Proteomic Analysis of PAXgene-Fixed Tissues Bilge Ergin,† Stephan Meding,‡ Rupert Langer,† Marcel Kap,§ Christian Viertler,| Christina Schott,† Uta Ferch,⊥ Peter Riegman,§ Kurt Zatloukal,| Axel Walch,‡ and Karl-Friedrich Becker*,† Institute of Pathology, Technische Universita¨t Mu ¨ nchen, Munich, Germany, Institute of Pathology, Helmholtz Center Munich, Neuherberg, Germany, Department of Pathology, Josephine Nefkens Institute, Rotterdam, The Netherlands, Institute of Pathology, Medical University of Graz, Graz, Austria, and Third Medical Department, Technische Universita¨t Mu ¨ nchen, Munich, Germany Received May 18, 2010
Formalin fixation and paraffin embedding is the standard technique for preserving biological material for both storage and histological analysis. Although recent progress has been made in the molecular analysis of formalin-fixed, paraffin-embedded (FFPE) tissues, proteomic applications are a special challenge due to the cross-linking property of formalin. Here we present the results of a new formalinfree tissue fixative, PAXgene, and demonstrate successful extraction of nondegraded and immunoreactive protein for subsequent standard protein assays, such as Western blot analysis and reversephase protein arrays. High amounts of protein can be obtained from PAXgene-fixed, paraffin-embedded (PFPE) mouse liver and human spleen, breast, duodenum, and stomach tissues, similar to frozen material. By Western blot analysis, we found that the detection of membrane, cytoplasmic, nuclear, and phosphorylated protein from PAXgene-fixed human tissue samples was comparable to cryopreserved samples. Furthermore, the distribution of protein in PAXgene-fixed human tissue specimens is adequate for matrix-assisted laser desorption/ionization (MALDI) imaging mass spectrometry for in situ proteomic analysis. Taken together, we demonstrate here that PAXgene has great potential to serve as a novel multimodal fixative for modern pathology, enabling extensive protein biomarker studies on clinical tissue samples. Keywords: PAXgene • alternative fixative • proteomics • MALDI imaging MS • RPPA • modern pathology
Introduction In recent years, molecular diagnostics and personalized medicine have become the major fields of investigation in cancer research. In addition to classical in vitro studies using cell culture systems, analysis of human specimens is required for the identification of novel biomarkers and targets for more specific cancer therapies. To date, the most abundant supply of human samples is represented by formalin-fixed and paraffin-embedded (FFPE) tissues, which are stored in huge archives at surgical pathology departments all over the world. The fixation of clinical tissues using neutral buffed formalin preserves the detailed cellular morphology, and the heat-induced antigen retrieval (HIAR) technique enables high-quality immunostaining by unmasking epitopes.1 However, molecular analyses are affected by the cross-linking property of formaldehyde. In FFPE tissue, nucleic acids are often fragmented, and the recovery of DNA and RNA from FFPE samples is limited, * Corresponding author. Technische Universita¨t Muenchen, Institut fu ¨ r Pathologie, Trogerstrasse 18, D-81675 Mu ¨ nchen. E-mail: kf.becker@ lrz.tu-muenchen.de. † Institute of Pathology, Technische Universita¨t Mu ¨ nchen. ‡ Helmholtz Center Munich. § Josephine Nefkens Institute. | Medical University of Graz. ⊥ Third Medical Department, Technische Universita¨t Mu ¨ nchen.
5188 Journal of Proteome Research 2010, 9, 5188–5196 Published on Web 09/02/2010
which impairs downstream high-quality molecular analyses.2 In addition, formalin fixation results in protein-protein crosslinking that limits the application of proteomic analysis. So far, formalin-fixed patient specimens have only been appropriate for a restricted number of protein approaches in routine clinical settings, including immunohistochemistry, although optimized extraction protocols for protein from FFPE tissue are available.3-6 Besides Western blotting7 and reversephase protein microarrays (RPPA),3 attempts have been made to develop protocols that make FFPE tissues accessible for matrix-assisted laser desorption/ionization (MALDI) imaging. Using fresh frozen material, MALDI imaging mass spectrometry (MS) is a powerful tool for investigating the localization and expression profile of proteins and small molecules within tissue sections while still preserving the histomorphological integrity of the tissue.8,9 By using this approach, cell type-specific proteomic patterns can be obtained, even in complex tissues. Recently, it was reported that in situ enzymatic digestion and automated microspotting of the MALDI matrix permitted spatial resolution of tissue mass spectrometry using FFPE material.10 However, it seemed that reliable biomarker discovery is only possible if nondegraded and chemically unmodified proteins are present in the tissue section. 10.1021/pr100664e
2010 American Chemical Society
Proteomic Analysis of PAXgene-Fixed Tissue Cryopreserved samples may represent the best choice at first glance to identify new biomarkers. The biomolecules are not chemically modified by the fixative, and subsequent molecular profile studies can be performed using high-throughput analyses. However, the architectural and morphological details are not preserved accurately, and consequently, reliable histological analysis may be impaired. Thus, the sample acquisition and longtime storage is complicated and cost intensive. With the recent interest in personalized medicine, there is a strong need for a new fixative that would enable the molecular analyses together with a reliable classical histological diagnosis from the same sample. There are several alternative fixatives to formalin on the market that have been proposed for standard molecular analysis, but none of these has replaced formalin in modern pathology to date.11-13 The PAXgene tissue system used here is a two-step approach that is commercially available (PreAnalytiX GmbH, Hilden, Germany). These new reagents are a mixture of different alcohols including methanol, an acid, and a soluble organic compound. Both solutions dehydrate and denature proteins and, in combination, prevent the shrinkage of tissue. After rapid fixation in PAXgene Tissue Fixation Reagent, the tissue is transferred into PAXgene Tissue Stabilization Reagent, which preserves the morphology and keeps the nucleic acids intact in the fixed tissue before paraffin embedding. As a partner of the EU project SPIDIA (www.spidia.eu), we aim to evaluate newly developed tissue-preservation techniques for their use in pathological institutes. We addressed the major fields of molecular pathology (morphology, immunohistochemistry, nucleic acid analysis, and protein analysis) with regard to the applicability of the new fixative. In this study, we present the first data for the extraction of nondegraded and immunoreactive proteins from PAXgene-fixed tissue samples for downstream applications, such as Western blotting and RPPA. In addition, we show that PAXgene can be utilized for reliable in situ proteomic analysis of human pancreatic tissue by MALDI imaging MS without digestion pretreatment. Altogether, our data demonstrate the potential of PAXgene to become a widely used tissue fixation system suitable for the extensive proteomic studies of clinical tissue samples.
