Unbiased high-throughput MALDI-TOF MS lipid and metabolite screening assays in primary human cells for drug discovery

Matrix-assisted laser/desorption ionization time-of-flight (MALDI-TOF) mass spectrometry (MS) has become a powerful tool for high-throughput screening (HTS) approaches in drug discovery, overcoming the shortcomings of conventional fluorescence label-based technologies. So far, most of the MALDI-TOF based HTS approaches have focused on in vitro assays with rather simple readouts, and have been limited mainly, but not exclusively, to peptide/protein-centric activity assays. Currently, comprehensive and unbiased HTS approaches to track metabolites with MALDI-TOF have not been explored for drug discovery applications. Metabolites are involved in every aspect of biology, and since the metabolome is highly dynamic in nature and a sensitive indicator of phenotype, we can take advantage of this to identify metabolic markers to be applied for treatment response in drug discovery. Although, phenotypic cellular assays using MALDI-TOF MS are possible using higher molecular masses, the capability of MALDI-TOF to detect compounds in the low mass range is generally considered limited due to interference peaks brought by the matrix. Metabolomics-based drug discovery presents therefore an exciting challenge for MS analysis as the system becomes inherently more complex. Herein, we apply this technology for cellular assays, specifically to detect metabolites and lipids in a comprehensive, untargeted, and unbiased HTS approach for drug discovery in idiopathic pulmonary fibrosis (IPF).

Primary human nasal epithelial cells were used to develop a cellular assay pipeline for untargeted metabolite phenotypic identification using MALDI-TOF MS.  Multiple IPF-relevant stimuli and inhibitors were tested to see if stimulation and inhibition could be distinguished in the assay. Next, different sample preparation conditions, such as testing different matrices, additives, derivatization reagents, and extraction protocols, were investigated to ensure the most effective analysis for metabolites and lipids. All the parameters were optimized using Mosquito-TTP Labtech liquid handling robot to allow a systematic screening of a large number of combinations to find the best conditions. Moreover, all the protocols were simplified and adapted to HTS compatible platforms to ensure a smooth translation from an academic laboratory to industry.

Preliminary testing revealed spectra that could be distinguished between the unstimulated, stimulated cells, and stimulated cells with inhibitor, in both the low and high mass region (m/z 200-1000 and 2-20 kDa). Using principal component analysis (PCA), hierarchical clustering, and machine learning strategies, a subset of peaks was identified to be unique to each condition. This data supports that it is possible to elucidate important metabolic features of cells in modelled pathophysiology. This approach has the potential to be further optimized as an automated HTS drug discovery assay in the industrial setting.