Department of Electrical and Electronic Engineering, Victorian Research Laboratory, National ICT Australia, The University of Melbourne, Victoria 3010, Australia.
J R Soc Interface. 2010 Jul 6;7(48):1049-59. doi: 10.1098/rsif.2009.0488. Epub 2010 Jan 6.
During the adaptive immune response, lymphocyte populations undergo a characteristic three-phase process: expansion through a series of cell divisions; cessation of expansion; and, finally, most of the accumulated lymphocytes die by apoptosis. The data used, thus far, to inform understanding of these processes, both in vitro and in vivo, are taken from flow cytometry experiments. One significant drawback of flow cytometry is that individual cells cannot be tracked, so that it is not possible to investigate interdependencies in the fate of cells within a family tree. This deficit in experimental information has recently been overcome by Hawkins et al. (Hawkins et al. 2009 Proc. Natl Acad. Sci. USA 106, 13 457-13 462 (doi:10.1073/pnas.0905629106)), who reported on time-lapse microscopy experiments in which B-cells were stimulated through the TLR-9 receptor. Cells stimulated in this way do not aggregate, so that data regarding family trees can be recorded. In this article, we further investigate the Hawkins et al. data. Our conclusions are striking: in order to explain the familial correlation structure in division times, death times and propensity to divide, a minimum of two distinct heritable factors are necessary. As the data show that two distinct factors are necessary, we develop a stochastic model that has two heritable factors and demonstrate that it can reproduce the key features of the data. This model shows that two heritable factors are sufficient. These deductions have a clear impact upon biological understanding of the adaptive immune response. They also necessitate changes to the fundamental premises behind the tools developed by statisticians to draw deductions from flow cytometry data. Finally, they affect the mathematical modelling paradigms that are used to study these systems, as these are widely developed based on assumptions of cellular independence that are not accurate.
在适应性免疫反应中，淋巴细胞群体经历一个特征性的三阶段过程：通过一系列细胞分裂进行扩增；停止扩增；最后，大部分积累的淋巴细胞通过细胞凋亡死亡。迄今为止，用于了解这些过程的资料，无论是体外还是体内，都来自流式细胞术实验。流式细胞术的一个显著缺点是单个细胞无法被追踪，因此无法研究细胞在家族树中的命运之间的相互依存关系。Hawkins 等人最近克服了这种实验信息的不足（Hawkins 等人，2009 年《美国国家科学院院刊》106,13457-13462（doi:10.1073/pnas.0905629106）），他们报告了关于通过 TLR-9 受体刺激 B 细胞的延时显微镜实验。以这种方式刺激的细胞不会聚集，因此可以记录关于家族树的数据。在本文中，我们进一步研究了 Hawkins 等人的数据。我们的结论令人震惊：为了解释分裂时间、死亡时间和分裂倾向的家族相关性结构，至少需要两个不同的可遗传因素。由于数据显示需要两个不同的因素，我们开发了一个具有两个可遗传因素的随机模型，并证明它可以再现数据的关键特征。该模型表明，两个可遗传因素就足够了。这些推论对适应性免疫反应的生物学理解有明显的影响。它们还需要改变统计学家开发的从流式细胞术数据中得出推论的工具背后的基本前提。最后，它们影响用于研究这些系统的数学建模范例，因为这些范例广泛基于不准确的细胞独立性假设。