Material and Methods Tissue Samples. Mouse Tissue. C57BL/6 mice were maintained at the Technische Universita¨t Mu ¨ nchen in accordance with national and institutional guidelines for animal care. Animals were used at 6-12 weeks of age. Mice were sacrificed, and the liver was dissected as short as 2 min post mortem and divided into equal parts. Samples were either snap-frozen in liquid nitrogen, fixed for 3 and 24 h in PAXgene Tissue Fixation Reagent (PreAnalytiX GmbH, Germany), or fixed for 24 h in formalin. The PAXgene-fixed samples were transferred into PAXgene Tissue Stabilization Reagent for 24 h at room temperature or as indicated in the text. After dehydration of PAXgene-fixed and formalin-fixed liver tissue, the samples were embedded in low melting paraffin. The protocol consists of stepwise dehydration in 70%, 80%, 90%, and 99% ethanol (two baths), followed by isopropanol (two baths) and xylene (two baths). PAXgene-fixed and formalin-fixed, paraffin-embedded tissue blocks were stored at 4 °C in the dark. All experiments with mice were performed in accordance with the guidelines for the use of living animals in scientific studies and the German Law for the Protection of Animals.
research articles Human Tissues. Nonmalignant human specimens from the spleen, breast, duodenum, kidney, pancreas, and stomach were equally divided into four samples and either snap-frozen, PAXgene-fixed (3 or 24 h fixation and 24 h stabilization), or formalin-fixed (24 h) and processed as described above for mouse tissue. Reference hematoxylin/eosin stained sections were histologically verified by a board-certified pathologist (R.L.). Subsequent unstained sections of the same paraffin blocks were used for protein extraction (see below). All patients gave informed consent, and the study was approved by the Ethics Committee of the Klinikum rechts der Isar of the Technische Universita¨t Mu ¨nchen, Germany (reference number 2336/09). PAXgene Tissue System. The novel two-step tissue fixation system, composed of the PAXgene Tissue Fixation Reagent and the PAXgene Tissue Stabilization Reagent, is commercially available for research use only (PreAnalytiX, Hilden, Germany). The system is based on a mixture of different alcohols including methanol (toxic, flammable), acetic acid (irritant), and a soluble organic compound. The fixation is carried out without the cross-linking of biomolecules, therefore intending to stabilize biomolecule profiles such as proteins and nucleic acids while also retaining histomorphology and antigenicity. Both solutions precipitate and denature proteins and together prevent tissue shrinkage and preserve nucleic acids. The term “PAXgene” is used throughout this manuscript instead of “PAXgene Tissue System” for readability. Protein Extraction. Protein was extracted according to a modified QProteome FFPE Tissue-Kit (Qiagen GmbH, Hilden, Germany) protocol. Briefly, after standardized deparaffinization of 10 µm tissue sections in xylene and rehydration in graded alcohol series, tissue was scratched from unstained slides using a needle. For PFPE tissue and snap-frozen specimens, the lysed samples were incubated for 15 min in extraction buffer on ice, followed by heating at 70 °C for 2 h with agitation. After centrifugation, the supernatant was stored at -20 °C. Protein from FFPE tissues was extracted as previously described3 according to the QProteome FFPE Tissue-Kit (Qiagen GmbH, Hilden, Germany). Protein Yield Calculation. Protein concentrations were determined using the Bradford reagent. The area of the section was measured using Image J (see http://rsbweb.nih.gov/ij/ docs/pdfs/ImageJ), and the yield was calculated as follows total protein amount ) X µg/mm3 area × section (thickness × number of slides) One-Dimensional SDS-PAGE and Western Blot. Equal amounts of protein lysates were separated by one-dimensional SDS-PAGE and visualized using Coomassie blue staining. For Western blotting, proteins were transferred onto the nitrocellulose membrane, and standard immunoblotting was performed using anti-β-actin (Sigma-Aldrich, Steinheim, Germany), anti-p-ERK, anti-ERK (Cell Signaling, Frankfurt, Germany), anti-E-cadherin (BD Transduction Laboratories, San Jose, USA), and anti-Hsp70 (Abcam, Cambridge, U.K.) antibodies. Proteins were visualized using the ECLplus detection reagent (Amersham/GE Healthcare Europe GmbH, Freiburg, Germany). The nitrocellulose membrane was also stained with Ponceau S to detect the protein pattern blotted onto the membrane. Reverse-Phase Protein Microarrays. Reverse-phase protein microarrays (RPPA) were generated using the BioOdyssey Journal of Proteome Research • Vol. 9, No. 10, 2010 5189
research articles
Ergin et al.
