Automation Highlights from the Literature

Abstract

In multiple industries, including the biopharmaceutical industry, automation is synonymous with increased productivity. Environments with high-throughput needs commonly employ automation for efficiency. However, in a discovery bioanalytical ligand-binding assay laboratory setting where the focus is not necessarily on sample analysis throughput but instead on assay development and characterization, is automation applicable? Can automation enhance productivity when tasks are more customized than routine? In this Perspective, the authors review the different categories of automation with ligand-binding assays with these questions in mind. In considering whether automation technology has progressed far enough to result in a positive return on investment in the discovery setting, the resource investment required to operate in this space was contrasted with the gain in productivity. The authors say that technology advancements in automated technology platforms, and especially personal automation, have allowed these categories to strike the right balance for investment in the discovery laboratory setting. (Leung, S. S., Dreher, E. A., Bioanalysis 2013, 5, 1775–1782) With the identification of vast numbers of novel proteins through genomic and proteomic initiatives, the need for efficient processes to characterize and target them has increased. Antibodies are naturally designed molecules that can fulfill this need, and in vitro methodologies for isolating them from either immune or naïve sources have been extensively developed. However, access to pure protein antigens for screening purposes is a major hurdle because of the limitations associated with recombinant production of eukaryotic proteins. Consequently, rational peptide design based on proteomic methodologies such as protein modeling, secondary sequence prediction, and hydrophobicity/hydrophilicity prediction, in combination with other bioinformatics data, is being explored as a viable solution to isolate specific antibodies against difficult antigens. Single-domain antibodies are becoming the ideal antibody format due to their structural advantages and ease of production compared with conventional antibodies and antibody fragments derived from conventional antibodies. For screening purposes, phage display technology is a well-established technique. With this technique, a repertoire of antibody fragments can be displayed on the surface of filamentous phages (f1, fd, M13) followed by screening against various antigenic targets. Furthermore, the technique can be expanded to a high-throughput scale using a magnetic-based, in-solution panning protocol, which allows for the screening of multiple target antigens simultaneously. In this chapter, Kumaran et al. describe a semiautomated panning method to screen a naïve Camelidae library against rationally designed peptide antigens, followed by preliminary characterization of isolated binders. (Kumaran, J., et al., Methods Mol. Biol. 2012, 911, 105–124) Capillary gel electrophoresis in the presence of sodium dodecyl sulfate is a well-established and widely used protein analysis technique in the biotechnology industry and is increasingly becoming the method of choice that meets the requirements of the standards of the International Conference of Harmonization. Automated single-channel capillary electrophoresis systems are usually equipped with ultraviolet absorbance and/or laser-induced fluorescent detection options offering general applicability and high detection sensitivity, respectively, but with limited throughput. This shortcoming is addressed by the use of multicapillary gel electrophoresis (mCGE) systems with LED-induced fluorescent detection, also featuring automation and excellent detection sensitivity, thus widely applicable to rapid and large-scale analysis of biotherapeutics, especially monoclonal antibodies. The methodology the authors report in this paper is readily applicable for rapid purity assessment and subunit characterization of IgG molecules including detection of nonglycosylated heavy chains and separation of possible subunit variations such as truncated light chains or alternative splice variants. Covalent fluorophore derivatization and the mCGE analysis of the labeled IgG samples with multicapillary gel electrophoresis are thoroughly described. Both reducing and nonreducing conditions are applied with and without peptide N-glycosidase F mediated deglycosylation. (Szekrenyes, A., et al., Anal. Bioanal. Chem. 2012, 404, 1485–1494) High-throughput proteomics is made possible by a combination of modern mass spectrometry instruments capable of generating many millions of tandem mass (MS(2)) spectra on a daily basis and the increasingly sophisticated associated software for their automated identification. Despite the growing accumulation of collections of identified spectra and the regular generation of MS(2) data from related peptides, the mainstream approach for peptide identification is still the nearly two