Proteins are the major class of effector molecules in cellular systems. For the identification of functional differences between normal and diseased tissues, a reliable analysis of their protein content is essential. Reproducible isolation and fractionation of intact proteins are important in this respect, but their complexity in structure and concentration, their close interaction, and their instability represent major challenges. For protein isolation in tissues, the breakdown of cell-cell and cell-matrix connections within a tissue without affecting protein quality is a critical factor. We compared different processes for a compartmental protein preparation from pancreatic tissue, one of the most challenging tissues for protein isolation because of its high protease content. Success of the different procedures varied greatly. Based on a scheme of tissue-slicing and subsequent cell isolation, we established a reliable workflow for the fractional extraction of cytosolic proteins, membrane and organelle proteins, nuclear proteins, and cytoskeletal filaments. The tissue slices also allow for a representative confirmation of individual samples’ cellular status by histochemical processes, and a proper separation or mixing of cellular material from across a tumor if required.

Introduction

The challenges of studying the human proteome are numerous due to the huge complexity in size, structure and functionality of proteins; their diverse and frequent interactions and modifications; the enormous dynamic range in protein concentration; and the variation in the abundance at different locations within cells and tissues It is estimated that there are several hundred thousand to several million different human protein molecules (1), and only a minority of these are present in relatively large quantities (2). A reliable protein analysis has to be able to deal with these points sufficiently in order to record a snapshot of the proteomic status of cells or tissues.

A recent increased interest in analyzing the complete protein content of tissues has sparked fresh developments in the area of protein extraction. In order to reduce the complexity of the analysis, sample fractionation methods are being used. Besides the classical centrifugation-techniques (3, 4), chemophysical properties have also been utilized for separation (5). One of these methods is differential detergent fractionation (6), a sequential extraction process with detergent-containing buffers yielding up to four different subproteomic extracts, which are enriched in (i) cytosolic proteins, (ii) membrane and organelle proteins, (iii) nuclear proteins and (iv) cytoskeletal filaments. In addition to reducing fraction complexity, the molecules within each fraction also have more similar biophysical properties. Another advantage is the option to define the subcellular localization of proteins and thus monitor their compartmental redistribution at basal and stimulated conditions. Despite its advantages, differential detergent fractionation was established for cultured cells. Cultured cells are clearly artificial systems, and significant differences between the gene expression profiles of cell cultures and tissues have been reported (7). Additionally, as a consequence of the cell-cell and cell-matrix contacts in tissues, there are many differences in protein extractions between tissue and cultured cells. In tissue, the contacts have to be disconnected and the cells isolated in a manner so that the detergent-containing buffer is able to reach the individual cells. Ideally, this should happen without affecting cell integrity. The nuclear membranes should not be disrupted during the process in order to avoid protein-compartment mixing prior to fractionation.

As part of a large-scale molecular analysis of pancreatic cancer [studying in each sample the methylation patterns of genomic DNA, transcript levels, and protein expression and modification (www.moldiagpaca.eu)], we compared different tissue preparation methods and established a workflow that allows protein extraction from pancreatic tissues. Pancreatic tumors are particularly difficult to handle due to the very high content of protein-degrading enzymes. Since working with various models, we aimed for a process that could be applicable to a number of human and other tissue sources. Initial analyses were therefore performed with pancreatic tissues from rat and pig in order to preserve less-abundant human samples. The established process yielded good-quality protein samples from both cell cultures and animal and human tissues, and the results were reproducible. In addition, the actual tissue samples could be checked for their tumor-cell content by histochemical analysis. Also, the process permits a split of each sample into three identical portions for a parallel analysis at the molecular level of DNA, RNA and protein.

Material and methods

Materials

All chemicals and solvents were purchased in extra-pure grade from Merck (Darmstadt, Germany) unless stated otherwise. Media and other solutions for cell culture were obtained from Invitrogen (Paisley, UK). Collagenase type XI was from Sigma (Taufkirchen, Germany).

Source of tissue samples

Fresh tissues from rat and pig were stored directly after resection in buffer containing protease inhibitors (Complete Mini Protease Inhibitor; Roche, Mannheim, Germany) and used for protein isolation the same day. For frozen porcine tissues, the samples were immediately snap-frozen in liquid nitrogen and stored at −80°C.

Human pancreatic specimens were collected during surgery on pancreatic cancer patients and samples were snap-frozen in liquid nitrogen directly after resection and subsequently stored at −80°C. Written informed consent was obtained from all patients. The study was approved by the local ethics committee of the University of Heidelberg.