Figure 1. Schematic overview of the fixation and extraction protocols. (A) Tissue samples were equally divided and fixed in either liquid nitrogen, PAXgene, or formalin. Stabilized PAXgene samples and formalin-fixed samples were paraffin-embedded. Subsequently, sections were cut from the resulting tissue blocks for protein extraction. (B) For protein extraction, three different protocols (I-III) were applied after the deparaffinization of sections: (I) 30 min on ice, (II) 15 min on ice and 2 h at 70 °C, or (III) 15 min on ice, 20 min in a water bath at 100 °C, and 2 h at 80 °C. After centrifugation, the samples were stored at -20 °C.
Calligrapher MiniArrayer according to the manufacturer’s instructions (BioRad, Hercules, USA). Human spleen lysates were spotted onto a nitrocellulose-coated glass slide (Oncyte Avid, Grace Bio-Laboratories, Bend, USA) using serial dilutions (1:2, 1:4, 1:8, 1:16) with three replicates for each dilution. The immunodetection for ERK was performed as described for the Western blot. To estimate total protein amounts, parallel arrays were stained with Sypro Ruby Protein Blot Stain (Molecular Probes, Eugene, USA) according to the manufacturer’s instructions and visualized on an Eagle Eye (Stratagene, La Jolla, USA). MALDI Imaging MS. MALDI imaging mass spectrometry (MALDI IMS) sample preparation and measurement was done as previously described.14 All chemicals, unless stated differently, were of pro analysi grade and purchased from SigmaAldrich, Steinheim, Germany. Cryopreserved pancreatic tissue was sectioned (12 µm; CM1950, Leica Microsystems, Wetzlar, Germany), thaw mounted onto conductive slides (Bruker Daltonik, Bremen, Germany), and washed with 70% and 100% ethanol for 1 min each. PAXgene- and formalin-preserved pancreatic tissue was sectioned (3.5 µm; HM340E, Microm International, Walldorf, Germany), mounted onto conductive slides (Bruker Daltonik, Bremen, Germany), and rehydrated by consecutive incubation (5 min) in xylene (twice), isopropanol, and 100%, 90%, 70%, and 50% ethanol, followed by drying (40 °C, 10 min). Sinapinic acid was applied as a matrix using the ImagePrep spray device (Bruker Daltonik, Bremen, Germany) according to the manufacturer’s protocol. MALDI imaging MS was performed on the Ultraflex III MALDI-TOF/TOF with FlexControl 3.0 and FlexImaging 2.1 software (Bruker Daltonik, Bremen, Germany) in positive linear mode with a detection range of m/z 2000-25 000, a sampling rate of 0.1 GS/s, a lateral resolution of 70 µm, and 200 laser shots per measuring point. A protein standard (Bruker Daltonik, Bremen, Germany) was spotted adjacent to the tissue sections and was used for spectra calibration. The matrix was rinsed off with 70% ethanol, and the slides were stained with hematoxylin and eosin (H&E), scanned with the MIRAX DESK system (Carl Zeiss MicroImaging, Go¨ttingen, Germany), and coregistered with the MALDI IMS results to correlate mass spectrometric data with the histological features of the same section. 5190
Journal of Proteome Research • Vol. 9, No. 10, 2010
RNA Extraction, PCR, and Gel Electrophoresis. For RNA extraction, 10-20 sections of 5 µm in thickness from either snap-frozen, PAXgene-fixed, or formalin-fixed and paraffinembedded human nonmalignant liver tissues were deparaffinized. RNA was extracted according to manufacturer’s instructions using the PAXgene Tissue RNA Kit for PAXgene-fixed, the RNeasy FFPE Kit (both Qiagen, Hilden, Germany) for FFPE samples, or the Trizol procedure (Invitrogen, Darmstadt, Germany) for cryopreserved samples. RNA was eluted in a final volume of 40 µL of kit buffer or RNase-free water. RNA yields were determined using a Nanodrop ND-1000 UV spectrophotometer (Nanodrop Technologies, Wilmington, Germany). Electropherograms were obtained from Agilent Bioanalyzer 2100 (Agilent, Palo Alto, USA) using the Agilent 6000 Nano Chip Kit, and the Agilent 2100 Expert software was used to calculate the RNA integrity number (RIN). RNA (1 µg) from each sample was transcribed into cDNA using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, USA) according to manufacturer’s instructions. An amount of 5 µL of a 1:20 cDNA dilution served as a template for PCR using SYBR Green PCR Master Mix (Applied Biosystems, Foster City, USA). PCR products were separated on a 2% agarose gel and were visualized by the addition of ethidium bromide. For the amplification of a 153-, 323-, and 530-base pair (bp) fragment of the human glyceraldehyde-3-phosphate dehydrogenase (GAPDH), the following primers were used: GAPDH forward: 5′-CCA CAT CGC TCA GAC ACC AT-3′, GAPDH reverse 153 bp: 5′-GTA AAC CAT GTA GTT GAG GTC-3′, GAPDH reverse 323 bp: 5′-AAG ACG CCA GTG GAC TCC A-3′; and GAPDH reverse 530 bp: 5′-ACG ATA CCA AAG TTG TCA TG-3′, for 45 cycles of PCR. All samples and no template controls were analyzed in triplicates. For depicting the degradation of RNA, total RNA was separated on a formaldehyde-containing 1% agarose gel and visualized by ethidium bromide.