decades old approach of matching one MS(2) spectrum at a time against a database of protein sequences. Moreover, database search tools overwhelmingly continue to require that users guess in advance a small set of four to six posttranslational modifications that may be present in their data in order to avoid incurring substantial false-positive and -negative rates. The spectral networks paradigm for analysis of MS(2) spectra differs from the mainstream database search paradigm in three fundamental ways. First, spectral networks are based on matching spectra against other spectra instead of against protein sequences. Second, spectral networks find spectra from related peptides even before considering their possible identifications. Third, spectral networks determine consensus identifications from sets of spectra from related peptides instead of separately attempting to identify one spectrum at a time. Even though spectral networks algorithms are still in their infancy, they have already delivered the longest and most accurate de novo sequences to date, revealed a new route for the discovery of unexpected posttranslational modifications and highly-modified peptides, enabled automated sequencing of cyclic nonribosomal peptides with unknown amino acids, and are now defining a novel approach for mapping the entire molecular output of biological systems that is suitable for analysis with tandem mass spectrometry. Here, Guthais reviews the current state of spectral networks algorithms and discusses possible future directions for automated interpretation of spectra from any class of molecules. (Guthais, A., Mol. Biosyst. 2012, 8, 2535–2544) Generation and interpretation of biotransformation data on drugs (i.e., identification of physiologically relevant metabolites, defining metabolic pathways and elucidation of metabolite structures) have become increasingly important to the drug development process. Profiling using (14)C or (3)H radiolabel is defined as the chromatographic separation and quantification of drug-related material in a given biological sample derived from an in vitro, preclinical in vivo, or clinical study. Metabolite profiling is a very time-intensive activity, particularly for preclinical in vivo or clinical studies that have defined limitations on radiation burden and exposure levels. A clear gap exists for certain studies that do not require specialized high-volume automation technologies, yet these studies would still clearly benefit from automation. Use of radiolabeled compounds in preclinical and clinical ADME studies, specifically for metabolite profiling and identification, is a very good example. The current lack of automation for measuring low-level radioactivity in metabolite profiling requires substantial capacity, personal attention, and resources from laboratory scientists. To help address these challenges and improve efficiency, Krauser et al. have innovated, developed, and implemented a novel and flexible automation platform that integrates a robotic plate-handling platform, HPLC or UPLC system, mass spectrometer, and an automated fraction collector. (Krauser, J., et al., PLoS One 2012, 7, e39070) Lipid extraction from biological samples is a critical and often tedious preanalytical step in lipid research. Primarily on the basis of automation criteria, Lofgren et al. have developed the butanol:methanol (BUME) method, a novel chloroform-free total lipid extraction method for blood plasma compatible with standard 96-well robots. In just 60 min, 96 samples can be automatically extracted with lipid profiles of commonly analyzed lipid classes almost identically and with absolute recoveries similar or better to what is obtained using the chloroform-based reference method. Lipid recoveries are linear from 10 to 100 µL plasma for all investigated lipids using the developed extraction protocol. The BUME protocol includes an initial one-phase extraction of plasma into a 300 µL BUME mixture (3:1) followed by two-phase extraction into 300 µL heptane:ethyl acetate (3:1) using 300 µL 1% acetic acid as buffer. The lipids investigated include the most abundant plasma lipid classes (e.g., cholesterol ester, free cholesterol, triacylglycerol, phosphatidylcholine, and sphingomyelin) as well as less abundant but biologically important lipid classes, including ceramide, diacylglycerol, and lyso-phospholipids. This novel method has been successfully implemented in the authors’ laboratory and is now used daily. The authors conclude that the fully automated, high-throughput BUME method can replace chloroform-based methods, saving both human and environmental resources. (Lofgren, L., et al., J. Lipid. Res. 2012, 53, 1690–1700) Glycoproteome contains valuable information as to where biomarkers may be discovered for disease diagnosis and monitoring. Nowadays, with the ever-increasing performances of mass spectrometers, the emphasis is shifting to sample preparation for better throughput and reproducibility. Therefore, to facilitate high-throughput N-linked glycopeptide