Economic growth and development come with increasing use of chemicals and subsequent exposure of humans both in professional and in consumer environments. Exposure to chemicals occurs through intended use of, for instance, agrochemicals, pharmaceuticals, or cosmetics. Exposure can also occur as environmental pollution due to spillage during production or use, or through waste at the end of use. New chemicals are designed every day and further developed for commercial distribution. Until now, regulation of chemicals for safe-guarding human and environmental health largely depended on toxicological studies in laboratory animals. Although still common practice, the disadvantages of testing in animals are widely acknowledged, as this comes with drawbacks relating to ethical (animal rights), economic (high cost), and scientific (human relevance) principles. For these reasons, the toxicological world has been working on a paradigm shift to move away from animal testing and to develop alternative testing methods. The initial approach to simply replace key target tissues in the animal by representative cell cultures appeared inadequate, and subsequent developments were to combine simple tests into batteries, to design more complex testing models, and to implement mechanistic considerations. Quantification of alternative testing was improved through dose–response modeling, translation through toxicokinetic modeling, and extrapolation through human relevance assessment. Also, exposure scenarios were considered to guide targeting of toxicological studies. All such developments were compiled into new concepts for chemical hazard and risk assessment, such as Toxicity Testing in the 21st Century and RISK21.(1,2) However, acceptance and actual implementation for regulatory purposes take a further step; it requires comprehensive testing of such concepts to see how new models and strategies work out in practice, to uncover associated uncertainties, and to define applicability and further refinements. A strategy to apply Next Generation Risk Assessment is depicted in Figure 1. It starts with problem formulation, in which initial considerations should be outlined, such as defining the population that is to be protected by the assessment, the associated probable exposure scenario, and the purpose of the assessment (prioritization for development, regulation). A first phase of assessment is aimed at predicting levels of external and internal exposure and of toxicological potential. In this phase, external exposure is modeled in a deterministic way, based on intended use of the chemical, for instance, by combining estimated concentrations in food with consumption patterns. Estimation of internal exposure (bioavailability) considers external exposure together with physicochemical properties of the compound, such as lipophilicity and molecular size, which affect its behavior in the body. Initial predictions of toxicological potential are based on in silico tools, including QSAR-directed read-across models, molecular docking, and Cramer classification, and on arrays of high-throughput in vitro assays, preferably interpreted using computational toxicity prediction models. Adverse outcome pathways may be helpful to structure and interpret the retrieved toxicological information in this and in the next phase. The initial assessment should lead to conservative estimation of risk, detect concerns and alerts, necessitating (or waiving, if absent or below a threshold of toxicological concern) more detailed analyses for confirmation and quantification in a subsequent phase. Figure 1. Flow of analysis in Next Generation Risk Assessment. The steps are explained in the text. QSAR, quantitative structure−activity relationship; MIE, molecular initiating event, applicable to adverse outcome pathway; TTC, threshold of toxicological concern; POD, point of departure. In the second phase, improved exposure assessment is achieved through more complex (probabilistic) models based on measured concentrations in vehicles of exposure, intake information, and measured concentrations in (human) biomaterials. Such refined exposure data can be used as input for toxicokinetic models to quantify internal exposure concentrations, notably at target tissues as indicated by first-phase toxicity models. First-phase toxicity alerts and predictions can be confirmed by more targeted, complex toxicological models, which should be more relevant to the situation in the human body, such as micro-organ cultures and nonlicensed whole organism cultures (vertebrate embryos, lower organisms). Eventually, and only as a last resort, traditional laboratory animals can be used for final confirmation. All such models should provide benchmark-based effective doses, which, again through toxicokinetic modeling, can be extrapolated to effective doses in humans. In a final assessment, the second-phase results are synthesized into a conclusive risk evaluation, specified to target tissues and to sensitive populations. Conducting case studies is an excellent way to test applicability of new strategies for chemical risk assessment. An example of such a case study along the flow outlined above is described in this issue.(3) The study, with a focus on the effect analysis of three triazoles, shows that an alternative approach for toxicity testing leads to similar conclusions as traditional in vivo testing, providing support for applying Next Generation Risk Assessment concepts for regulatory purposes. This article references 3 other publications.