Results Proteins Can Be Efficiently Recovered from PAXgeneFixed Tissue. To compare protein extraction efficiency, cryopreserved, formalin-fixed, and PAXgene-fixed tissue samples
Proteomic Analysis of PAXgene-Fixed Tissue
research articles
Figure 2. Evaluation of the protocol for protein extraction from PFPE tissue. Heating scratched tissues from glass slides in extraction buffer for 2 h at 70 °C is sufficient to extract protein from PAXgene-fixed tissue but not from FFPE tissue. (A) Mouse liver tissue was either cryopreserved, PAXgene-fixed (3 and 24 h), and paraffin-embedded or formalin-fixed and paraffin-embedded. Protein was extracted according to the protocols shown, and protein yield was calculated as described in the Materials and Methods section. (B) Liver lysates were separated on a one-dimensional gel and stained with Coomassie blue. (C) Comparison of the final extraction protocols of frozen, PAXgene, and formalin-fixed stomach specimen. After separation on a one-dimensional gel and blotting, proteins on nitrocellulose membrane were stained with Ponceau S. Ice ) protocol I; 70 ° C ) protocol II; 100 °C ) protocol III.
were analyzed in parallel (Figure 1A). Three different protocols were initially tested to evaluate the optimal extraction method for protein from PAXgene-fixed tissue (Figure 1B). Therefore, PAXgene-fixed and formalin-fixed samples (both paraffinembedded) and cryopreserved samples from mouse liver tissue were processed in parallel through three incubation steps (I, 15 min on ice; II, 70 °C for 2 h; III, 20 min at 100 °C and 2 h at 80 °C). The quantitative analysis of extracted protein showed that all of the protocols could be performed with frozen samples to recover a high amount of protein. Boiling of the samples, however, resulted in a decreased yield from frozen material (Figure 2A). For PAXgene-fixed samples, a reliably high amount of protein when compared to frozen samples could be extracted using all three protocols, and heating during extraction even elevated protein recovery. In contrast to FFPE tissue, incubation of the PAXgene tissue lysate on ice, representing protocol I, resulted in a recovery of protein from PAXgene tissue. PAXgene fixation for 24 h seemed to decrease the protein amount, but this effect was not consistent and varied between different tissue types (Supporting Information,
Figure 1). For FFPE tissues, as expected, only extended heating (protocol III) resulted in sufficient protein recovery. However, the yield was lower when compared to frozen or PAXgene-fixed samples. These experiments clearly show that PAXgene fixation allows for the extraction of a high quantity of protein from a variety of mouse and human tissues. PAXgene-Fixed and Cryopreserved Tissues Revealed Similar Protein Pattern. To test the pattern of extracted protein from PAXgene-fixed, paraffin-embedded (PFPE) tissue, the lysate was separated by SDS-PAGE, and the protein profile was visualized by Coomassie staining. As seen in Figure 2B, the general profile of protein bands in the PAXgene-fixed samples is comparable to that of frozen and superior to formalin-fixed tissues. Distinct bands of full-length mouse liver proteins were observed in PAXgene tissue in the range of 25-250 kDa for protocol II and III, both including heating, where only few bands were visible for FFPE tissue. Overall, protein can be reliably extracted from PAXgene-fixed, paraffin-embedded tissue, and the protein ladder is clearly visible with 2 h of heating at 70 °C. Therefore, protocol II was chosen for protein extracJournal of Proteome Research • Vol. 9, No. 10, 2010 5191
research articles
Figure 3. Effect of stabilization time on protein recovery. Protein recovery was slightly reduced for tissue that had been stabilized for 48 h when compared to 2 h. Mouse liver tissue was fixed in PAXgene for 3 h and incubated in PAXgene Tissue Stabilizer Reagent for 2, 24, or 48 h at room temperature. Protein was extracted, and the protein amount was determined using the Bradford assay. The areas of HE-stained sections were measured, and the micrograms of total protein per cubed millimeter of tissue was calculated. Error bars show the standard deviation of four independent samples.
