Generation of high‐quality protein extracts from formalin‐fixed,paraffin‐embedded tissues

A wealth of information on proteins involved in many aspects of disease is encased within formalin‐fixed paraffin‐embedded (FFPE) tissue repositories stored in hospitals worldwide. Recently, access to this “hidden treasure” is being actively pursued by the application of two main extraction strategies: digestion of the entangled protein matrix with generation of tryptic peptides, or decrosslinking and extraction of full‐length proteins. Here, we describe an optimised method for extraction of full‐length proteins from FFPE tissues. This method builds on the classical “antigen retrieval” technique used for immunohistochemistry, and allows generation of protein extracts with elevated and reproducible yields. In model animal tissues, average yields of 16.3 μg and 86.8 μg of proteins were obtained per 80 mm² tissue slice of formalin‐fixed paraffin‐embedded skeletal muscle and liver, respectively. Protein extracts generated with this method can be used for the reproducible investigation of the proteome with a wide array of techniques. The results obtained by SDS‐PAGE, western immunoblotting, protein arrays, ELISA, and, most importantly, nanoHPLC‐nanoESI‐Q‐TOF MS of FFPE proteins resolved by SDS‐PAGE, are presented and discussed. An evaluation of the extent of modifications introduced on proteins by formalin fixation and crosslink reversal, and their impact on quality of MS results, is also reported.     

Integrated and convenient procedure for protein extraction from formalin‐fixed,paraffin‐embedded tissues for LC‐MS/MS analysis

Because fresh‐frozen tissue samples associated with long‐term clinical data and of rare diseases are often unobtainable at the present time, formalin‐fixed paraffin‐embedded (FFPE) tissue samples are considered a highly valuable resource for researchers. However, protein extraction from FFPE tissues faces challenges of deparaffinization and cross‐link reversion. Current procedures for protein extraction from FFPE tissue require separate steps and toxic solvents, resulting in inconvenience in protein extraction. To overcome these limitations, an integrated method was developed using nontoxic solvents in four types of FFPE tissues. The average amount of proteins from three replicates of bladder, kidney, liver, and lung FFPE tissues were 442.6, 728.9, 736.4, and 694.7 μg with CVs of 7.5, 5.8, 2.4, and 4.5%, respectively. Proteomic analysis showed that 348, 417, 607, and 304 unique proteins were identified and quantified without specification of isoform by a least two peptides from bladder, kidney, liver, and lung FFPE tissue samples, respectively. The analysis of individual protein CV demonstrated that 97–99% of the proteins were quantified with a CV ≤ 30%, verifying the reproducibility of the integrated protein extraction method. In summary, the developed method is high‐yield, reproducible, convenient, simple, low cost, nonvolatile, nonflammable, and nontoxic.     

Validation of a robust proteomic analysis carried out on formalin‐fixed paraffin‐embedded tissues of the pancreas obtained from mouse and human

A number of reports have recently emerged with focus on extraction of proteins from formalin‐fixed paraffin‐embedded (FFPE) tissues for MS analysis; however, reproducibility and robustness as compared to flash frozen controls is generally overlooked. The goal of this study was to identify and validate a practical and highly robust approach for the proteomics analysis of FFPE tissues. FFPE and matched frozen pancreatic tissues obtained from mice (n = 8) were analyzed using 1D‐nanoLC‐MS(MS)² following work up with commercially available kits. The chosen approach for FFPE tissues was found to be highly comparable to that of frozen. In addition, the total number of unique peptides identified between the two groups was highly similar, with 958 identified for FFPE and 1070 identified for frozen, with protein identifications that corresponded by approximately 80%. This approach was then applied to archived human FFPE pancreatic cancer specimens (n = 11) as compared to uninvolved tissues (n = 8), where 47 potential pancreatic ductal adenocarcinoma markers were identified as significantly increased, of which 28 were previously reported. Further, these proteins share strongly overlapping pathway associations to pancreatic cancer that include estrogen receptor α. Together, these data support the validation of an approach for the proteomic analysis of FFPE tissues that is straightforward and highly robust, which can also be effectively applied toward translational studies of disease.     

mTRAQ‐based quantification of potential endometrial carcinoma biomarkers from archived formalin‐fixed paraffin‐embedded tissues