isolation, in this study, a novel hydrazide tip is devised and an integrated workflow of N-linked glycopeptide isolation using hydrazide tips is presented. With the use of bovine fetuin as a standard glycoprotein, the incubation time is determined for each major step of glycopeptide isolation. With the use of commercially available human serum, multiple parallel isolations of glycopeptides are performed using hydrazide tips with a liquid-handling robotic system. Chen et al. demonstrate that, with the hydrazide tips, the processing time is significantly decreased from 3 to 4 days to less than 8 h with excellent reproducibility. The hydrazide pipet tips have great potential in achieving automation of N-linked glycopeptide isolation for high-throughput sample preparation when used in combination with liquid-handling robotic systems. (Chen, J. et al., Anal. Chem. 2013, 85, 10670–10674) Real-time tracking of implanted fiducials in cine megavoltage (MV) imaging during volumetric modulated arc therapy (VMAT) delivery is complicated because of the inherent low contrast of MV images and the potential blockage of dynamic leaves configurations. The purpose of the work reported in this paper is to develop a clinically practical autodetection algorithm for motion management during VMAT. The expected field-specific segments and the planned fiducial position from the Eclipse (Varian Medical Systems, Palo Alto, CA) treatment-planning system are projected onto the MV images. The fiducials are enhanced by applying a Laplacian of Gaussian filter in the spatial domain for each image, with a blob-shaped object as the impulse response. The search of implanted fiducials is then performed on a region of interest centered on the projection of the fiducial when it is within an open field, including the case in which it is close to the field edge or partially occluded by the leaves. A universal template formula is proposed for template matching, and normalized cross-correlation is employed for its simplicity and computational efficiency. The search region for every image is adaptively updated through a prediction model that employs the 3D position of the fiducial estimated from the localized positions in previous images. This prediction model allows the actual fiducial position to be tracked dynamically and is used to initialize the search region. The artifacts caused by electronic interference during the acquisition are effectively removed. A score map is computed by combining both morphological information and image intensity. The pixel location with the highest score is selected as the detected fiducial position. The sets of cine MV images taken during treatment are analyzed with in-house–developed software written in MATLAB (The Mathworks, Inc., Natick, MA). Five prostate patients are analyzed to assess the algorithm’s performance by measuring their positioning accuracy during treatment. The algorithm is able to localize accurately the fiducial position on MV images with success rates of more than 90% per case. The percentage of images in which each fiducial is localized in the studied cases varied between 23% and 65%, with at least one fiducial having been localized between 40% and 95% of the images. This depends mainly on the modulation of the plan and fiducial blockage. The prostate movement in the presented cases varies between 0.8 and 3.5 mm (mean values). The maximum displacement detected among all patients is of 5.7 mm. An algorithm for automatic detection of fiducial markers in cine MV images has been developed and tested with five clinical cases. Despite the challenges posed by complex beam aperture shapes, fiducial localization close to the field edge, partial occlusion of fiducials, fast leaf and gantry movement, and inherently low MV image quality, good localization results are achieved in patient images. This work provides a technique for enabling real-time accurate fiducial detection and tumor tracking during VMAT treatments without the use of an extra imaging dose. (Azcona, J. D., et al., Med. Phys. 2013, 40, 031708) Surface microscopy of individual biological cells is essential for determining the patterns of cell migration to study the tumor formation or metastasis. This paper presents a correlated and effective theoretical and experimental technique to automatically address the biophysical and mechanical properties and acquire live images of biological cells that are of interest in studying cancer. In the theoretical part, a distributed-parameters model as the comprehensive representation of the microcantilever is presented along with a model of the contact force as a function of the indentation depth and mechanical properties of the biological sample. Analysis of the transfer function of the whole system in the frequency domain is carried out to characterize the stiffness and damping coefficients of the sample. In the experimental section, unlike the conventional atomic force microscope techniques that basically use the laser for determining the deflection of microcantilever’s