tion from frozen and PAXgene samples. As protein from FFPE specimens can only be sufficiently recovered after extended heating, protocol III was used to extract protein from formalinfixed tissues. We then applied the optimal protocols for protein extraction (protocol II for frozen and PAXgene-fixed samples and protocol III for FFPE samples) to human stomach tissue. The separation of the protein visualized by Ponceau S staining of the nitrocellulose membrane clearly showed a comparable protein pattern for frozen, PAXgene, and FFPE tissue when using the optimized protein extraction protocols for each fixative (Figure 2C). In the next step, a possible effect of the PAXgene stabilization solution on protein recovery was investigated. After 3 h of PAXgene fixation, mouse liver samples were either stabilized for 2, 24, or 48 h at room temperature. Quantitative analysis of protein yield revealed a slightly reduced recovery of protein
Ergin et al. from samples that have been incubated in PAXgene stabilizer for 48 h when compared to 2 h of stabilization (Figure 3). However, the decrease in protein yield with a prolonged duration of stabilization was not statistically significant. Proteins Extracted from PAXgene-Fixed Tissue Are Nondegraded and Immunoreactive. To test the influence of the new fixative on protein integrity, electrophoretic mobility, and immunoreactivity, cryopreserved, PAXgene-fixed, and formalinfixed nonmalignant human specimens from spleen, breast, duodenum, and stomach tissues were analyzed in parallel. The protein profile was studied using Western blot with antibodies specific for membrane, cytoplasmic, and nuclear proteins. Clear bands of actin, E-cadherin, Hsp70, and ERK were detected for PAXgene-fixed tissues. No additional degraded protein bands appeared, and no differences in protein size were observed when compared to frozen samples (Figure 4A). The protein bands obtained for FFPE tissues were mainly less intense than those from PAXgene-fixed and frozen samples. When we performed reverse-phase protein microarray (RPPA), clear spots for PAXgene-fixed protein lysates detected by ERK antibody staining were visualized (Figure 4B), similar to the cryopreserved sample and FFPE. Using this approach, FFPE samples may even show stronger spot intensities than PFPE and cryopreserved samples. These data show that full-length, nondegraded, and immunoreactive proteins can be extracted from PAXgene tissues. The recovered proteins have the same properties when compared to cryopreserved samples by Western blotting and RPPA experiments. In addition, an important protein modification is preserved with the new fixative because phosphorylated ERK could be detected (Figure 4A). PAXgene-Fixed and Cryopreserved Tissues Revealed Similar Proteomic Signatures for MALDI IMS. Pancreatic tissue can be used as a model system for assessing the compatibility of different preservation methods with mass spectrometry (MS), namely, MALDI imaging mass spectrometry (MALDI IMS), because the peptide hormones located in the endocrine pancreas (Islets of Langerhans) can be measured. Therefore, pancreatic tissue preserved either by formalin, PAXgene, or cryopreservation was analyzed with MALDI IMS. For FFPE pancreatic tissue, no peptide peaks could be observed in the exocrine part of the formalin-fixed tissue, and in the
Figure 4. Western blot (A) and RPPA (B) analysis of PAXgene-fixed samples. PAXgene-fixed samples are comparable to cryopreserved samples in Western blotting and RPPA experiments. (A) Nonmalignant human specimens of spleen, breast, duodenum, and stomach tissues were equally divided into four samples and either snap-frozen, PAXgene-fixed (3 or 24 h), and paraffin-embedded or formalinfixed and paraffin-embedded. Protein extracts were separated by SDS-PAGE, and Western blot analysis was performed using indicated antibodies. (B) Human spleen tissue lysates were spotted in serial dilution onto the nitrocellulose membrane and immunodetected by ERK antibody. Sypro Ruby staining detected the total protein amount. 5192
Journal of Proteome Research • Vol. 9, No. 10, 2010
Proteomic Analysis of PAXgene-Fixed Tissue
research articles
Figure 5. MALDI imaging mass spectrometry. The display of the average m/z spectra of differently fixed pancreatic tissues. In contrast to formalin-fixed tissue, PFPE and cryopreserved pancreatic samples display a multitude of peaks. Average m/z spectra (2500-25 000 m/z) for exocrine pancreas in blue and for endocrine pancreas in yellow are shown. Marked peaks (asterisk, circle, diamond, or square) are enlarged.
Figure 6. Visualization of Insulin (m/z 3485) and Glucagon (m/z 5808) expression in pancreatic tissue using MALDI imaging. Image tiles show the visualization of Insulin (top left, in red), of Glucagon (top right, in green), and of Insulin and Glucagon (middle left) expression in the tissue section of (A) PAXgene-fixed and (B) cryopreserved tissue and the respective hematoxylin and eosin staining (middle right). The endocrine pancreas regions are circled. Spectra show the differential expression of Insulin (left) and Glucagon (right) comparing exocrine pancreas (blue) and endocrine pancreas (yellow).
endocrine part, two peaks with low intensity and m/z sharpness were detected (Figure 5). In contrast, the spectra of PAXgene and cryopreserved tissues were comprised of a multitude of peaks (n > 100). Furthermore, the PAXgene-
fixed and the cryopreserved tissue sections yielded similar spectra (Figure 5). We then focused on two known peptides, Insulin and Glucagon, which can be identified and visualized in pancreas Journal of Proteome Research • Vol. 9, No. 10, 2010 5193
research articles sections. These peptides are localized in the endocrine pancreas, and the correct localization can be judged by comparing to H&E stained slices. Insulin (m/z 3485) and Glucagon (m/z 5808) can only be detected in the endocrine pancreas of both PAXgene- and cryofixed tissues and not in the formalin-fixed tissue. Revisualization of the respective masses showed a colocalization of the respective m/z species with the endocrine pancreas (Figure 6A and B, the upper part). Comparing the average spectra from the endocrine and exocrine pancreas shows differential expression of Insulin and Glucagon (Figure 6A and B lower part). Signal intensity in PAXgene-fixed tissues was in average 15% lower than in cryopreserved tissues. Nucleic Acids Are Preserved in PAXgene-Fixed Tissue. To complete the use of PAXgene for multimodal tissue analysis in modern pathology, PFPE tissue was used for PCR-based assays. RNA was extracted from human liver samples that had been preserved in parallel with formalin, PAXgene fixative, or fresh frozen in liquid nitrogen, and RNA concentration was measured by Nanodrop. A PCR amplification assay using different amplicon lengths was performed to test for RNA fragmentation in fixed samples. All tissue preservation methods allowed for the PCR amplification of shorter fragments, but longer amplicons could only be amplified from PAXgene and cryopreserved tissue samples (Figure 7A). These data demonstrate better RNA preservation in PAXgene-fixed when compared to formalinfixed tissue. In addition, the integrity of extracted RNA was determined on the Agilent Bioanalyzer 2100, and total RNA was separated on a 1% agarose gel. RNA extracted from PFPE tissue showed clearly visible but lower rRNA peaks and a slight increase of shorter fragments compared to cryopreserved tissue (RIN 8.50), resulting in a calculated RIN value of 5.00 and 5.30, respectively (Supporting Information Figure 2). On a 1% agarose gel, separated total RNAs showed two bands at approximately 4.5 and 1.9 kb for Pax 3 h, 24 h, and the snap frozen sample (Figure 7 B). In comparison, RNA from formalin-fixed tissue revealed no 18S and 28S peaks on the electropherograms (RIN 2.30) and showed strongly degraded RNA on the agarose gel. Within our EU consortium which aims to tackle the standardization and improvement of preanalytical procedures for in vitro diagnostics (www.spidia.eu), over 30 routinely used diagnostic antibodies were tested on PFPE tissue, and staining patterns in PFPE sections were compared to the patterns in adjacent FFPE sections. Our preliminary results are very promising (data not shown); however, a detailed immunohistochemical analysis will be reported elsewhere (Kap et al., in preperation).