Formalin‐fixed paraffin‐embedded (FFPE) tissues are the primary and preferred medium for archiving patients' samples. Here we demonstrate relative quantifications of protein biomarkers in extracts of laser microdissected epithelial cells from FFPE endometrial carcinoma tissues versus those from normal proliferative endometria by means of targeted proteomic analyses using LC–multiple reaction monitoring (MRM) MS with MRM Tags for Relative and Absolute Quantitation (mTRAQ) labeling. Comparable results of differential expressions for pyruvate kinase isoform M2 (PK‐M2) and polymeric Ig receptor were observed between analyses on laser microdissected epithelial cells from FFPE tissues and corresponding homogenates from frozen tissues of the same individuals that had previously been analyzed and reported. We also identified PK‐M2 in the normal proliferative phase of the endometrium. Other biomarkers in addition to PK‐M2 and polymeric Ig receptor were also observed but not consistently and/or were at levels below the threshold for quantification.     

Proteomic analysis of doxorubicin‐induced changes in the proteome of HepG2cells combining 2‐D DIGE and LC‐MS/MS approaches

HepG‐2 cells are widely used as a cell model to investigate hepatocellular carcinomas and the effect of anticancer drugs such as doxorubicin, an effective antineoplastic agent, which has broad antitumoral activity against many solid and hematological malignancies. To investigate the effect of doxorubicin on the protein pattern, we used complementary proteomic workflows including 2‐D gel‐based and gel‐free methods. The analysis of crude HepG2 cell extracts by 2‐D DIGE provided data on 1835 protein spots which was then complemented by MS‐centered analysis of stable isotope labeling by amino acids in cell culture‐labeled cells. The monitoring of more than 1300 distinct proteins, including proteins of the membrane fraction provides the most comprehensive overview on the proteome of the widely used model cell line HepG2. Of the proteins monitored in total, 155 displayed doxorubicin‐induced changes in abundance. Functional analysis revealed major influences of doxorubicin on proteins involved in protein synthesis, DNA damage control, electron transport/mitochondrial function, and tumor growth. The strongest decrease in level was found for proteins involved in DNA replication and protein synthesis, whereas proteins with a function in DNA damage control and oxidative stress management displayed increased levels following treatment with doxorubicin compared with control cells. Furthermore, the doxorubicin‐associated increase in levels of multiple forms of keratins 8, 18, and 19 and other structural proteins revealed an influence on the cytoskeleton network.     

Detergent addition to tryptic digests and ion mobility separation prior to MS/MS improves peptide yield and protein identification for in situ proteomic investigation of frozen and formalin‐fixed paraffin‐embedded adenocarcinoma tissue sections

The identification of proteins involved in tumour progression or which permit enhanced or novel therapeutic targeting is essential for cancer research. Direct MALDI analysis of tissue sections is rapidly demonstrating its potential for protein imaging and profiling in the investigation of a range of disease states including cancer. MALDI‐mass spectrometry imaging (MALDI‐MSI) has been used here for direct visualisation and in situ characterisation of proteins in breast tumour tissue section samples. Frozen MCF7 breast tumour xenograft and human formalin‐fixed paraffin‐embedded breast cancer tissue sections were used. An improved protocol for on‐tissue trypsin digestion is described incorporating the use of a detergent, which increases the yield of tryptic peptides for both fresh frozen and formalin‐fixed paraffin‐embedded tumour tissue sections. A novel approach combining MALDI‐MSI and ion mobility separation MALDI‐tandem mass spectrometry imaging for improving the detection of low‐abundance proteins that are difficult to detect by direct MALDI‐MSI analysis is described. In situ protein identification was carried out directly from the tissue section by MALDI‐MSI. Numerous protein signals were detected and some proteins including histone H3, H4 and Grp75 that were abundant in the tumour region were identified.     

