|
Abstract Introduction: Central laboratory facilities, large-scale biomarker identification campaigns, and drug discovery studies, all rely on the high-throughput analysis of complex biological samples, from non-volatile buffers with high concentrations of detergents or salts, to tissue biopsies, cell cultures, bacterial colonies, or diverse biofluids. Traditionally the use of mass spectrometry (MS) for the study of these classes of samples has heavily relied on sample preparation and purification approaches prior to analysis. This, although effective at dealing with the complexity of biological matrices, drastically limits the inherent throughput of MS analysis and generates multiple challenges when dealing with large-scale sample cohorts. Here we describe the operation and several applications of an automated robotic platform based on desorption electrospray ionization (DESI) MS for the rapid and direct (i.e. no sample workup needed) analysis of arrays of biological samples, with throughputs better than 1 sample per second. Relevant examples of the screening of biosamples for biomarker discovery, routine bacterial identification, as well as bioactivity assessment using purified targets or cell cultures will be overviewed.
Methods: High-density biological arrays (up to 6,144 samples/array) are rapidly generated on PTFE-coated glass slides using a fluid handling workstation (Beckman Biomek i7 or Hamilton Vantage) equipped with a slotted floating 384 pin-tool. Samples (50 nL) are spotted from 384-well plates either in replicates or from multiple sources, with the spotting being completely user-defined in each experiment. Prepared plates are automatically transferred to a DESI stage and analyzed by MS(/MS) in a spot-to-spot fashion with effective analysis times in the order of 500 ms per spot. The large data volumes produced by this approach are automatically processed in real time by a combination of custom Python- and MATLAB-based software. For biological assays samples are automatically prepared using the fluid handler capabilities or externally with Multidrop Combi setups. For cell studies, prostate cancer cell lines (PC3, LNCaP) were grown and enzymatic target expression was induced. Target inhibition was assessed after a 24-hour incubation followed by media removal and the addition of methanol spiked with an internal standard. All cell cultures were performed in 96-well plates before transfer to the final plate for spotting. For bacteria arrays, single colonies are picked, deposited in 384-well plates pre-filled with solvent (for inactivation of the bacteria and safe analysis in open air), and 50-nL aliquots are spotted onto the PTFE array. Alternatively, analysis of the bacterial colonies can proceed directly from liquid culture media or from solid agar plates after image recognition and route definition algorithms are applied and interface with the control system of the platform.
Results: This general DESI-MS methodology has been applied to a variety of complex samples, including biological assays, tissue biopsies, human cell cultures, and bacterial colonies, with no need for sample work-up. Sub-second analysis times, low sample consumption (typically sub-ng), remarkable matrix tolerance, and excellent quantitative performance are important features of this approach. Several examples including those used for the efficient generation of large spectral libraries, the characterization of enzymatic processes and the evaluation of binding affinities, as well as the profiling and classification of biological specimens will be overviewed. Examples include the label-free characterization of novel potential targets for drug discovery in cancer and neurodegenerative diseases, as well as screening campaigns against such targets for the identification of hits and their subsequent confirmation in cell cultures. Similarly, the use of tissue instead of purified targets to assess overall metabolic pathway dynamics as well as downstream inhibition effects will be overviewed. Additionally, we will showcase the generation of bacterial sample arrays for rapid lipid and metabolite profiling. The unique spectral signatures obtained allow for high-accuracy (>95%) classification of ca. 30 species of microorganisms based entirely on small molecules, in a complementary fashion to commercial protein-based MALDI biotyping but with a simplified workflow and potential expansion to simultaneous identification of enzyme-based antibiotic resistance. Finally, current efforts towards the use of image recognition techniques for the direct analysis of random arrays such as bacterial colony plates, will be discussed.
Discussion/Conclusion: The use of DESI allows for direct analysis of a wide range of biological samples, despite large concentrations of non-volatile salts or detergents, without the need for any sample work-up. This advantage comes from the intrinsic mechanism of DESI which involves a contactless online microextraction event on the surface of analysis. Additionally, the intrinsic control over the sampling event that DESI provides allows for facile and rapid introduction of discrete samples in a high-density array by precise movement of the surface of analysis. These advantages, added to both custom and commercial robotics, software, and analytical instrumentation, make for a powerful automated platform well suited to handle large sample sets, such as that commonplace in central laboratory facilities, drug discovery campaigns, or biomarker screening studies. Additionally, the depth of the data acquired is well suited to provide valuable information, for instance combined with AI approaches, to guide drug development or even clinical decision-making.
|
= Discovery stage. (16.60%, 2024)
= Translation stage. (37.02%, 2024)
= Clinically available. (46.38%, 2024)