Discussion Over decades, formalin has been the fixative of choice in surgical pathology departments, as it preserves the cellular and morphological details and enables long-term storage of tissues at room temperature. However, molecular analyses using FFPE tissues are limited due to the chemical modifications that formaldehyde introduces into biomolecules.15,16 Nucleic acids are fragmented, and protein-protein cross-linking interferes with efficient protein extraction procedures. In recent years, downstream methods have been improved, and attempts to make FFPE material accessible for molecular analyses and in particular proteomic analyses have been successful.3,5,17 Using fresh-frozen tissue samples, the discovery of disease-related RNA profiles which could be translated into diagnostics like MammaPrint18,19 and the detection of specific DNA mutations 5194
Journal of Proteome Research • Vol. 9, No. 10, 2010
Ergin et al.
Figure 7. PAXgene stabilizes RNA. (A) Electrophoresis of PCR products on a 2% TAE agarose gel showing amplicons with 153, 323, and 530 bp in length for human GAPDH amplified from human nonmalignant liver tissue either formalin-fixed and paraffin-embedded (FF 24 h), PAXgene-fixed and paraffin-embedded (Pax 3 and 24 h), or cryopreserved. (B) Electrophoresis of total RNA isolated from nonmalignant human liver samples (Formalinfixed, PAXgene-fixed, and snap frozen) on a 1% agarose gel showing two bands at approximately 4.5 and 1.9 kb for Pax 3 h, 24 h, and the snap frozen sample (Cryo). These bands represent 28S and 18S rRNA indicating intact RNA. For the formalin-fixed sample (FFPE), these bands are missing.
that have been linked to cancer types20,21 demonstrate the success of molecular approaches in modern pathology. However, there is a strong need for alternative fixatives that enable reliable molecular analyses and preserve the morphological and cellular structure of fixed samples to enable multimodal approaches from one surgical specimen. PAXgene, a novel formalin-free fixative, is now on the market and combines all properties of an ideal fixative. PAXgene fixation allows storage of tissue blocks in archives like FFPE material, and no costintensive logistics, liquid nitrogen tanks, etc. are needed. Thus, the samples can be collected at one place, transferred into a stabilizer, and then be safely transported to the analyzing laboratory. The great advantage of PAXgene fixation compared to formalin fixation and frozen material is that it allows IHC, morphology analysis, and detection of biomolecules from the same sample. In this study, we focused on the protein analysis of PAXgenefixed tissues and presented data that PFPE tissues can be used for proteomic studies, such as Western blotting, RPPA, and MALDI IMS. In the first step, the efficient recovery of protein from PAXgene-fixed tissue was tested using three different
research articles
Proteomic Analysis of PAXgene-Fixed Tissue protocols. The extraction protocols varied mainly in the incubation temperature of the lysates. We demonstrated the recovery of protein from PFPE tissue when using the standard cryopreserved protocol (protocol I), including a 30 min incubation step on ice (Figure 2A). To extract protein from FFPE samples, a high temperature is needed to reverse cross-linking.1,4 Consequently, protein recovery from FFPE tissue was inefficient using protocol I, which clearly showed that protein can be easily extracted from PAXgene-fixed tissue. The quantity of protein recovery could be increased by heating the samples for 2 h at 70 °C with protein amounts higher than those from FFPE tissue. In addition, our findings show that the extracted protein from PFPE tissue separated by SDS-PAGE displayed the same protein pattern as protein extracted from cryopreserved samples (Figure 2B). Finally, extraction protocol II including a 2 h incubation at 70 °C is recommended for the efficient extraction of full-length proteins from PFPE tissue samples. As the PAXgene system consists of two solutions (fixative and stabilizer), the possible effect of PAXgene stabilizer solution on the quantity of recovered protein was investigated. We found that tissue samples could be stabilized for several hours without severe negative effects on protein recovery (Figure 3) but with great advantages for tissue handling and transportation. We successfully tested five different antibodies against membrane, nuclear, and cytoplasmic proteins by Western blot analysis (Figure 4A). PFPE and cryopreserved materials were comparable regarding the intensity, pattern, and resolution of protein in all tested tissues. Although equal amounts of protein were loaded, the signal intensity of FFPE tissue was lower for some tissues and antibodies, which suggests a lower overall protein recovery of full-length proteins from FFPE material. However, with RPPA approaches we detected a higher intensity for FFPE material than for cryopreserved, which is a common observation for several antibodies and tissue types in our lab (data not shown). The reason for this result remains unclear. Nevertheless, clear spots and a reliable dilution series can be detected for PAXgene tissue, indistinguishable from cryopreserved tissue (Figure 4B). We