Development of a monoclonal antibody that specifically detects tissue inhibitor of metalloproteinase‐4 (TIMP‐4) in formalin‐fixed,paraffin‐embedded human tissues

Overexpression of the extracellular metalloproteinase inhibitor TIMP‐4 in estrogen receptor‐negative breast cancers was found recently to be associated with a poor prognosis for survival. To pursue exploration of the theranostic applications of TIMP‐4, specific antibodies with favorable properties for immunohistochemical use and other clinical assays are needed. Here we report the characterization of a monoclonal antibody (clone 9:4–7) specific for full‐length human TIMP‐4 with suitable qualities. The antibody was determined to be an IgG_2b immunoglobulin. In enzyme‐linked immunosorbent assay (ELISA) and immunoblotting assays, it did not exhibit any detectable crossreactivity with recombinant forms of the other human TIMPs 1, 2, and 3. In contrast, the antibody displayed high specificity and sensitivity for TIMP‐4 including in formalin‐fixed and paraffin‐embedded specimens of human breast specimens. An analysis of tissue microarrays of human cancer and corresponding normal tissues revealed specific staining patterns with excellent signal‐to‐noise ratios. This study documents TIMP‐4 monoclonal antibody clone 9:4–7 as an effective tool for preclinical and clinical investigations. J. Cell. Biochem. 110: 1255–1261, 2010. Published 2010 Wiley‐Liss, Inc.     

Proteomics analysis of BHK‐21 cells infected with a fixed strain of rabies virus

Rabies is a neurotropic virus that causes a life threatening acute viral encephalitis. The complex relationship of rabies virus (RV) with the host leads to its replication and spreading toward the neural network, where viral pathogenic effects appeared as neuronal dysfunction. In order to better understand the molecular basis of this relationship, a proteomics study on baby hamster kidney cells infected with challenge virus standard strain of RV was performed. This cell line is an in vitro model for rabies infection and is commonly used for viral seed preparation. The direct effect of the virus on cellular protein machinery was investigated by 2‐DE proteome mapping of infected versus control cells followed by LC‐MS/MS identification. This analysis revealed significant changes in expression of 14 proteins, seven of these proteins were viral and the remaining were host proteins with different known functions: cytoskeletal (capping protein, vimentin), anti‐oxidative stress (superoxide dismutase), regulatory (Stathmin), and protein synthesis (P0). Despite of limited changes appeared upon rabies infection, they present a set of interesting biochemical pathways for further investigation on viral‐host interaction.     

Preparation of C60‐functionalized magnetic silica microspheres for the enrichment of low‐concentration peptides and proteins for MALDI‐TOF MS analysis

In this work, for the first time, a novel C60‐functionalized magnetic silica microsphere (designated C60‐f‐MS) was synthesized by radical polymerization of C60 molecules on the surface of magnetic silica microspheres. The resulting C60‐f‐MS microsphere has magnetite core and thin C60 modified silica shell, which endow them with useful magnetic responsivity and surface affinity toward low‐concentration peptides and proteins. As a result of their excellent magnetic property, the synthesized C60‐f‐MS microspheres can be easily separated from sample solution without ultracentrifuge. The C60‐f‐MS microspheres were successfully applied to the enrichment of low‐concentration peptides in tryptic protein digest and human urine via a MALDI‐TOF MS analysis. Moreover, they were demonstrated to have enrichment efficiency for low‐concentration proteins. Due to the novel materials maintaining excellent magnetic properties and admirable adsorption, the process of enrichment and desalting is very fast (only 5 min), convenient and efficient. As it has been demonstrated in the study, newly developed fullerene‐derivatized magnetic silica materials are superior to those already available in the market. The facile and low‐cost synthesis as well as the convenient and efficient enrichment process of the novel C60‐f‐MS microspheres makes it a promising candidate for isolation of low‐concentration peptides and proteins even in complex biological samples such as serum, plasma, and urine or cell lysate.     

Analysis of membrane–protein complexes of the marine sulfate reducer Desulfobacula toluolica Tol2 by 1D blue native‐PAGE complexome profiling and 2D blue native‐/SDS‐PAGE

