Enabling rapid liquid and freeze-dried formulation design for the manufacture and delivery of novel biopharmaceuticals

生物药物已被批准用于治疗癌症、类风湿性关节炎、多发性硬化症、糖尿病、白血病和中性粒细胞减少症等疾病。下一代基于蛋白质的疗法，或生物制药，是日益复杂的工程形式，具有不可预测的溶液性质。蛋白质以高浓度配制用于临床，经常导致不希望的聚集体形成、高粘度、乳浊度或相分离，使它们不安全，或难以注射或制造。这对生物制药行业是一个重大挑战，因为临床试验中约50%的蛋白质是冷冻干燥的，因为它们不是在开发过程中所需的几个月的时间尺度上随时进行液体配方的。配方是一种使用组合筛选的经验性过程，旨在优化稳定性、效力和给患者的易用性。在伦敦大学学院EPSRC创新制造中心的研讨会上，与36位行业领袖进行了接触，确定了最重要的蛋白质配方挑战，例如在材料不多的情况下预测两年内的货架稳定性。目前加速蛋白质降解和最大限度减少样品消耗的替代技术提供了两年保质期的糟糕指标。行业将显著受益于i)更准确地确定长期保质期的快速分析，ii)指示高浓度溶液行为的低浓度分析，以及iii)更好地使用计算的蛋白质和赋形剂性质来预测最有可能满足所需属性的配方的能力。最先进的自动化微板和微流体分析仪最近通过UCL和UOM的EPSRC和BSRC/BRIC奖购买或建立，为生成跨越许多不同时间的聚集动力学、构象和胶体稳定性、流变性、相变和玻璃化转变温度的大型实验数据提供了一个及时的平台。EPSRC最近在伦敦大学学院资助的GB 500k中试冷冻干燥设备(EP/M028100/1)，结合能源部和3D工艺模拟，将生成冷冻干燥过程模型，阐明将关键工艺参数与新配方的关键质量属性联系起来的机制。从最近的计量单位工作中出现的新型二肽将极大地扩大可供行业接受的配方辅料的范围。新的微流控分析将为配方需求量身定做，带来对配方蛋白质异质性和储存动力学的更早、更灵敏和更低容量的评估，最终以使用密封微孔的高通量格式进行评估。所有数据将填充到单位的网络访问数据库中，以使建模小组能够访问急需的实验数据集。在BioProNet PoC奖的UOM启动的信息学技术将使新的蛋白质能够与数据库中的蛋白质进行比较(通过根据序列和结构计算的性质)，并以预测的方式使用它们通过实验确定的配方行为。计算出的蛋白质性质和关键配方属性之间的相关性将确定赋形剂行为的分子基础。总体而言，这将使生物制药配方界受益，因为它能够：a)确定用于配方筛选的更好的赋形剂组合；b)预测那些最容易在当前辅料和溶液条件下形成的候选蛋白质；c)通过定义的化学修饰为新型多肽辅料的合理设计提供信息；d)使用最小样本从可获得的实验中预测长期储存稳定性和浓缩溶液行为。

Biopharmaceuticals have been approved to treat diseases including cancers, rheumatoid arthritis, multiple sclerosis, diabetes, leukemia and neutropenia. The next-generation of protein-based therapies, or biopharmaceuticals, are of increasingly complex engineered forms, with unpredictable solution properties. Proteins are formulated at high concentrations for clinical use, leading often to undesirable aggregate formation, high viscosity, opalescence, or phase separation, rendering them unsafe, or difficult to inject or manufacture. This is a major challenge to the biopharmaceuticals industry as approximately 50% of proteins in clinical trials have been freeze-dried as they were not readily liquid-formulated on timescales of months required during development. Formulation is an empirical process using combinatorial screens that aim to optimise stability, potency and ease of delivery to patients. Engagement with 36 industry leaders at a UCL EPSRC Centre for Innovative Manufacturing workshop identified the most significant protein formulation challenges such as the prediction of shelf stability over a two year period, at a time in development when not much material is available. Current surrogate techniques that accelerate protein degradation and minimize sample consumption provide poor indicators of 2-year shelf-life. Industry would benefit significantly from i) rapid analyses that more accurately determine long-term shelf-life, ii) low concentration analyses that indicate high-concentration solution behaviour, and iii) a better ability to use calculated protein and excipient properties to predict those formulations that are most likely to meet the required attributes.State-of-the-art automated microplate and microfluidic analytics, purchased or established recently via EPSRC and BBSRC/BRIC awards at UCL and UoM provide a timely platform for generating large experimental datasets of aggregation kinetics spanning many different timescales, conformational and colloidal stabilities, rheological properties, phase-transition and glass transition temperatures, for liquid and freeze-dried formulations. A recent EPSRC funded £500k pilot-scale freeze-drying facility at UCL (EP/M028100/1), combined with DoE and 3D process simulations, will generate freeze-drying process models that elucidate the mechanisms linking critical process parameters to critical quality attributes for new formulations. Novel dipeptides emerging from recent UoM work will significantly expand the range of industry-accepted formulation excipients available. Novel microfluidic analytics will be tailored for formulation needs, bringing earlier, more sensitive, and lower-volume assessments of formulated protein heterogeneity and storage kinetics, ultimately in a high-throughput format using sealed microwells.All data will populate a web-access database at UoM to provide modeling groups access to a much-needed experimental dataset. Informatics techniques initiated at UoM in a BioProNet PoC award will enable new proteins to be compared (via properties calculated from sequence and structure) to those in the database, and use their experimentally determined formulation behaviours in a predictive manner. Correlations between calculated protein properties and critical formulation attributes will identify the molecular basis of excipient behaviour. Overall, this will benefit the biopharmaceutical formulation community with an ability to: a) identify better excipient combinations for input into formulation screens; b) predict those protein candidates most readily formulatable with current excipients and solution conditions; c) inform the rational design of novel peptide-based excipients through defined chemical modifications, d) predict long-term storage stability and concentrated solution behaviour from accessible experiments using minimal sample.