demonstrated here for the first time the recovery of nondegraded and immunoreactive proteins from PAXgene-fixed, paraffin-embedded tissue suitable for Western blotting and reverse-phase protein microarray. These findings display central qualifications for using this fixative in clinical settings and protein biomarker discovery. Thus, we present in this study the first data for MALDI imaging mass spectrometry on PFPE material. A multitude of peaks could be detected in PFPE tissue sections, and the peak pattern was comparable to the one from cryopreserved-fixed material (Figure 5). Two proteins in the pancreas, Insulin and Glucagon, were visualized in endocrine cells. The expression matched well with the histological staining of the Islets of Langerhans (Figure 6). Because MALDI imaging MS can be performed successfully using PFPE tissues, this fixation method may also be suitable for other mass spectrometric methods. So far, there are two publications demonstrating MALDI imaging MS on alternative fixed tissues. The first article describes the feasibility of MALDI-IMS using ethanol-fixed lung tumors.22 The second article tests whether RCL2/CS100 fixation can be used as an alternative fixative instead of formalin23 and showed a high level of agreement between fresh-frozen and RCL2/CS100-fixed tissue specimens. Using the new alcoholbased and formalin-free fixative PAXgene, we also observed similarities between cryopreserved and PFPE samples regarding the detected peaks and their local distribution. PAXgene fixation
also resulted in lowered signal intensity (∼15%) compared with cryopreserved fixation. The reason for this result remains unclear. However, the reduction of signal intensity is less pronounced than in the case of the RCL2/CS100 fixative. Due to innovative techniques, the discovery of diseaserelated biomolecules has revolutionized cancer diagnosis and therapies in recent years, which corresponds with the need for alternative fixatives enabling more precise multimodal analyses from the same surgical specimens. Here we report that PAXgene, a novel formalin substitute, can be used for standard proteomic analyses, e.g., Western blotting and RPPA, but can also be used for innovative in situ proteomic investigations including MALDI IMS. The properties of this fixative, with regard to analysis of nucleic acids and preservation of morphology, will be reported in much more detail by C. Viertler et al. and M. Kap et al. (in preparation), respectively. The PAXgene fixative enables the successful application of common molecular techniques to analyze clinical specimens regarding nucleic acids, proteins, and histology from the same sample, thereby gaining novel insights into cancer diagnosis, progression, and therapies. Further analysis of additional proteomic investigations, including two-dimensional gel electrophoresis, mass spectrometry, and ELISA (enzyme-linked immunosorbent assay), will be performed in the near future along with the utilization of longtime-stored tissue blocks and inactivation of pathogens to establish the benefit of PAXgene-fixed tissues.
Acknowledgment. The authors have declared no conflict of interest. The authors thank Claudia-Mareike Pflu ¨ ger for excellent technical assistance. The research leading to these results has received funding from the European Union Seventh Framework Programme [FP7/ 2007-2013] under grant agreement n° 222916 and from the Federal Ministry of Education and Research (BMBF) Germany (Grant 0101-31P5873 and Grant 01GR0805). We thank all the SPIDIA partners for their support and discussions. Supporting Information Available: Figure showing the calculated protein yield of human kidney, spleen, breast, duodenum; and stomach tissue samples and figure with the electropherograms of nonmalignant human liver tissue samples that were either snap-frozen, PAXgene-fixed, or formalin-fixed. This material is available free of charge via the Internet at http://pubs.acs.org. References (1) Shi, S. R.; Key, M. E.; Kalra, K. L. Antigen retrieval in formalinfixed, paraffin-embedded tissues: an enhancement method for immunohistochemical staining based on microwave oven heating of tissue sections. J. Histochem. Cytochem. 1991, 39 (6), 741–8. (2) Lehmann, U.; Kreipe, H. Real-time PCR analysis of DNA and RNA extracted from formalin-fixed and paraffin-embedded biopsies. Methods 2001, 25 (4), 409–18. (3) Becker, K. F.; Schott, C.; Hipp, S.; Metzger, V.; Porschewski, P.; Beck, R.; Nahrig, J.; Becker, I.; Hofler, H. Quantitative protein analysis from formalin-fixed tissues: implications for translational clinical research and nanoscale molecular diagnosis. J. Pathol. 2007, 211 (3), 370–8. (4) Ikeda, K.; Monden, T.; Kanoh, T.; Tsujie, M.; Izawa, H.; Haba, A.; Ohnishi, T.; Sekimoto, M.; Tomita, N.; Shiozaki, H.; Monden, M. Extraction and analysis of diagnostically useful proteins from formalin-fixed, paraffin-embedded tissue sections. J. Histochem. Cytochem. 1998, 46 (3), 397–403. (5) Grantzdorffer, I.; Yumlu, S.; Gioeva, Z.; von Wasielewski, R.; Ebert, M. P.; Rocken, C. Comparison of different tissue sampling methods for protein extraction from formalin-fixed and paraffin-embedded tissue specimens. Exp. Mol. Pathol. 2010, 88 (1), 190–6.