Sulfate‐reducing bacteria (SRB) obtain energy from cytoplasmic reduction of sulfate to sulfide involving APS‐reductase (AprAB) and dissimilatory sulfite reductase (DsrAB). These enzymes are predicted to obtain electrons from membrane redox complexes, i.e. the quinone‐interacting membrane‐bound oxidoreductase (QmoABC) and DsrMKJOP complexes. In addition to these conserved complexes, the genomes of SRB encode a large number of other (predicted) membrane redox complexes, the function and actual formation of which is unknown. This study reports the establishment of 1D Blue Native‐PAGE complexome profiling and 2D BN‐/SDS‐PAGE for analysis of the membrane protein complexome of the marine sulfate reducer Desulfobacula toluolica Tol2. Analysis of normalized score profiles of >800 proteins in combination with hierarchical clustering and identification of 2D BN‐/SDS‐PAGE separated spots demonstrated separation of membrane complexes in their native form, e.g. ATP synthase. In addition to the QmoABC and DsrMKJOP complexes, other complexes were detected that constitute the basic membrane complexome of D. toluolica Tol2, e.g. transport proteins (e.g. sodium/sulfate symporters) or redox complexes involved in Na⁺‐based bioenergetics (RnfABCDEG). Notably, size estimation indicates dimer and quadruple formation of the DsrMKJOP complex in vivo. Furthermore, cluster analysis suggests interaction of this complex with a rhodanese‐like protein (Tol2_C05230) possibly representing a periplasmic electron transfer partner for DsrMKJOP.     

iTRAQ‐coupled 2‐D LC‐MS/MS analysis of protein profile associated with HBV‐modulated DNA methylation

The development of hepatocellular carcinoma (HCC) is believed to be associated with multiple risk factors, including the infection of hepatitis B virus (HBV). Based on the analysis of individual genes, evidence has indicated the association between HCC and HBV and has also been expanded to epigenetic regulation, with an involvement of HBV in the DNA methylation of the promoter of cellular target genes leading to changes in their expression. Proteomic study has been widely used to map a comprehensive protein profile, which in turn could provide a better understanding of underlying mechanisms of disease onset. In the present study, we performed a proteomic profiling by using iTRAQ‐coupled 2‐D LC/MS‐MS analysis to identify cellular genes down‐regulated in HBV‐producing HepG2.2.15 cells compared with HepG2 cells. A total of 15 proteins including S100A6 and Annexin A2 were identified by our approach. The significance of these cellular proteins as target of HBV‐mediated epigenetic regulation was supported by our validation assays, including their reactivation in cells treated with 5‐aza‐2′‐deoxycytidine (a DNA methyltransferase inhibitor) by real‐time RT‐PCR and Western blot analysis, as well as the DNA methylation status analysis by bisulfite genome sequencing. Our approach provides a comprehensive analysis of cellular target proteins to HBV‐mediated epigenetic regulation and further analysis should facilitate a better understanding of its involvement in HCC development.     

Proteomic profiling of Cronobacter turicensis 3032, a food‐borne opportunistic pathogen

Members of the genus Cronobacter are opportunistic pathogens for neonates and are often associated with contaminated milk powder formulas. At present little is known about the virulence mechanisms or the natural reservoir of these organisms. The proteome of Cronobacter turicensis 3032, which has recently caused two deaths, was mapped aiming at a better understanding of physiology and putative pathogenic traits of this clinical isolate. Our analyses of extracellular, surface‐associated and whole‐cell proteins by two complementary proteomics approaches, 1D‐SDS‐PAGE combined with LC‐ESI‐MS/MS and 2D‐LC‐MALDI‐TOF/TOF MS, lead to the identification of 832 proteins corresponding to a remarkable 19% of the theoretically expressed protein complement of C. turicensis. The majority of the identified proteins are involved in central metabolic pathways, translation, protein folding and stability. Several putative virulence factors, whose expressions were confirmed by phenotypic assays, could be identified: a macrophage infectivity potentiator involved in C. turicensis persistence in host cells, a superoxide dismutase protecting the pathogen against reactive oxygen species and an enterobactin‐receptor protein for the uptake of siderophore‐bound iron. Most interestingly, a chitinase and a metalloprotease that might act against insects and fungi but no casein hydrolysing enzymes were found, suggesting that there is an environmental natural habitat of C. turicensis 3032.     

