Application of a novel particle coating technology for paediatric formulation development

背景：儿科人群缺乏合适的药物意味着未经许可的制剂往往是由医疗保健专业人员配制和管理的，这导致这一弱势群体面临安全性和临床疗效方面的风险。制约儿科药物开发进展的众多挑战包括对提供剂量灵活性、安全性和商业可行性的剂型缺乏了解。儿科制剂的翻译一直受到限制，因为现有技术有限，无法配制可用于各种大剂量药物的商业可行产品[1]、[2]、[3]。此外，儿科药物中安全包含的辅料选择有限，加上味道可接受性问题，构成了额外的一层配方困难。导致这一建议的工作：阿斯顿开发的干颗粒包衣是一种新的颗粒工程技术，将首次允许通过流态化和对工艺参数的细致精确控制将细颗粒分层在粗颗粒上。这种方法允许高载药量、最低赋形剂载量和可接受的味道，并且可以量身定做以处理小分子和生物制品。在数据分析方面，我们建立了用于数据可视化、数据抽象、概率建模和预测技术以及植物动态分析和控制的新方法[4]、[5]。显然，我们使用机器学习方法的方法，例如用于自动响应面方法的径向基函数网络[6]，以前从未在公式开发中开发过。它在收敛速度和解的精度方面优于现有的响应面方法和群体优化等机器学习方法，并且具有不局限于局部二次曲面的优点。该提案的首要假设是，机器学习的应用将产生用于生产包括生物制品在内的高剂量、脆弱剂型的预测工具。该项目将包括了解干燥包衣过程中的工艺参数对小药物分子和生物制剂(如多肽和蛋白质)稳定性和性能的影响。在优化配方后，该项目的工作将涉及基于细胞培养的转运和毒性研究，以发展对配方在体内性能的机理理解。此外，还将进行包括高效液相色谱、蛋白质/多肽稳定性指示分析、激光衍射研究在内的表征研究，以研究颗粒大小和电荷，以表征粉末混合物。“

"Background: The lack of suitable medicines for the paediatric population means that unlicensed formulations are often prepared and administered by healthcare professionals, which entails exposing this vulnerable population to risks in terms of safety and clinical efficacy. The multitude of challenges restricting progress in paediatric medicine development include lack of understanding of dosage forms that offer dose flexibility, safety and commercial feasibility. The translation of paediatric formulations has been limited, due to the limited availability of technologies to formulate commercially viable products that would be applicable for a wide range of high dose drugs[1],[2],[3]. In addition, a limited choice of excipients for safe inclusion in paediatric medicines, together with taste acceptability issues, presents an additional layer of formulation difficulty.Work leading to this proposal: Dry particle coating developed at Aston is a novel particle engineering technology that, for the first time, will permit layering of fine particles over coarse particles through fluidisation and meticulous precision control of processing parameters. This method allows high drug loading, minimal excipient load and acceptable taste and can be tailored to handle small molecules and biologicals. In data analytics, we have established new approaches for data visualisation, data abstractions, probabilistic modelling and prediction techniques, and analysis and control of plant dynamics[4],[5]. Explicitly, our approach for using machine learning methods, such as radial basis function networks[6] for automated response surface methodology, has never been developed before in formulation development. It is superior to existing machine learning approaches of response surface methodology and swarm optimisation in terms of speed of convergence and accuracy of solution, and has the advantage of not being limited to locally quadratic surfaces. The overarching hypothesis of the proposal is that application of machine learning will produce predictive tools for the manufacture of high dose, fragile dosage forms including biologicals. This project will involve understanding the impact of processing parameters during dry coating on the stability and performance of small drug molecules and biologics such as peptides and proteins. Following the optimisation of formulations, work in this project will involve cell culture based transport and toxicity studies to develop a mechanistic understanding of formulation performance in vivo. Additionally, characterisation studies including HPLC, protein/peptide stability indicating assays, laser diffraction studies to study particle size and charge will be carried out to characterise powder blends. "