Toxicity predictions were performed using Derek Nexus (Knowledge Base: v6.4.1; Nexus: v2.7.2; Lhasa Limited, Leeds, UK), a knowledge-based expert system for mechanistic toxicity assessment. The software’s rule-based framework incorporates structural alerts derived from empirical toxicological data, peer-reviewed literature, and regulatory guidelines. Metabolic predictions were performed using Meteor Nexus (Lhasa Limited, Leeds, UK), a knowledge-based expert system for Phase I/II biotransformation analysis. The software integrates empirical data, peer-reviewed literature, and proprietary metabolic rules to predict potential metabolites and their accumulation pathways.

The pyrrole-based hit compound (structure detailed in Table 1) was input into Derek Nexus using its canonical SMILES string, generated and validated via ChemAxon’s MarvinSketch (v22.15). To ensure reproducibility, the compound’s stereochemistry and tautomeric forms were explicitly defined prior to analysis.

All toxicity endpoints available in the Derek Prediction Set Up menu were evaluated, including carcinogenicity (e.g., IARC Group classifications); genotoxicity (mutagenicity, clastogenicity); organ toxicity (hepatotoxicity, nephrotoxicity, neurotoxicity); reproductive/developmental toxicity; sensitization (skin, respiratory); and miscellaneous endpoints – phototoxicity and phospholipidosis. Predictions were generated for bacteria (primarily for Ames test relevance) and mammals (rat, human) to assess species-specific toxicological profiles. The prediction confidence levels were categorized as probable: strong structural or mechanistic evidence (≥70% likelihood); plausible: moderate evidence (30–69% likelihood); Equivocal: Insufficient or conflicting evidence (<30%); negative predictions (no alerts) were assigned when the compound lacked structural motifs associated with toxicity in Derek’s knowledge base (e.g., no phototoxic arylpropanone or polycyclic aromatic fragments).

Metabolic pathways were visualized hierarchically as metabolic trees, annotating parent compounds and their derivatives with reaction types (e.g., hydroxylation, ester hydrolysis). Mechanistic rationales, such as CYP3A4 affinity for aromatic oxidation, and literature citations supporting predicted biotransformations were curated. Competing reactions, such as glucuronidation versus sulfation, were flagged for further analysis.

The initial calculations were aimed at determining the chemical toxicity after in-depth evaluations with the Derek Nexus module of Lhasa. The module provides data about the potential safety concerns of the tested structures. These assessments are vital in pharmaceutical development and regulatory submissions, as they identify chemical features – known as toxicophores – that are linked to harmful biological effects (Anwar et al. 2021). Several studies have demonstrated a strong correlation between experimental results and predictions made by Derek Nexus, supporting the reliability and established status of the module (Judson et al. 2013). Derek Nexus utilizes data from both the Ames bacterial reverse mutation assay and transgenic rodent mutation models to inform its toxicity predictions.

Derek Nexus assesses chemical compounds across more than 50 toxicological endpoints, providing a likelihood rating for each prediction that ranges from high certainty to more tentative assessments. Negative results are classified from “inactive” to “improbable.” For example, if a compound shares structural features with known mutagens, it may be designated as a “probable mutagen” and assigned a specific confidence level. The platform also offers mechanistic explanations for its predictions, such as detailing cytochrome P450-mediated metabolic pathways in cases of hepatotoxicity or referencing established assays for endpoints like skin sensitization. Notably, the analysis indicated that compound 7b did not show toxicity across any of the 57 evaluated endpoints, with the full list of endpoints summarized in Table 2.

Moreover, compound 7b was evaluated using the Carcinogenic Potency Categorization Approach (CPCA) for N-nitrosamines. This approach is a structure-activity relationship (SAR)-based framework that categorizes N-nitrosamine impurities into potency categories, each associated with an acceptable intake limit. The categorization is based on the presence of specific structural features within the molecule that are known to influence carcinogenic potential. The CPCA is particularly relevant for N-nitrosamines containing a carbon atom on both sides of the N-nitroso group and is widely used to predict carcinogenic risk when experimental data are unavailable or limited.

