Department of Pharmaceutics, Utrecht Institute for Pharmaceutical Sciences-UIPS, Utrecht University, Universiteitsweg 99, 3584 CG Utrecht, The Netherlands.
Pharm Res. 2013 Jul;30(7):1719-28. doi: 10.1007/s11095-013-1062-z. Epub 2013 May 7.
All therapeutic proteins are potentially immunogenic. Antibodies formed against these drugs can decrease efficacy, leading to drastically increased therapeutic costs and in rare cases to serious and sometimes life threatening side-effects. Many efforts are therefore undertaken to develop therapeutic proteins with minimal immunogenicity. For this, immunogenicity prediction of candidate drugs during early drug development is essential. Several in silico, in vitro and in vivo models are used to predict immunogenicity of drug leads, to modify potentially immunogenic properties and to continue development of drug candidates with expected low immunogenicity. Despite the extensive use of these predictive models, their actual predictive value varies. Important reasons for this uncertainty are the limited/insufficient knowledge on the immune mechanisms underlying immunogenicity of therapeutic proteins, the fact that different predictive models explore different components of the immune system and the lack of an integrated clinical validation. In this review, we discuss the predictive models in use, summarize aspects of immunogenicity that these models predict and explore the merits and the limitations of each of the models.
所有治疗性蛋白都具有潜在的免疫原性。针对这些药物产生的抗体可能会降低疗效，导致治疗成本大幅增加，在极少数情况下还会导致严重甚至危及生命的副作用。因此，人们付出了很多努力来开发免疫原性最小的治疗性蛋白。为此，在药物开发的早期阶段就需要对候选药物的免疫原性进行预测。目前有多种计算、体外和体内模型可用于预测药物先导物的免疫原性，从而改变潜在的免疫原性，并继续开发预期免疫原性低的候选药物。尽管这些预测模型被广泛应用，但它们的实际预测价值存在差异。造成这种不确定性的重要原因是对治疗性蛋白免疫原性相关免疫机制的了解有限/不足，不同的预测模型探索了免疫系统的不同组成部分，以及缺乏综合的临床验证。在这篇综述中，我们讨论了目前使用的预测模型，总结了这些模型所预测的免疫原性的各个方面，并探讨了每种模型的优缺点。
