Paving the way to 3D spatial omics: Methodological evolution of nanosecond infrared laser (NIRL)-based tissue sampling and further processing for low-input mass spectrometric analysis

Abstract

In recent years, there has been a significant increase in the number of mass spectrometry-based spatial omics studies. Considering the spatial resolution allows for a deeper understanding of the cellular organization and tissue interactions, whereby physiological and pathological processes can be investigated more comprehensively. Despite the growing demand, there is currently a lack of suitable workflows in which both the spatial information in the tissue is retained during sampling and comprehensive coverage is achieved in the subsequent proteome and lipidome analysis. A novel approach utilizes short pulse mid-infrared laser ablation for the simultaneous sampling and homogenization of tissues. Biomolecules are released from their cellular tissue compartments via a gentle and cold vaporization process. Moreover, this technology preserves the spatial information of the sample location and provides highly precise sampling by maintaining the integrity of the surrounding tissue area. In this dissertation, the methodology of nanosecond infrared laser (NIRL)-based tissue sampling and subsequent processing for low-input mass spectrometric omics was further developed, paving the way for an application in three-dimensional spatial omics. Given that common sample processing workflows for bottom-up proteomics and shotgun lipidomics necessitate the use of tissue amounts of several milligrams for robust processing, workflows were developed enabling the robust processing of miniaturized tissue volumes of 500 nL, corresponding to low-input amounts of 500 μg. The implementation of a novel aerosol collection method, as described in the study performed by Hahn & Moritz et al., enabled the successful bottom-up proteome analysis of miniaturized tissue volumes of murine spleen and colon tissue. This resulted in the identification of 1,889 proteins with quantitative information. In comparison to a previous PIRL-based study conducted by our research group, in which a 50-fold larger tissue volume was utilized, a comparable number of proteins was identified. The results of differential quantitative proteome analysis revealed significantly different protein abundances in the tissues displaying expected proteomes of spleen and colon tissue. Accordingly, the applicability of NIRL ablation for soft and muscle-rich tissue types was confirmed. In a further methodological development, NIRL-based tissue sampling was combined with quantitative shotgun lipidome analysis, for the first time. In the study of Stadlhofer & Moritz et al., biopsies from human squamous cell carcinomas of the oropharynx (OPSCC) and surrounding mucosa tissue were ablated by NIRL. In total, 755 lipid species from 13 lipid classes were quantified. The data demonstrated not only intra- but also interpatient alterations in lipid composition. The findings confirm similarities in the lipid profiles of the OPSCC samples from different tumor locations. For instance, the lipid classes phosphatidylcholine (PC) and phosphatidylethanolamine (PE) were found to be present in higher concentrations in the OPSCC samples than in the respective non-tumorous oropharyngeal tissue. In contrast, mucosa samples from the base of tongue and tonsil exhibit distinct lipid compositions. Since comprehensive analysis with high coverage was achieved for both bottom-up proteomics and shotgun lipidomics, the suitability of the developed processing workflows for the sampled tissue volumes was verified and provided the foundation for further development towards spatial omics. In the study of Voß & Moritz et al., NIRL ablation was utilized for the tissue sampling of three different colon parts (ascending, transverse, descending) and for layer-by-layer sampling of eight consecutive intestinal tissue layers. The samples were processed by the previously developed bottom-up proteomics workflow. The optimization of the experimental setup further enhanced the number of identified proteins to 3,053 proteins for the region-resolved and 2,882 proteins for the layer-resolved proteome analysis. Based on quantitative bottom-up proteome analysis, colon regions were distinguished by significantly differentially abundant proteins mainly corresponding to metabolic processes. Moreover, for the first time, spatially resolved proteome analysis of murine intestine was performed with an axial resolution of 117 μm. While the top and bottom ablation layers could be clearly assigned to the mucosa and the muscularis propria respectively, the cellular identity of the intermediate layers could not be verified, possibly due to insufficient spatial resolution. To further improve the spatial resolution, the sample processing protocol was optimized by minimizing sample loss due to potential adsorption effects. The ablation volume, which is necessary for reproducible and global analysis, was reduced by a factor of 20, from 600 nL to 30 nL. Accordingly, the study by Navolić & Moritz et al. allowed for the spatially resolved proteome analysis of intact embryonic mouse head with an even higher axial resolution of 40 μm. Targeting the forebrain region, nine consecutive tissue layers were directly ablated from the scalp to the cerebral cortex. Despite the significantly reduced ablation volume, 5,031 proteins were identified including 3,126 proteins with quantitative information, underlining the successful method optimization. While the superficial layers were clearly assigned to skin and bone structures, specific marker proteins for the meninges were also identified, which are difficult to access by conventional techniques. Moreover, gradual changes in spatial protein abundances were analyzed in the highly complex cortex lamination at developmental stage, for the first time. According to this, potential new marker proteins for the cortical layers were proposed. To conclude, the methodological evolution of NIRL-based tissue sampling and further processing for low-input mass spectrometric analysis has enabled the successful application for 3D spatial omics. The region of interest can be sampled directly from intact fresh-frozen tissues, whereby the threedimensional spatial information is maintained. As this outstanding technique has no limitation in tissue type and the layer-by-layer sampling is performed independently of the given tissue architecture, spatially resolved analysis is not only enabled for layered but also complex and sensitive tissue structures. The future incorporation of image guidance will enable cross-validation and targeted ablation, therefore further strengthening the application of NIRL-based tissue sampling for 3D spatial omics and paving the way for answering research questions, which previously remained unanswered due to methodological limitations.