During the assessment of compound 7b, special attention was given to the identification of dialkyl N-nitrosamine substructures, as these are recognized as highly potent carcinogens due to their metabolic activation pathways. Structural analysis of 7b demonstrated that it does not contain any dialkyl N-nitrosamine motifs, which are typically associated with a higher risk of carcinogenicity. The absence of these hazardous structural features in compound 7b suggests a lower predicted risk for nitrosamine-related carcinogenicity.

Several structural alerts were identified during the in silico toxicity assessment of compound 7b using the Derek Nexus platform, with most of these alerts related to the skin sensitization endpoint. Derek Nexus is a knowledge-based expert system that predicts toxicity by identifying chemical substructures, known as toxicophores, which are associated with specific toxicological outcomes. Fig. 1 illustrates the specific alert generated for the substituted phenol ring in compound 7b, highlighting the correspondence between this structural feature and the system’s prediction.

Figure 1. 

Description of the alert image and its occurrence in 7b (highlighted in grey).

The primary alert triggered for compound 7b is associated with the presence of a substituted phenol ring. Phenolic structures are recognized as potential skin sensitizers, and their presence in a molecule often prompts further evaluation. Derek Nexus incorporates this knowledge into its alert system, flagging compounds that contain such substructures due to their potential to induce skin sensitization reactions. This is based on evidence that phenol derivatives can undergo metabolic activation or direct interaction with skin proteins, ultimately leading to immune responses characteristic of sensitization.

While phenol itself has generally not shown skin sensitization in animal studies, several of its substituted derivatives have demonstrated sensitizing potential. For instance, compounds such as 2,5-dimethylphenol, 3,4-dimethylphenol, pentachlorophenol, and 2,4,5-trichlorophenol have been identified as skin sensitizers in the local lymph node assay (LLNA) (Ashby et al. 1995). Similarly, other substituted phenols, including 4-styrylphenol (also known as 4-hydroxystilbene) and chloroxylenol (4-chloro-2,3-dimethylphenol), have produced positive results in the guinea pig maximization test (GPMT) (Cronin and Basketter 1994). However, it is important to recognize that not all substituted phenols act as sensitizers; some, such as 4-tert-butylphenol, 2-methylphenol, and 3,4,5-trimethylphenol, have yielded negative results in both the GPMT and LLNA, indicating variability in sensitization potential among phenol derivatives, as discussed by Yamano et al. (2007).

Hydrazine and phenylhydrazine are classified as significant contact allergens (category A) based on an analysis of the available human and animal data, and positive human patch test results were reported for hydrazine and selected derivatives (Foulds and Koh 1991). Conversely, some human patch test studies have reported negative results (Rothe 1988). Hydrazides such as carbonohydrazide showed positive activity in the GPMT, although the structural analogue adipohydrazide was negative in the LLNA. The alert description image and its match with the evaluated structure, 7b, are given in Fig. 2.

Figure 2. 

Description of the alert image and its occurrence in 7b (highlighted in grey).

In the context of skin sensitization, research by Chilton et al. (2018) has validated the use of confidence metrics for non-sensitizer predictions, drawing on data from established assays such as the Local Lymph Node Assay (LLNA) and the Human Repeat Insult Patch Test (HRIPT). These studies confirm the reliability of Derek Nexus in predicting sensitization potential based on a chemical structure.

The structural configuration of the query molecule does not correspond to any known structural alerts or reference compounds linked to bacterial in vitro mutagenicity in the Derek knowledge base. Additionally, the compound lacks any unclassified or ambiguous features, further supporting the computational prediction that it will be inactive in the bacterial reverse mutation (Ames) assay.

Overall, analysis using Derek indicated that the title compound 7b is predicted to have plausible potential for skin irritation or corrosion in mammalian models. However, the compound was predicted to be inactive with respect to in vitro mutagenicity in bacterial assays and showed no activity in 58 other evaluated toxicity endpoints. Therefore, it could be used for future hit-to-lead and lead optimization processes.