tip, a piezoresistive microcantilever serving as a force sensor is implemented to produce the appropriate voltage and measure the deflection of the microcantilever. A micromanipulator robotic system is integrated with the MATLAB and programmed to automatically control the microcantilever mounted on the tip of the micromanipulator to achieve the topography of biological samples including the human corneal cells. For this purpose, the human primary corneal fibroblasts are extracted and adhered on a sterilized culture dish and prepared to attain their topographical image. This proposed methodology allows an approach to obtain 2D quality images of cells being comparatively cost-effective and extendable to obtain 3D images of individual cells. The characterized mechanical properties of the human corneal cell are furthermore established by comparing and validating the phase shift of the theoretical and experimental results of the frequency response. (Eslami, S., Rev. Sci. Instrum. 2012, 83, 105002). A single-cell transcriptome contains reliable gene regulatory relationships because gene-gene interactions happen within only a mammalian cell. Although the study of gene-gene interactions enables the understanding of the molecular mechanism of cellular events and the evaluation of molecular characteristics of a mammalian cell population, its complexity requires an analysis of a large number of single cells at various stages. However, many existing microfluidic platforms cannot process single cells effectively for routine molecular analysis. To address these challenges, Chen et al. develop an integrated system with an individual controller for effective single-cell transcriptome analysis. In this paper, the authors report an integrated microfluidic approach to rapidly measure gene expression in individual cells for genetic stability assessment of a cell population. Inside this integrated microfluidic device, the cells are individually manipulated and isolated in an array using micro sieve structures, then transferred into different nanoliter reaction chambers for parallel processing of single-cell transcriptome analysis. This device enables the manipulation of individual single cells into a nanoliter reactor with high recovery rates. Chen et al. perform gene expression analysis for a large number of HeLa cells and 293T cells expanded from a single cell. The data show that even the housekeeping genes are expressed at heterogeneous levels within a clone of cells. The heterogeneity of actin expression reflects the genetic stability, and the expression distribution is different between cancer cells (HeLa) and immortalized 293T cells. The result demonstrates that this platform has the potential for assessment of genetic stability in cancer diagnosis. (Chen, Y., et al., Lab Chip 2012, 12, 3930–3935) In this study, fine bubbles are successfully generated and used as a simple, low-cost driving force for mixing fluids in an integrated microfluidic bead-based enzyme-linked immunosorbent assay (ELISA) to rapidly and quantitatively detect apolipoprotein A1 (APOA1), a biomarker highly correlated with bladder cancer. A wooden gas diffuser is embedded underneath a microfluidic chip to refine injected air and generate bubbles of less than 0.3 mm. The rising bubbles cause disturbances and convection in the fluid, increasing the probability of analyte interaction. This setup not only simplifies the micromixer design but also achieves rapid mixing with a small airflow as a force. Lin et al. use this bubble-driven micromixer in a bead-based ELISA that targets APOA1. The results indicate that this micromixer reduces the time for each incubation from 60 min in the conventional assay to 8 min with the chip, resulting in a reduction of total ELISA reaction time from 3-4 h to 30-40 min. Furthermore, the concentration detection limit is 9.16 ng/mL, which is lower than the detection cutoff value (11.16 ng/mL) for bladder cancer diagnosis reported in the literature. Therefore, this chip can be used to achieve rapid low-cost bladder cancer detection and may be used in point-of-care cancer monitoring. (Lin, Y. H., et al., Biomed. Microdevices. 2013) Since the first Food and Drug Administration approval of a PEGylated product in 1990, so-called random PEGylation reactions are still used to increase the efficacy of biopharmaceuticals and represent the major technology of all approved PEG-modified drugs. However, the great influence of process parameters on PEGylation degree and the PEG-binding site results in a lack of reaction specificity that can have a severe impact on the product profile. Consequently, reproducible and well-characterized processes are essential to meet increasing regulative requirements resulting from the quality-by-design initiative, especially for this kind of modification type. In this study, Maiser et al. present a general approach that combines the simple chemistry of random PEGylation reactions with high-throughput experimentation (HTE) to achieve a well-defined process. Robotic-based