中文翻译：

---

下一代化学品对人体健康的风险评估的实际应用

经济增长和发展伴随着化学药品使用的增加以及随后在专业环境和消费环境中人类的接触。化学品的暴露是通过例如农药，药品或化妆品的预期使用而发生的。由于生产或使用过程中的溢出或使用结束时的废物，也可能由于环境污染而导致接触。每天都会设计新的化学品，并进一步开发以用于商业分销。迄今为止，为保护人类和环境健康而对化学品的管理很大程度上取决于对实验动物的毒理学研究。尽管仍是普遍做法，但动物测试的弊端已广为人知，因为它伴随着道德（动物权利），经济（高成本），和科学（与人类相关）的原则。由于这些原因，毒理学界一直在努力进行范式转变，以摆脱动物测试并开发替代测试方法。用代表性的细胞培养物简单地替换动物中关键靶组织的最初方法似乎是不充分的，随后的发展是将简单的测试结合到电池中，设计出更复杂的测试模型，并实现了机械方面的考虑。通过剂量反应模型，通过毒物动力学模型进行翻译以及通过人类相关性评估进行推断，改进了替代测试的量化。此外，还考虑了暴露情况以指导毒理学研究的目标。所有这些进展都被纳入化学危害和风险评估的新概念中，例如21世纪的毒性测试和RISK21。（1,2）但是，出于监管目的的接受和实际实施又向前迈进了一步；它要求对此类概念进行全面测试，以了解新模型和策略在实践中如何工作，发现相关的不确定性以及定义适用性和进一步完善。图1中描述了应用下一代风险评估的策略。它从问题制定开始，其中应概述最初的考虑因素，例如定义要由评估保护的人群，相关的可能的暴露场景以及风险评估。评估的目的（优先发展，监管）。评估的第一阶段旨在预测外部和内部暴露水平以及毒理学潜力。在这个阶段 根据化学品的预期用途，例如通过将食物中的估计浓度与消费模式相结合，以确定性的方式对外部暴露进行建模。内部暴露（生物利用度）的估算将外部暴露与化合物的物理化学特性（如亲脂性和分子大小）一起考虑在内，这会影响其在体内的行为。毒理学潜力的初步预测是基于在计算机工具中，包括QSAR指导的跨模型，分子对接和Cramer分类，以及体外高通量阵列分析，优选使用计算毒性预测模型解释。不利的结果途径可能有助于在此阶段和下一阶段中构造和解释检索到的毒理学信息。初步评估应导致风险的保守估计，发现问题和警报，有必要（或放弃（如果不存在或低于毒理学问题的话）进行更详细的分析，以在后续阶段进行确认和量化。图1.下一代风险评估中的分析流程。在文本中解释了这些步骤。QSAR，定量构效关系；MIE，分子引发事件，适用于不良结局途径；TTC，毒理学关注阈值；POD，出发点。在第二阶段 通过更复杂的（概率）模型，基于暴露媒介物的测得浓度，摄入信息和（人类）生物材料中的测得浓度，可以改善暴露评估。这种精炼的暴露数据可以用作毒物动力学模型的输入，以量化内部暴露浓度，尤其是如第一阶段毒性模型所示的靶组织。第一阶段的毒性警报和预测可以通过更有针对性的，复杂的毒理学模型进行确认，该模型应与人体状况更为相关，例如微生物培养和未经许可的全生物培养（脊椎动物，低等生物）。最终，只有在不得已时，才能使用传统的实验动物进行最终确认。所有这些模型都应提供基于基准的有效剂量，再次通过毒物动力学模型，可以将其推断为对人体有效的剂量。在最终评估中，将第二阶段的结果综合为结论性风险评估，专门针对目标组织和敏感人群。进行案例研究是测试化学风险评估新策略适用性的绝佳方法。本期中描述了一个沿上述流程进行案例研究的例子。（3）研究重点在于对三种三唑的效果分析，结果表明，替代性的毒性测试方法得出的结论与传统方法相似。指定用于目标组织和敏感人群。进行案例研究是测试化学风险评估新策略适用性的绝佳方法。本期中描述了一个沿上述流程进行案例研究的例子。（3）研究重点在于对三种三唑的效果分析，结果表明，替代性的毒性测试方法得出的结论与传统方法相似。指定用于目标组织和敏感人群。进行案例研究是测试化学风险评估新策略适用性的绝佳方法。本期中描述了一个沿上述流程进行案例研究的例子。（3）研究重点在于对三种三唑的效果分析，结果表明，替代性的毒性测试方法得出的结论与传统方法相似。体内测试，为将下一代风险评估概念应用于监管目的提供支持。本文引用了其他3个出版物。