Meteor Nexus is a commercially available metabolite prediction software that is part of a larger suite of tools developed by Lhasa Ltd. Meteor generates metabolites using a knowledge-based expert system, which identifies structural patterns within the parent compound and correlates these patterns to encoded biotransformations. The Meteor software employs a knowledge-based approach, using curated scientific data and reasoning algorithms to predict possible metabolic transformations (Phase I and II) for a given chemical structure. This approach is distinct from data-driven or purely statistical models, as it relies on encoded expert knowledge and mechanistic rules to generate its predictions (Boyce et al. 2022). Moreover, recent validation studies published in journals such as Xenobiotica and the Journal of Medicinal Chemistry support the integration of Derek Nexus with metabolic prediction platforms like Meteor Nexus. These studies highlight the advantages of combining mechanistic toxicity predictions with metabolic fate analysis, demonstrating that such an approach can improve early risk assessment in drug discovery by providing a more comprehensive evaluation of both parent compounds and their metabolites. The results provided by Meteor are generalized in Fig. 3.

Figure 3. 

Predicted biotransformation pathway of 7b.

The most probable metabolic pathway for compound 7b is hydrolysis of the acyclic carboxylic ester located at the 4^th position of the pyrrole core motif (Fig. 4). Ester hydrolysis is a common metabolic reaction, typically catalyzed by carboxylesterase enzymes in various tissues, especially in the liver. This process involves the cleavage of the ester bond, resulting in the formation of a carboxylic acid and an alcohol as the primary metabolites. These hydrolysis products are generally more polar than the parent ester, which increases their hydrophilicity and promotes their excretion via the kidneys (Laizure et al. 2013).

Figure 4. 

Hydrolysis of the ester group in 7b.

For compound 7b, hydrolysis at the 4^th position would introduce a polar carboxylic group, likely increasing water solubility and reducing membrane permeability. This transformation may impact the pharmacokinetic profile of 7b by decreasing its bioavailability while facilitating its elimination. Thus, the presence of the acyclic carboxylic ester at this position represents a key metabolic liability that should be taken into account during further development and optimization of the compound. Table 3 presents the chemical structures identified as the closest metabolic analogues to compound 7b. The reference compounds were selected based on their structural similarity to 7b, particularly with respect to core motifs and key functional groups relevant to metabolic transformation. The comparative analysis of these analogues enables a more informed prediction of the metabolic pathways that the pyrrole-based compound 7b may undergo.

Table 3.

Download as

CSV

XLSX

Nearest metabolic analogues of compound 7b undergoing hydrolysis of the acyclic carboxylic ester.

	19%	Positive	Ma et al. 2005
	17%	Positive	Roberts et al. 2012
	16%	Positive	Roberts et al. 2012

Furthermore, the predictions generated by the Meteor software indicated a high likelihood for the initial hydrolysis of the acyclic carboxylic ester moiety in compound 7b, followed by subsequent hydroxylation of the indole ring (Fig. 5). The reliability of these predicted metabolic transformations is supported by the knowledge-based algorithms employed by Meteor, which utilize curated mechanistic rules and expert-derived data to forecast plausible biotransformation pathways.

Figure 5. 

Hydrolysis of the ester group followed by hydroxylation of the tryptophan ring in 7b.

The initial hydrolytic cleavage of the ester group is a well-recognized Phase I metabolic reaction, typically mediated by carboxylesterase enzymes, resulting in the formation of a more polar carboxylic acid derivative. Following this transformation, the indole ring is predicted to undergo hydroxylation, a common oxidative process catalyzed by cytochrome P450 enzymes. Such hydroxylation reactions generally increase the hydrophilicity of the molecule, further facilitating its excretion.

The sequential nature of these metabolic events – beginning with ester hydrolysis and proceeding to aromatic hydroxylation – reflects typical metabolic routes for compounds containing both ester and indole functionalities. This metabolic profile provides valuable insight into the likely biotransformation fate of compound 7b and highlights potential sites for structural modification to improve metabolic stability or modulate pharmacokinetic properties. Table 4 presents the chemical structures identified as the closest metabolic analogues to compound 7b.

Table 4.

Download as

CSV

XLSX

Representative compounds undergoing ester hydrolysis and subsequent tryptophan ring hydroxylation analogous to 7b.

	66%	Positive	Barbuch et al. 2006
	65%	Positive	Barbuch et al. 2006
	62%	Negative	Sun and Yost 2008

By examining the documented metabolic profiles of these structurally related compounds, it is possible to anticipate potential metabolites, assess metabolic liabilities, and guide the rational design of derivatives with enhanced metabolic stability. This approach provides a valuable framework for understanding the likely biotransformation routes of 7b and supports the overall strategy for optimizing its pharmacokinetic properties.