batch experiments are established in a 96-well plate format and analyzed to investigate the influence of different PEGylation conditions for lysozyme as model protein. With common SEC analytics, highly reproducible reaction kinetics are measured, and a significant influence of PEG excess, buffer pH, and reaction time can be investigated. Additional mono-PEG-lysozyme analytics show the impact of varying buffer pH on the isoform distribution, which allows the identification of optimal process parameters to get a maximum concentration of each isoform. Employing Micrococcus lysodeikticus–based activity assays, PEG-lysozyme33 is identified to be the isoform with the highest residual activity, followed by PEG-lysozyme1. Based on these results, a control space for a PEGylation reaction is defined with respect to an optimal overall volumetric activity of mono-PEG-lysozyme isoform mixtures. (Maiser, B., et al., Biotechnol. Bioeng. 2013) Biodegradable nanoparticles have been receiving attention for pharmaceutical applications as well as applications in the food industry. Weiss et al. demonstrate chip electrophoresis of fluorescently (FL) labeled gelatin nanoparticles (gelatin NPs) on a commercially available instrument. FL labeling includes a step for the removal of low-molecular-mass material (especially excess dye molecules). Nevertheless, for the investigated gelatin NP preparation, two analyte peaks, one very homogeneous with an electrophoretic net mobility of µ = –24.6 ± 0.3 × 10(–9) m(2)/Vs at the peak apex (n = 17) and another more heterogeneous peak with µ between approximately –27.2 ± 0.2 × 10(–9) m(2)/Vs and –36.6 ± 0.2 × 10(–9) m(2)/Vs at the peak beginning and end point (n = 11, respectively), are recorded. Filtration allows enrichment of particles in the size range of approximately 35 nm (pore size employed for concentration of gelatin NPs) to 200 nm (pore size employed during FL labeling). This corresponds to the very homogeneous peak, linking it to gelatin NPs, whereas the more heterogeneous peak probably corresponds to gelatin not cross-linked to such a high degree (NP building blocks). Several further gelatin NP preparations are analyzed according to the same protocol, yielding peaks with electrophoretic net mobilities between –23.3 ± 0.3 × 10(–9) m(2)/Vs and –28.9 ± 0.2 × 10(–9) m(2)/Vs at peak apexes (n = 15 and 6). Chip electrophoresis allows analyte separation in less than 2 min (including electrophoretic sample injection). Together with the high sensitivity of the FL detection, the LOD as derived for the first main peak of the applied dye from the threefold standard deviation of the background noise values of 80 pM for determined separation conditions leads to a very promising high-throughput separation technique, especially for the analysis of bionanoparticles. For gelatin NP preparations, chip electrophoresis allows, for example, the comparison of preparation batches concerning the amount of NPs and gelatin building blocks as well as the indirect assessment of the degree of gelatin cross-linking (from obtained FL signals). (Weiss, V. U., et al., Electrophoresis 2013, 34, 2152–2161) Advances in human antibody discovery allow for the selection of hundreds of high-affinity antibodies against many therapeutically relevant targets. This has necessitated the development of reproducible, high-throughput analytical techniques to characterize the output from these selections. Among these characterizations, epitopic coverage and affinity are among the most critical properties for lead identification. Biolayer interferometry (BLI) is an attractive technique for epitope binning due to its speed and low antigen consumption. Although surface-based methods such as BLI and surface plasmon resonance (SPR) are commonly used for affinity determinations, sensor chemistry and surface-related artifacts can limit the accuracy of high-affinity measurements. When comparing BLI and solution equilibrium–based kinetic exclusion assays, significant differences in measured affinity (10-fold and above) are observed. KinExA direct association (k(a)) rate constant measurements suggest that this is mainly caused by inaccurate k(a) measurements associated with BLI-related surface phenomena. Based on the kinetic exclusion assay principle used for KinExA, Estep et al. developed a high-throughput 96-well plate format assay, using a Meso Scale Discovery instrument to measure solution equilibrium affinity. This improved method combines the accuracy of solution-based methods with the throughput formerly achievable only with surface-based methods. (Estep, P., et al., MAbs 2013, 5, 270–278) Over the past decades, microfabricated bioanalytical platforms have gained enormous interest because of their potential to revolutionize biological analytics. Their popularity is based on several key properties, such as high flexibility of design, low sample consumption, rapid analysis time, and minimization of manual handling steps, which are of interest for proteomics analyses. An ideal totally integrated