Characterisation of spray-dried protein/carbohydrate formulations using the combination technique of water vapour sorption and near infrared spectroscopy

Protein drug formulations are commonly produced in the solid form due to their degradation in solution. Drying may alter protein conformation and cause loss of activity. Stabilising excipients, particularly disaccharides, are used to protect proteins during drying. The mechanism of this protection remains unclear. The aims of this work were to investigate the use of the combined technique of dynamic vapour sorption and near infrared spectroscopy (DVS/NIRS) for the analysis of spray-dried protein/trehalose formulations and to examine the mechanism of protein stabilisation by trehalose upon drying. Anhydrous forms of trehalose (a disaccharide) were produced according to methods in the literature and by drying in the DVS analyser. Characterisation was performed using Differential Scanning Calorimetry (DSC), X-Ray Powder Diffraction analysis (XRPD) and Thermogravimetric Analysis (TGA). NIR spectra of samples were recorded. A new anhydrous form of trehalose was produced in the DVS analyser. Assignment of peaks in the NIR spectra for the polymorphic forms of trehalose was achieved. Spray-dried trehalose was prepared by spray-drying solutions of trehalose dihydrate in water, of varying concentration, using parameters appropriate for the drying of proteins. Crystallisation was induced at 75% RH and NIR spectra recorded during the experiments. Spray-dried trehalose was also produced using parameters commonly used for drying sugars. Crystallisation was induced by repetitive exposure to 75% RH in the DVS analyser and NIR spectra were recorded. Variability was observed in the behaviour of trehalose samples dried from solutions of low trehalose concentration. Small amounts of anhydrous trehalose in otherwise amorphous samples were identified by NIRS. The cause of variability/anhydrous nature was proposed to be the less uniform droplets atomised from solutions of low concentration. DVS/NIR allowed the examination of the crystallisation of amorphous trehalose in real-time, which was shown not to be instantaneous. The presence of anhydrous trehalose in otherwise amorphous samples was proposed to act as a seed for subsequent crystallisation, causing the formation of an unstable dihydrate with a tendency to an anhydrous state. Co-spray-dried samples of catalase (a protein) and trehalose were prepared and DVS/NIR experiments were performed in the same way as for spray-dried trehalose samples. The activity of the samples was determined before and after the experiments. Mathematically produced theoretical NIR spectra were compared with the spectra of co-spray-dried samples. A lower ratio of catalase: trehalose was required for the effective stabilisation of catalase during drying. Upon exposure to 75% RH, (mimicking storage), the presence of trehalose in the formulation was detrimental to the stability of the protein. Catalase: trehalose 50:50 was the most effective ratio of components for the overall stabilisation of catalase during/following spray drying. The data supported the water replacement hypothesis of protein stabilisation upon drying because the greatest interaction between components was expected at a 50:50 ratio and evidence of hydrogen bonding between co-spray-dried components was shown in the NIR spectra. Multiple Linear Regression (MLR) and Partial Least Squares Regression (PLSR) were used to determine the feasibility of NIRS for the quantification of components in co- spray-dried catalase/trehalose formulations. Feasibility was demonstrated and PLSR gave a more specific calibration model. The simultaneous use of DVS/NIRS was shown to be very useful to analyse sample transitions in real-time. The method allowed conclusions to be drawn that would be difficult, or impossible to arrive at by the use of either method in isolation