N-nitrosamines (N–NAs) are a group of organic compounds that are known to be potentially carcinogenic [1]. They are formed when amines react with nitrites in the presence of an acid [2,3]. N–NAs can be found in a variety of foods, including smoked meats [4], cured meats, fish, dairy products, and vegetables [5] EFSA Panel on Contaminants in the Food Chain [6]. The reaction typically occurs during food processing, especially in situations where nitrites are used as preservatives or flavour enhancers [7,8]. The concentration of N–NAs in these foods can differ depending on the type of food, the processing methods used, and the storage conditions [9]. N–NAs can be produced when certain conditions allow the combination of an amine with a nitrosating agent. However, there are alternative pathways for their formation, including oxidation and reduction processes involving hydrazine-type compounds and N-nitro derivatives [10].

Why is this so important from the point of view of regulatory toxicology? There is no doubt that consumers may be exposed to N–NAs through multiple routes, further exacerbating the regulatory challenges. In food products, the ingestion of N–NAs can occur through contaminated meat products, cured meats, and various processed foods containing nitrite-based preservatives [11,12]. For cosmetics, topical application can lead to dermal exposure, while aerosolised product inhalation can be an additional concern [13]. In pharmaceuticals, especially those with N–NAs impurities, oral ingestion is the primary exposure route [14,15]. These diverse exposure pathways require comprehensive risk assessments to accurately assess potential health hazards.

The health implications of these xenobiotics are a significant public concern due to their potentially carcinogenic nature [16]. Prolonged exposure to these compounds can lead to various health issues, including the development of cancer. In addition, N–NAs have been associated with liver damage and reproductive problems in certain studies [17]. The International Agency for Research on Cancer (IARC), an authoritative body for cancer research, has classified N–NAs as Group 1 carcinogens. This classification denotes that there is sufficient evidence to conclude that these compounds are capable of causing cancer in humans. The tumorigenic dose rate 50 (TD50) has been used as a Point of Departure (PoD) when the toxic substance was administered chronically for the standard lifetime of the species, but it is not recommended for low-dose extrapolation [18].

The importance of the N-NA issue has recently been noticed by the EFSA in a comprehensive assessment of N-NA in food EFSA Panel on Contaminants in the Food Chain [6]. This document takes into account various aspects of the toxicological evaluation of these xenobiotics. Of particular concern within this assessment were two problematic N–NAs: NMAMPA (N-2-methylpropyl-N-1-methylacetonylnitrosamin; CAS: 93755-83-0) and NMAMBA (N-3-Methylbutyl-N-1-methylacetonylnitrosamine; CAS: 71016-15-4) which are briefly characterised in Table 1 based on PubChem.

The problematic nature of these compounds is due to the lack of important toxicological data. For NMAMPA, no genotoxicity and metabolic data were available EFSA Panel on Contaminants in the Food Chain [6]. However, for NMAMBA, limited data indicated positive results in Salmonella Typhimurium (TA1535 and TA100 strains) only in the presence of metabolic activation by rat liver S9 EFSA Panel on Contaminants in the Food Chain [6]. Based on the presence of α-hydrogens suitable for hydroxylation, similar to other acyclic N–NAs [19], both NMAMPA and NMAMBA are expected to be carcinogenic (EFSA Panel on Contaminants in the Food Chain (CONTAM) et al., 2020).

In the course of the EFSA study, the Carcinogenic Potency Database (CPDB) within the OECD QSAR Toolbox (Organization for Economic Co-operation and Development- Quantitative structure–activity relationship) was consulted, revealing acyclic N–NAs analogues (n = 11) with tumorigenic dose rate 50 (TD50) values for the rat. While some of these compounds were not included in the present EFSA analysis, they served to expand the database. Notably, a correlation was observed between TD50 and LogKow (a hydrophobic parameter) with an R-square value of 0.53, supporting the dependency of TD50 on the hydrophobicity of N–NAs, consistent with previous research by Wishnok et al. [20]. Using this correlation, the EFSA tentatively estimated the TD50 values for NMAMPA and NMAMBA: 0.242 mg/kg bdwt/day and 0.34 mg/kg bdwt/day, respectively. It is important to note that these TD50 estimates are conservative, as the bulkiness and branching of the alkyl moieties can potentially decrease the carcinogenic risk associated with these N–NAs. It should be noted that the conducted estimations have been carried out in a rather enigmatic mannerthe published document includes only a screenshot and concise information about the performed predictions EFSA Panel on Contaminants in the Food Chain [6]. The confusion of designations (NMAMP is described twice) is also questionable. It is also surprising that only estimations were made using only one toxicology in silico software (QSAR Toolbox v 4.4). It is essential to note that this estimate should ideally be performed using various methods, as recommended by the European Medicines Agency (EMA) which requires the use of two complementary (Q)SAR methodologies in the assessment of pharmaceutical impurities [21]. It is important to realise that the prediction of carcinogenic potency is a challenging and demanding process. In this situation very important is the question – what exactly is TD50?

