Wertheim Kenneth Y, Chisholm Robert, Richmond Paul, Walker Dawn
Insigneo Institute for in Silico Medicine, University of Sheffield, Sheffield, United Kingdom.
School of Computer Science, University of Sheffield, Sheffield, United Kingdom.
PLoS Comput Biol. 2024 Dec 23;20(12):e1012648. doi: 10.1371/journal.pcbi.1012648. eCollection 2024 Dec.
Neuroblastoma is the most common extra-cranial solid tumour in children. Over half of all high-risk cases are expected to succumb to the disease even after chemotherapy, surgery, and immunotherapy. Although the importance of MYCN amplification in this disease is indisputable, the mechanistic details remain enigmatic. Here, we present a multicellular model of neuroblastoma comprising a continuous automaton, discrete cell agents, and a centre-based mechanical model, as well as the simulation results we obtained with it. The continuous automaton represents the tumour microenvironment as a grid-like structure, where each voxel is associated with continuous variables such as the oxygen level therein. Each discrete cell agent is defined by several attributes, including its cell cycle position, mutations, gene expression pattern, and more with behaviours such as cell cycling and cell death being stochastically dependent on these attributes. The centre-based mechanical model represents the properties of these agents as physical objects, describing how they repel each other as soft spheres. By implementing a stochastic simulation algorithm on modern GPUs, we simulated the dynamics of over one million neuroblastoma cells over a period of months. Specifically, we set up 1200 heterogeneous tumours and tracked the MYCN-amplified clone's dynamics in each, revealed the conditions that favour its growth, and tested its responses to 5000 drug combinations. Our results are in agreement with those reported in the literature and add new insights into how the MYCN-amplified clone's reproductive advantage in a tumour, its gene expression profile, the tumour's other clones (with different mutations), and the tumour's microenvironment are inter-related. Based on the results, we formulated a hypothesis, which argues that there are two distinct populations of neuroblastoma cells in the tumour; the p53 protein is pro-survival in one and pro-apoptosis in the other. It follows that alternating between inhibiting MDM2 to restore p53 activity and inhibiting ARF to attenuate p53 activity is a promising, if unorthodox, therapeutic strategy. The multicellular model has the advantages of modularity, high resolution, and scalability, making it a potential foundation for creating digital twins of neuroblastoma patients.
神经母细胞瘤是儿童最常见的颅外实体瘤。即使经过化疗、手术和免疫治疗，所有高危病例中仍有超过一半预计会死于该疾病。尽管MYCN扩增在这种疾病中的重要性无可争议，但其机制细节仍然不明。在此，我们展示了一个神经母细胞瘤的多细胞模型，该模型包括一个连续自动机、离散细胞主体和一个基于中心的力学模型，以及我们用它获得的模拟结果。连续自动机将肿瘤微环境表示为类似网格的结构，其中每个体素与诸如其中的氧气水平等连续变量相关联。每个离散细胞主体由几个属性定义，包括其细胞周期位置、突变、基因表达模式等，细胞周期和细胞死亡等行为随机依赖于这些属性。基于中心的力学模型将这些主体的属性表示为物理对象，描述它们如何作为软球体相互排斥。通过在现代图形处理器上实现随机模拟算法，我们在数月的时间内模拟了超过一百万个神经母细胞瘤细胞的动态。具体而言，我们建立了1200个异质性肿瘤，并追踪每个肿瘤中MYCN扩增克隆的动态，揭示有利于其生长的条件，并测试其对5000种药物组合的反应。我们的结果与文献报道一致，并为MYCN扩增克隆在肿瘤中的生殖优势、其基因表达谱、肿瘤的其他克隆（具有不同突变）以及肿瘤微环境之间的相互关系提供了新的见解。基于这些结果，我们提出了一个假设，即肿瘤中存在两种不同的神经母细胞瘤细胞群体；p53蛋白在其中一种中具有促生存作用，而在另一种中具有促凋亡作用。因此，在抑制MDM2以恢复p53活性和抑制ARF以减弱p53活性之间交替是一种有前景的（如果非传统的话）治疗策略。该多细胞模型具有模块化、高分辨率和可扩展性的优点，使其成为创建神经母细胞瘤患者数字孪生体的潜在基础。
