xenobiotics. The analyzed compound was found to be a CYP3А4 substrate and inhibitor (Obach 2000; Lee et al. 2006) and a CYP2C9 inhibitor. Compounds that are inhibitors of relevant isoenzymes should not be administered with other drugs or xenobiotics substrates of cytochrome P450 isoenzymes due to the risk of increased plasma concentrations or loss of effect.

To analyze and predict the excretion profile of Hyp, the MDCK cell permeability model was evaluated. The phytochemical was found to have medium permeability to MDCK cells in this in-silico assay, suggesting that it would have a longer renal excretion time.

When assessing toxicity, the mutagenicity and carcinogenicity should also be analyzed. The model used for predicting the mutagenicity, the Ames test, confirmed that Hyp has a non-mutagenic effect. The in silico carcinogenicity prediction in rats and mice showed that the analyzed compound is a carcinogen. These results partially overlap with those obtained by the QSAR Toolbox software.

Another parameter in evaluating new drug compounds is cardiotoxicity. Inhibition of the hER gene leads to impaired expression of potassium channels and the subsequent occurrence of heart problems. In some cases, a fatal outcome is possible. Hyp was observed to have a medium risk in inhibiting the hER gene and, therefore, is not potentially cardiotoxic according to in silico tests.

In summary, Hyp has a good oral ADME profile according to the used PreADME/Tox software.

For evaluation of the drug-likeness, Molinspiration Cheminformatics and SwissADME software were used. Drug-likeness can be defined as a complex balance of different molecular properties and structural features that determine whether a given molecule is similar to a known drug. The variety of possible drug targets (each requiring a different combination of matching molecular features) is so vast that it is nearly impossible to test all in vitro or in vivo. This necessitates using in silico methods to predict the molecule’s bioactivity. Results for bioactivity score from Molinspiration Cheminformatics software are presented in Table 3.

As values less than 0 mean the molecule has no affinity for the corresponding target, it can be assumed that Hyp has an average affinity for nuclear receptors and a low affinity for enzymes.

The bioactivity of the acylphloroglucinol was determined in more detail with SwissADME software. Results for bioactivity score (Target Prediction) from SwissADME software are presented in Table 4.

Data showed that Hyp has activity towards various receptors and enzymes, explaining the high biological activity and the possibility of applying Hyp for multiple indications. The activity towards the nuclear Pregnane X receptor and the enzyme Prostaglandin E synthase was observed to be most significant, confirming the results obtained from Molinspiration software.

For evaluation of the metabolic activation and further toxicological analysis, the OECD QSAR Toolbox software was used. In silico study showed that Hyp cannot bind to DNA or proteins. However, according to the Toxic hazard classification by Cramer, Hyp is positioned in Class III. Due to alpha, beta-unsaturated carbonyl structural alert, it has carcinogenic and mutagenic (in vitro and in vivo) effects.

Applying a mathematical metabolism prediction model allowed us to identify and determine the metabolic activation of Hyp’s structure. The resulting metabolites were also analyzed for their ability to bind to DNA and proteins, Toxic hazard classification by Cramer, carcinogenicity, and in vitro and in vivo mutagenicity.

As a result of the mathematical prediction performed using the in vivo rat metabolism simulator for Hyp, 103 metabolites were obtained, presented in Appendix 1: Table A1. In the scientific work of Hokkanen et al. (2011), it was found that Hyp gives 57 metabolites formed by monohydroxylation, hydroxylation, dehydrogenation, or a combination of these reactions.

As a result of the mathematical prediction performed using the Rat liver S9 metabolism simulator of Hyp, a total of 16 metabolites were obtained. They are presented in Appendix 1: Table A2.