In the context of classical acute toxicity, the LD50 (median lethal dose) is a well-established parameter used to quantify the acute toxicity of a chemical substance [22]. To establish a roughly analogous measure for the tumorigenicity of a particular agent, the concept of tumorigenic dose rate 50 – TD50 was defined [23]. In the simplest terms, it can be stated that TD50 represents the daily dose rate in mg/kg bdwt/day required to induce tumours in 50 % of the test animals that would otherwise have remained tumour-free at a zero dose [24]. In cases where multiple positive experiments are conducted on a species, the reported TD50 value is determined as the harmonic mean [25], derived by considering the TD50 value of the most potent target site observed in each of the positive experiments. It may seem that the parameter known as TD50 is analogous to LD50 [23]; however, it is essential to recognise that these dose descriptors pertain to entirely different situations – non-threshold and threshold responses, respectively. Measures of carcinogenic potency, such as TD50, are instrumental in the determination of metrics such as the threshold of toxicological concern (TTC), acceptable intake (AI) and permitted daily exposure (PDE), which in turn impact on human exposure [26]. In general, TD50 is defined as the dose required to halve the probability that a subject will remain without tumours throughout the lifetime of exposure [25]. The calculation or estimation of TD50 values is a complex process, as evidenced by various publications, notably Cox (1972), Peto et al. [23], and Sawyer et al. [23,27,27]. The articles mentioned outline the methodology for determining TD50; however, its implementation requires the use of ‘lifetable’ data, where the incidence of tumours is meticulously tracked over time. Unfortunately, such comprehensive data are seldom accessible because of the constraints imposed by current experimental protocols. Therefore, the determination of TD50 is a difficult procedure, often requiring specialised software for the appropriate estimations. This approach considers the possibility of nonlinearities in the dose-response relationship, which can play a crucial role in estimating the dose at which 50 % of the test animals develop tumours [26]. Furthermore, the TD50 calculations can be tailored to specific categories of neoplastic lesions or encompass all types of tumours observed in the experimental study. This flexibility allows researchers to focus on particular tumour types or obtain an overall assessment of the tumorigenic potential of a substance. By incorporating advanced mathematical algorithms and considering various aspects of the dose-response relationship, the process of calculating the TD50 values becomes a sophisticated and data-driven approach. This comprehensive analysis provides valuable information on dose-dependent tumour development, allowing a more accurate assessment of the potential carcinogenicity of the agents tested. Hence, the significance of employing diverse in silico tools that rely on various methodologies to yield accurate predictions for this parameter cannot be overstated [28]. In modern toxicology, it is important to apply suitable in silico software, applied in a skilful manner, and meticulously documented to determine the TD50 values [29].

The aim of this research was to determine the carcinogenic potency and tumorigenic dose (TD50) for two problematic N–NAs: NMAMPA (CAS: 93755-83-0) and NMAMBA (CAS: 71016-15-4) using various available in silico tools: qualitative methods (ToxTree, ProTox II, and CarcinoPred-EL) and quantitative methods (QSAR Toolbox, LAZAR). The idea as a workflow of the conducted studies is presented schematically in Fig. 1. The inspiration for conducting this research comes from the under-use of the vast potential of known in silico toxicology methods in practical applications. While numerous theoretical studies have been conducted to introduce and validate these methodologies, there remains a lack of publications that demonstrate their practical implementation (especially including the prediction of TD50). Additionally, the motivation is driven by the insufficient documentation of the results of TD50 prediction studies by EFSA for the two problematic N–NAs mentioned (NMAMPA and NMAMBA) the TD50 prediction studies for these compounds lack comprehensive documentation and transparency (and even contain inaccuracies and discrepancies in relation to our results).