On‐target and nanoparticle‐facilitated selective enrichment of peptides and proteins for analysis by MALDI‐MS

Over the course of the last decade, a number of investigators have come to appreciate that the surface of a MALDI target, after suitable modification, can be used for selective enrichment of peptides and proteins. More recently, surface‐modified nanoparticles (NPs) that readily co‐crystallize in MALDI matrix, are not ionized by laser desorption/ionization, and do not interfere with MS have attracted interest as alternatives to surface‐modified targets for selective enrichment of peptides and proteins. Surface‐modified targets and NPs facilitate parallel processing of samples, and when used in conjunction with MALDI mass spectrometers with kHz lasers enable development of high‐throughput proteomics platforms. Targets and NPs for reversed phase and ion exchange retention, selective enrichment of glycopeptides, selective enrichment of phosphopeptides, and immunoaffinity MS are described in conjunction with details regarding their preparation and utility. Commercial availability of the reagents and substrates required to prepare surface‐modified targets and NPs is also discussed.     

Isoforms of wild type proteins often appear as low molecular weight bands on SDS‐PAGE

Immunoblotting, after polyacrylamide gel electrophoresis with sodium dodecyl sulfate (SDS‐PAGE), is a technique commonly used to detect specific proteins. SDS‐PAGE often results in the visualization of protein band(s) in addition to the one expected based on the theoretical molecular mass (TMM) of the protein of interest. To determine the likelihood of additional band(s) being nonspecific, we used liquid chromatography – mass spectrometry to identify proteins that were extracted from bands with the apparent molecular mass (MM) of 40 and 26 kD, originating from protein extracts derived from non‐malignant HEK293 and cancerous MDA‐MB231 (MB231) cells separated using SDS‐PAGE. In total, approximately 57% and 21% of the MS/MS spectra were annotated as peptides in the two cell samples, respectively. Moreover, approximately 24% and 36.2% of the identified proteins from HEK293 and MB231 cells matched their TMMs. Of the identified proteins, 8% from HEK293 and 26% from MB231 had apparent MMs that were larger than predicted, and 67% from HEK293 and 37% from MB231 exhibited smaller MM values than predicted. These revelations suggest that interpretation of the positive bands of immunoblots should be conducted with caution. This study also shows that protein identification performed by mass spectrometry on bands excised from SDS‐PAGE gels could make valuable contributions to the identification of cancer biomarkers, and to cancer‐therapy studies.     

A reference map and identification of porcine testis proteins using 2-DE and MS

The development of the testis is essential for maturation of male mammals. A complete understanding of proteins expressed in the testis will provide biological information on many reproductive dysfunctions in males. The purposes of this study were to apply a proteomic approach to investigating protein composition and to establish a 2-D PAGE reference map for porcine testis proteins. MALDI-TOF MS was performed for protein identification. When 1 mg of total proteins was assayed by 2-D PAGE and stained with colloidal CBB, more than 400 proteins with a pI of pH 3-10 and M(r) of 10-200 kDa could be detected. Protein expression varied among individuals, with CV between 4.7 and 131.5%. A total of 447 protein spots were excised for identification, among which 337 spots were identified by searching the mass spectra against the NCBInr database. Identification of the remaining 110 spots was unsuccessful. A 2-D PAGE-based porcine testis protein database has been constructed on the basis of the results and will be published on the WWW. This database should be valuable for investigating the developmental biology and pathology of porcine testis.     

Optimization of a phenol extraction‐based protein preparation method amenable to downstream 2DE and MALDI‐MS based analysis of bacterial proteomes

2DE is one of the most efficient and widely used methods for resolving complex protein mixtures. For efficient analysis of complex samples, high‐resolution separation of proteins on 2D gel is essential, and for that purpose good sample preparation is crucial. In this study, we have improvized a method for preparing bacterial total cellular proteome, from a strategy applied earlier to recalcitrant plant tissues, which gave high‐quality resolution on 2DE. The method involving phenol extraction followed by methanol/ammonium acetate precipitation was first optimized for the chemolithotrophic proteobacteria Tetrathiobacter kashmirensis WT001 and Pseudaminobacter salicylatoxidans KCT001 that did not yield quality protein preps in conventional trichloroacetic acid/acetone precipitation method. Subsequently, to validate its general applicability, the method was evaluated against the trichloroacetic acid/acetone precipitation method for two other model bacteria, i.e. Escherichia coli DH5α and Mycobacterium smegmatis mc²6. Identification of at least four proteins each from the outer membrane, periplasm, and cytoplasm of T. kashmirensis by MALDI‐MS not only proved the efficiency of the method in extracting proteins from the different cellular compartments but also the amenability of the obtained protein spots toward MALDI‐MS based identification.     