Journal of Proteome Research • Vol. 9, No. 10, 2010 5195
research articles (6) Berg, D.; Hipp, S.; Malinowsky, K.; Bo¨llner, C.; Becker, K. F., Molecular profiling of signalling pathways in formalin-fixed and paraffin-embedded cancer tissues. Eur. J. Cancer 20104647. 55. (7) Nirmalan, N. J.; Harnden, P.; Selby, P. J.; Banks, R. E. Development and validation of a novel protein extraction methodology for quantitation of protein expression in formalin-fixed paraffinembedded tissues using western blotting. J. Pathol. 2009, 217 (4), 497–506. (8) Walch, A.; Rauser, S.; Deininger, S. O.; Hofler, H. MALDI imaging mass spectrometry for direct tissue analysis: a new frontier for molecular histology. Histochem. Cell Biol. 2008, 130 (3), 421–34. (9) Cornett, D. S.; Reyzer, M. L.; Chaurand, P.; Caprioli, R. M. MALDI imaging mass spectrometry: molecular snapshots of biochemical systems. Nat. Methods 2007, 4 (10), 828–33. (10) Lemaire, R.; Desmons, A.; Tabet, J. C.; Day, R.; Salzet, M.; Fournier, I. Direct analysis and MALDI imaging of formalin-fixed, paraffinembedded tissue sections. J. Proteome Res. 2007, 6 (4), 1295–305. (11) Delfour, C.; Roger, P.; Bret, C.; Berthe, M. L.; Rochaix, P.; Kalfa, N.; Raynaud, P.; Bibeau, F.; Maudelonde, T.; Boulle, N. RCL2, a new fixative, preserves morphology and nucleic acid integrity in paraffin-embedded breast carcinoma and microdissected breast tumor cells. J. Mol. Diagn. 2006, 8 (2), 157–69. (12) Stanta, G.; Mucelli, S. P.; Petrera, F.; Bonin, S.; Bussolati, G. A novel fixative improves opportunities of nucleic acids and proteomic analysis in human archive’s tissues. Diagn. Mol. Pathol. 2006, 15 (2), 115–23. (13) Vincek, V.; Nassiri, M.; Nadji, M.; Morales, A. R. A tissue fixative that protects macromolecules (DNA, RNA, and protein) and histomorphology in clinical samples. Lab. Invest. 2003, 83 (10), 1427–35. (14) Rauser, S.; Marquardt, C.; Balluff, B.; Deininger, S. O.; Albers, C.; Belau, E.; Hartmer, R.; Suckau, D.; Specht, K.; Ebert, M. P.; Schmitt, M.; Aubele, M.; Hofler, H.; Walch, A. Classification of HER2 Receptor Status in Breast Cancer Tissues by MALDI Imaging Mass Spectrometry. J. Proteome Res. 2010, 9, 1854–1863.
5196
Journal of Proteome Research • Vol. 9, No. 10, 2010
Ergin et al. (15) Puchtler, H.; Meloan, S. N. On the chemistry of formaldehyde fixation and its effects on immunohistochemical reactions. Histochemistry 1985, 82 (3), 201–4. (16) Nirmalan, N. J.; Harnden, P.; Selby, P. J.; Banks, R. E. Mining the archival formalin-fixed paraffin-embedded tissue proteome: opportunities and challenges. Mol. Biosyst. 2008, 4 (7), 712–20. (17) Addis, M. F.; Tanca, A.; Pagnozzi, D.; Rocca, S.; Uzzau, S. 2-D PAGE and MS analysis of proteins from formalin-fixed, paraffin-embedded tissues. Proteomics 2009, 9 (18), 4329–39. (18) van ’t Veer, L. J.; Dai, H.; van de Vijver, M. J.; He, Y. D.; Hart, A. A.; Mao, M.; Peterse, H. L.; van der Kooy, K.; Marton, M. J.; Witteveen, A. T.; Schreiber, G. J.; Kerkhoven, R. M.; Roberts, C.; Linsley, P. S.; Bernards, R.; Friend, S. H. Gene expression profiling predicts clinical outcome of breast cancer. Nature 2002, 415 (6871), 530– 6. (19) Glas, A. M.; Floore, A.; Delahaye, L. J.; Witteveen, A. T.; Pover, R. C.; Bakx, N.; Lahti-Domenici, J. S.; Bruinsma, T. J.; Warmoes, M. O.; Bernards, R.; Wessels, L. F.; Van’t Veer, L. J. Converting a breast cancer microarray signature into a high-throughput diagnostic test. BMC Genomics 2006, 7, 278. (20) Bardelli, A.; Velculescu, V. E. Mutational analysis of gene families in human cancer. Curr. Opin. Genet. Dev. 2005, 15 (1), 5–12. (21) Futreal, P. A.; Coin, L.; Marshall, M.; Down, T.; Hubbard, T.; Wooster, R.; Rahman, N.; Stratton, M. R. A census of human cancer genes. Nat. Rev. Cancer 2004, 4 (3), 177–83. (22) Chaurand, P.; Latham, J. C.; Lane, K. B.; Mobley, J. A.; Polosukhin, V. V.; Wirth, P. S.; Nanney, L. B.; Caprioli, R. M. Imaging mass spectrometry of intact proteins from alcohol-preserved tissue specimens: bypassing formalin fixation. J. Proteome Res. 2008, 7 (8), 3543–55. (23) Mange, A.; Chaurand, P.; Perrochia, H.; Roger, P.; Caprioli, R. M.; Solassol, J. Liquid chromatography-tandem and MALDI imaging mass spectrometry analyses of RCL2/CS100-fixed, paraffin-embedded tissues: proteomics evaluation of an alternate fixative for biomarker discovery. J. Proteome Res. 2009, 8 (12), 5619–28.
PR100664E
