Flow injection analysis with diode array absorbance detection and dynamic surface tension detection for studying denaturation and surface activity of globular proteins

Interest of flow injection spectrophotometry as an orthogonal method for analyzing biomolecule aggregates: Application to stressed monoclonal antibody study

This study aimed to explore the suitability of flow injection spectrophotometry (FIS) to analyze three degraded therapeutic monoclonal antibodies (bevacizumab, nivolumab, and rituximab). For this purpose, aggregates were generated with stirring, freeze-thaw, and heat stresses. The intact and stressed mab samples were filtered with 0.22 µm hydrophilic filters and analyzed by size exclusion chromatography (SEC), cation-exchange chromatography (CEX), and FIS. In terms of quantitative and qualitative analysis, protein loss and structural changes were assessed. Various aggregates profiles were obtained according to the mabs and the stresses. FIS allowed performing very satisfactory quantifications for each mab with intermediate precision RSD < 3.0 % and recovery between 97.9 and 102.0 %. From the protein loss measurements, it appears that SEC underestimates the mab aggregate proportions up to two times less as compared with FIS since the latter avoids any non-specific interactions (electrostatic or hydrophobic interactions). Using second derivative spectroscopy and multivariate data analysis, we noticed apparent structural differences, located in the regions 245-265 nm for rituximab and nivolumab and 280-300 nm for bevacizumab, depending on the stress. The FIS complementarity with the other techniques used in this study allowed us to demonstrate that the three mabs behave differently for a given stress condition. While extreme mechanical stress formed large aggregates irrespective of the mabs, rituximab showed to be less stable and more sensitive than the two other mabs under freeze-thaw and heat stresses, generating large aggregates (>200 nm) and partial unfolding. Nivolumab tends to form small aggregates less than 50 nm when heated and freeze-thawed. Moreover, freeze-thaw seems to generate native IgG-1 aggregates with rituximab. Similarly, bevacizumab showed to form these IgG-1 aggregates and was resistant to freeze-thaw, likely thanks to trehalose cryoprotectant from its formulation. Finally, FIS associated with multivariate analysis can provide rich information in one single run and appears to be a fast, simple, and reliable method to set complementary and orthogonal approaches for protein aggregates monitoring.

Simple and ultra-fast recognition and quantitation of compounded monoclonal antibodies: Application to flow injection analysis combined to UV spectroscopy and matching method

Compounding of monoclonal antibody (mAbs) constantly increases in hospital. Quality control (QC) of the compounded mAbs based on quantification and identification is required to prevent potential errors and fast method is needed to manage outpatient chemotherapy administration. A simple and ultra-fast (less than 30 s) method using flow injection analysis associated to least square matching method issued from the analyzer software was performed and evaluated for the routine hospital QC of three compounded mAbs: bevacizumab, infliximab and rituximab. The method was evaluated through qualitative and quantitative parameters. Preliminary analysis of the UV absorption and second derivative spectra of the mAbs allowed us to adapt analytical conditions according to the therapeutic range of the mAbs. In terms of quantitative QC, linearity, accuracy and precision were assessed as specified in ICH guidelines. Very satisfactory recovery was achieved and the RSD (%) of the intermediate precision were less than 1.1%. Qualitative analytical parameters were also evaluated in terms of specificity, sensitivity and global precision through a matrix of confusion. Results showed to be concentration and mAbs dependant and excellent (100%) specificity and sensitivity were reached within specific concentration range. Finally, routine application on "real life" samples (n = 209) from different batch of the three mAbs complied with the specifications of the quality control i.e. excellent identification (100%) and ± 15% of targeting concentration belonging to the calibration range. The successful use of the combination of second derivative spectroscopy and partial least square matching method demonstrated the interest of FIA for the ultra-fast QC of mAbs after compounding using matching method.

Protein conformational transitions at the liquid-gas interface as studied by dilational surface rheology

Experimental results on the dynamic dilational surface elasticity of protein solutions are analyzed and compared. Short reviews of the protein behavior at the liquid-gas interface and the dilational surface rheology precede the main sections of this work. The kinetic dependencies of the surface elasticity differ strongly for the solutions of globular and non-globular proteins. In the latter case these dependencies are similar to those for solutions of non-ionic amphiphilic polymers and have local maxima corresponding to the formation of the distal region of the surface layer (type I). In the former case the dynamic surface elasticity is much higher (>60 mN/m) and the kinetic dependencies are monotonical and similar to the data for aqueous dispersions of solid nanoparticles (type II). The addition of strong denaturants to solutions of bovine serum albumin and β-lactoglobulin results in an abrupt transition from the type II to type I dependencies if the denaturant concentration exceeds a certain critical value. These results give a strong argument in favor of the preservation of the protein globular structure in the course of adsorption without any denaturants. The addition of cationic surfactants also can lead to the non-monotonical kinetic dependencies of the dynamic surface elasticity indicating destruction of the protein tertiary and secondary structures. The addition of anionic surfactants gives similar results only for the protein solutions of high ionic strength. The influence of cationic surfactants on the local maxima of the kinetic dependencies of the dynamic surface elasticity for solutions of a non-globular protein (β-casein) differs from the influence of anionic surfactants due to the heterogeneity of the charge distribution along the protein chain. In this case one can use small admixtures of ionic surfactants as probes of the adsorption mechanism. The effect of polyelectrolytes on the kinetic dependencies of the dynamic surface elasticity of protein solutions is weaker than the effect of conventional surfactants but exceeds the error limits.

Batch foam separation of a soluble protein

Removal of protein dissolved in water by batch foam separation was conducted with using ovalbumin (OA) as a model protein in the light of wastewater treatment reducing organic loading. The removal efficiency had a maximum value near the i.e.p. of OA (pH 4.6); thus, most experiments were conducted at pH 4.6. Typical experimental conditions; superficial gas velocity, U(g): 1.97 x 10(-2)-5.37 x 10(-2)cm/s; initial bulk concentration of OA, C(i): ca. 0.05-0.25 g/L; liquid volume, V: 600 cm(3). A model estimating bulk concentration profile was proposed by taking into account a mass balance of the present system. The model predicted that OA could be removed perfectly, however, was not all removed experimentally. The residual OA concentration of the bulk liquid within the column reached plateau value, which correspond to ca. 18% of the initial OA concentration. The plateau value of the bulk concentration was attained for ca. 100-500 min with U(g)=1.97 x 10(-2)-5.37 x 10(-2)cm/s. Foaming ability test revealed that the foaming limit concentration of OA at pH 4.6 was 9.72 x 10(-3)g/L. These results suggested that OA molecules could be damaged by interaction of bubble surface in the dispersed phase, since there were the residual OA concentrations over the limit concentration. To take account of this phenomena and correct the model, average surface density, X(d), which should convert protein molecule into the denatured protein molecule, was introduced. The corrected model could explain well the time profile of OA bulk concentration.