LC‐MS/MS as an alternative for SDS‐PAGE in blue native analysis of protein complexes

Two‐dimensional blue native/SDS‐PAGE is widely applied to investigate native protein–protein interactions, particularly those within membrane multi‐protein complexes. MS has enabled the application of this approach at the proteome scale, typically by analysis of picked protein spots. Here, we investigated the potential of using LC‐MS/MS as an alternative for SDS‐PAGE in blue native (BN) analysis of protein complexes. By subjecting equal slices from BN gel lanes to label‐free semi‐quantitative LC‐MS/MS, we determined an abundance profile for each protein across the BN gel, and used these profiles to identify potentially interacting proteins by protein correlation profiling. We demonstrate the feasibility of this approach by considering the oxidative phosphorylation complexes I–V in the native human embryonic kidney 293 mitochondrial fraction, showing that the method is capable of detecting both the fully assembled complexes as well as assembly/turnover intermediates of complex I (NADH:ubiquinone oxidoreductase). Using protein correlation profiling with a profile for subunits NDUFS2, 3, 7 and 8 we identified multiple proteins possibly involved in the biogenesis of complex I, including the recently implicated chaperone C6ORF66 and a novel candidate, C3ORF60.     

2‐D DIGE analysis implicates cytoskeletal abnormalities in psychiatric disease

The mechanisms underlying white matter changes in psychiatric disease are not known. We aimed to characterise the differential protein expression in deep white matter from the dorsolateral prefrontal cortex from 35 schizophrenia, 35 bipolar disorder, and 35 control subjects, from the Stanley Array Collection. We used 2‐D DIGE to profile for protein expression changes in the brain. We found 70 protein spots to be significantly differentially expressed between disease and control subjects (ANCOVA, p<0.05), 46 of which were subsequently identified by LC‐MS/MS. The proteins identified included novel disease candidates as well as proteins that have previously been reported as abnormal in schizophrenia, thus reinforcing their association with the disease. Furthermore, we confirmed the direction of change for three proteins using ELISA, namely neurofilament‐light, amphiphysin II, and Rab‐GDP‐α, in a subset of the Stanley Array Collection. In addition, altered expression of neurofilament‐light, amphiphysin II, and Rab‐GDP‐α was not observed in the cortex of mice chronically treated with haloperidol, making it less likely that these alterations are a consequence of neuroleptic medication. The data presented here strongly suggest disruption of the cytoskeleton and its associated signal transduction proteins in schizophrenia, and to a lesser extent in bipolar disorder.     

Differential proteomic analysis of white adipose tissues from T2D KKAy mice by LC‐ESI‐QTOF

Type 2 diabetes (T2D) has become a worldwide increasingly social health burden for its high morbidity and heightened prevalence. As one of the main tissues involved in uptake of glucose under the stimulation of insulin, WAT plays very important role in metabolic and homeostasis regulation. We performed a differential proteomics study to investigate alterations in epididymis fat pad of high fat diet fed T2D KKAy mice compared to normal fed C57BL/6J mice, by ¹⁸O‐labeling relative quantitative technique. Among 329 confidently identified proteins, 121 proteins showed significant changes with CV ≤ 20% (fold changes of >2 or <0.5 as threshold). According to GO classification, we found that altered proteins contained members of biological processes of metabolic process, oxidative stress, ion homeostasis, apoptosis and cell division. In metabolic, proteins assigned to fatty acid biosynthesis (FAS etc.) were decreased, the key enzyme (ACOX3) in β‐oxidation process was increased. Increased glycolysis enzymes (ENOB etc.) and decreased TCA cycle related enzymes (SCOT1 etc.) suggested that glucose metabolism in mitochondria of T2D mice might be impaired. Elevated oxidative stress was observed with alterations of a series of oxidordeuctase (QSOX1 etc.). Besides, alterations of ion homeostasis (AT2C1 etc.) proteins were also observed. The enhancement of cell proliferation associated proteins (ELYS etc.) and inhibition of apoptosis associated proteins (RASF6 etc.) in WAT might contributed to the fat pad and body weight gain. Overall, these changes in WAT may serve as a reference for understanding the functional mechanism of T2D.