Hemostasis Branch, Division of Plasma Protein Therapeutics, Office of Tissues and Advanced Therapies, Center for Biologics Evaluation and Research, U.S. Food and Drug Administration, Silver Spring, MD, United States.
Office of Biostatistics & Epidemiology, Center for Biologics Evaluation and Research, U.S. Food and Drug Administration, Silver Spring, MD, United States.
Front Immunol. 2019 Dec 20;10:2894. doi: 10.3389/fimmu.2019.02894. eCollection 2019.
Immune responses to therapeutic proteins and peptides can adversely affect their safety and efficacy; consequently, immunogenicity risk-assessments are part of the development, licensure and clinical use of these products. In most cases the development of anti-drug antibodies is mediated by T cells which requires antigen presentation by Major Histocompatibility Complex Class II (MHCII) molecules (also called Human Leucocyte Antigen, HLA in humans). Immune responses to many protein therapeutics are thus HLA-restricted and it is important that the distribution of HLA variants used in the immunogenicity assessments provides adequate coverage of the target population. Due to biases inherent to the collection of samples in a blood bank or donor pool, simple random sampling will not achieve a truly representative sample of the population of interest. To help select a donor cohort we introduce SampPick, an implementation of simulated annealing which optimizes cohort selection to closely match the frequency distribution of a target population or subpopulation. With inputs of a target background frequency distribution for a population and a set of available, HLA-typed donors, the algorithm will iteratively create a cohort of donors of a user selected size that will closely match the target population rather than a random sample. In addition to optimizing the HLA types of donor cohorts, the software presented can be used to optimize donor cohorts for any other biallelic or monoallelic trait.
治疗性蛋白和肽的免疫反应可能会对它们的安全性和疗效产生不利影响；因此，免疫原性风险评估是这些产品开发、许可和临床应用的一部分。在大多数情况下，抗药物抗体的产生是由 T 细胞介导的，这需要主要组织相容性复合物 II 类 (MHCII) 分子（也称为人类白细胞抗原，HLA）呈递抗原。因此，许多蛋白质治疗药物的免疫反应受到 HLA 的限制，重要的是，免疫原性评估中使用的 HLA 变体的分布为目标人群提供了足够的覆盖。由于血库或供体库中样本采集固有的偏见，简单的随机抽样不会实现对目标人群的真正代表性样本。为了帮助选择供体队列，我们引入了 SampPick，这是一种模拟退火的实现，它可以优化队列选择，以紧密匹配目标人群或亚人群的频率分布。该算法的输入是目标人群的背景频率分布和一组可用的 HLA 分型供体，它将迭代创建一个用户选择大小的供体队列，该队列将紧密匹配目标人群，而不是随机样本。除了优化供体队列的 HLA 类型外，所提出的软件还可用于优化任何其他双等位基因或单等位基因特征的供体队列。