chip-based microfluidic device could allow rapid automated workflows starting from cell cultivation and ending with MS-based proteome analysis. By reducing or eliminating sample handling and transfer steps and increasing the throughput of analyses, these workflows would dramatically improve the reliability, reproducibility, and throughput of proteomic investigations. Although these complete devices do not exist for routine use yet, many improvements have been made in the translation of proteomic sample-handling and separation steps into microfluidic formats. In this review, Chao and Hansmeier focus on recent developments and strategies to enable and integrate proteomic workflows into microfluidic devices. (Chao, T. C., and Hansmeier, N., Proteomics 2013, 13, 467–479) Research on plant metabolism is currently experiencing the common use of various omics methods creating valuable information on the concentrations of the cell’s constituents. However, little is known about in vivo reaction rates, which can be determined by metabolic flux analysis (MFA), a combination of isotope-labeling experiments and computer modeling of the metabolic network. Large-scale applications of this method so far have been hampered by tedious procedures of tissue culture, analytics, modeling, and simulation. By streamlining the workflow of MFA, the throughput of the method could be significantly increased. Huege et al. propose strategies for these improvements on various substeps that will move flux analysis to the medium-throughput range and closer to established methods such as metabolite profiling. Furthermore, this may enable novel applications of MFA, for example, screening plant populations for traits related to the flux phenotype. (Huege, J., et al., Mol. Biosyst. 2012, 8, 2466–2469) Mathematical deconvolution methods can separate co-eluting peaks in samples for which (chromatographic) separation fails. However, these methods often heavily rely on manual user input and interpretation. This is not only time-consuming but also error-prone, and automation is needed if such methods are to be applied in a routine manner. One major hurdle when automating deconvolution methods is the selection of the correct number of components used for building the model. Peters et al. propose a new method for the automatic determination of the optimum number of components when applying multivariate curve resolution (MCR) to comprehensive two-dimensional gas chromatography–mass spectrometry (GC×GC-MS) data. It is based on a twofold cross-validation scheme. The obtained overall cross-validation error decreases when adding components and increases again once overfitting of the data starts to occur. The turning point indicates that the optimum number of components has been reached. Overall, the method is at least as good as and sometimes superior to the inspection of the eigenvalues when performing singular-value decomposition. However, its strong point is that it can be fully automated and is thus more efficient and less prone to subjective interpretation. The method is applied to two different-sized regions in a GC×GC-MS chromatogram. In both regions, the cross-validation scheme results in selecting the correct number of components for applying MCR. The pure concentration and mass spectral profiles obtained are then used for identification and/or quantification of the compounds. Although the method is for applying MCR to GC×GC-MS data, a transfer to other deconvolution methods and other analytical systems should require only minor modifications. (Peters, S., et al., Anal. Chim. Acta. 2013, 799, 29–35) Rhodococcus opacus strain PD630 (R. opacus PD630) is an oleaginous bacterium and also one of few prokaryotic organisms that contain lipid droplets (LDs). LD is an important organelle for lipid storage but also intercellular communication regarding energy metabolism and yet is a poorly understood cellular organelle. To understand the dynamics of LD using a simple model organism, Chen et al. conduct a series of comprehensive omics studies of R. opacus PD630, including complete genome, transcriptome, and proteome analysis. The genome of R. opacus PD630 encodes 8947 genes that are significantly enriched in the lipid transport, synthesis, and metabolic, indicating a super ability of carbon source biosynthesis and catabolism. The comparative transcriptome analysis from three culture conditions reveals the landscape of gene-altered expressions responsible for lipid accumulation. The LD proteomes further identify the proteins that mediate lipid synthesis, storage and other biological functions. Integrating these three omics uncovers 177 proteins that may be involved in lipid metabolism and LD dynamics. An LD structurelike protein LPD06283 is further verified to affect the LD morphology. The omics studies provide not only a first integrated omics study of prokaryotic LD organelle but also a systematic platform for facilitating further prokaryotic LD research and biofuel development. (Chen, Y., et al., Nucleic Acids Res. 2013)