Reproducible proteomics sample preparation for single FFPE tissue slices using acid-labile surfactant and direct trypsinization

Melanie Christine Föll,Martin Werner,Matthias Fahrner,Martin Lothar Biniossek,Markus Kühs,Peter Bronsert,Oliver Schilling,Victor Oginga Oria

doi:10.1186/s12014-018-9188-y

Abstract

BackgroundProteomic analyses of clinical specimens often rely on human tissues preserved through formalin-fixation and paraffin embedding (FFPE). Minimal sample consumption is the key to preserve the integrity of pathological archives but also to deal with minimal invasive core biopsies. This has been achieved by using the acid-labile surfactant RapiGest in combination with a direct trypsinization (DTR) strategy. A critical comparison of the DTR protocol with the most commonly used filter aided sample preparation (FASP) protocol is lacking. Furthermore, it is unknown how common histological stainings influence the outcome of the DTR protocol.MethodsFour single consecutive murine kidney tissue specimens were prepared with the DTR approach or with the FASP protocol using both 10 and 30 k filter devices and analyzed by label-free, quantitative liquid chromatography–tandem mass spectrometry (LC–MS/MS). We compared the different protocols in terms of proteome coverage, relative label-free quantitation, missed cleavages, physicochemical properties and gene ontology term annotations of the proteins. Additionally, we probed compatibility of the DTR protocol for the analysis of common used histological stainings, namely hematoxylin & eosin (H&E), hematoxylin and hemalaun. These were proteomically compared to an unstained control by analyzing four human tonsil FFPE tissue specimens per condition.ResultsOn average, the DTR protocol identified 1841 ± 22 proteins in a single, non-fractionated LC–MS/MS analysis, whereas these numbers were 1857 ± 120 and 1970 ± 28 proteins for the FASP 10 and 30 k protocol. The DTR protocol showed 15% more missed cleavages, which did not adversely affect quantitation and intersample comparability. Hematoxylin or hemalaun staining did not adversely impact the performance of the DTR protocol. A minor perturbation was observed for H&E staining, decreasing overall protein identification by 13%.ConclusionsIn essence, the DTR protocol can keep up with the FASP protocol in terms of qualitative and quantitative reproducibility and performed almost as well in terms of proteome coverage and missed cleavages. We highlight the suitability of the DTR protocol as a viable and straightforward alternative to the FASP protocol for proteomics-based clinical research.

Highlights

Proteomic analyses of clinical specimens often rely on human tissues preserved through formalin-fixation and paraffin embedding (FFPE)
We investigated the compatibility of the standard histological tissue stainings hematoxylin & eosin (H&E), hematoxylin and hemalaun on the direct trypsinization (DTR) protocol, that we consider as a useful tool for studying small FFPE tissues in clinical proteomic research
Non-fractionated LC–liquid chromatography–tandem mass spectrometry (MS)/MS analysis the DTR protocol identified on average 1841 ± 22 proteins whereas slightly higher numbers were found with the filter aided sample preparation (FASP) 10 and 30 k approach, namely 1857 ± 120 and 1970 ± 28 proteins (Fig. 2a)

Summary

Introduction

Proteomic analyses of clinical specimens often rely on human tissues preserved through formalin-fixation and paraffin embedding (FFPE). Minimal sample consumption is the key to preserve the integrity of pathological archives and to deal with minimal invasive core biopsies. For a long time the highly crosslinked FFPE tissue specimens were considered to not be amenable to proteomic studies using liquid chromatography–tandem mass spectrometry (LC–MS/MS). HIAR enabled mass-spectrometry based proteomics studies of FFPE tissues with similar protein extraction efficiencies and numbers of identified peptides as observed for the first time by Hood et al [13]. All SDS based protocols have in common that protein loss is inevitable due to the SDS cleanup step This is typically offset by increasing the amount of input material

Methods

Results

Conclusion