.JPG?ad=18481&adview=true)
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Molecular Oncology, Markers, Clinical Correlates |
1 Departments of Biochemistry, 2 Neurosurgery, and 3 Pathology, and the 4 Mass Spectrometry Research Center, Vanderbilt University School of Medicine, Nashville, Tennessee; 5 Brain Tumor Institute, Cleveland Clinic Foundation, Cleveland, Ohio; and 6 Surgical Neurology Branch, National Institutes of Neurological Disorders and Stroke, NIH, Bethesda, Maryland
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Experimental Design: We examined 20 prospectively collected, snap-frozen normal brain and brain tumor specimens using matrix-assisted laser desorption/ionization (MALDI) mass spectrometry (MS), and compared peptide and protein expression in primary brain tumor and nontumor brain tissues.
Results: MS can be used to identify protein expression patterns in human brain tissue and tumor specimens. The mass spectral patterns can reliably identify glial neoplasms of similar histological grade and differentiate them from tumors of different histological grades as well as from nontumor brain tissues. Initial bioinformatics cluster analysis algorithms classified tumor and nontumor tissues into similar groups comparable with their histological grade.
Conclusions: We describe a novel tool for the analysis of protein expression patterns in human glial neoplasms. Initial results demonstrate that MALDI-MS technology can significantly aid in the process of unraveling and understanding the molecular complexities of gliomas. MALDI-MS accurately and reliably identified normal and neoplastic tissues, and could be used to discriminate between tumors of increasing grades.
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Protein profiling mass spectrometry (MS) is a new technology that takes advantage of the methodology and instrumentation of matrix-assisted laser desorption/ionization (MALDI) MS (10, 11, 12, 13, 14 , 15, 16, 17 ). MALDI-MS can be used to detect the molecular weight of peptides and proteins over Mr 100,000 directly from a variety of in vitro and in vivo samples including protein solutions, fresh-frozen tissue sections, microdissected fresh tissues, or from cells derived in culture (10, 11, 12, 13, 14, 15, 16, 17) . In this method, a sample is deposited or mounted on a metal plate, U-desorbing matrix is deposited on the sample, and each matrix droplet is analyzed. A UV laser is fired for a predetermined number of times over each matrix droplet surface, desorbing and ionizing the protein/peptide analytes from the sample. These charged molecules are accelerated down a flight tube and detected, and the m/z ratio of each molecule is determined (10, 11, 12, 13, 14, 15, 16, 17) .
Data obtained from such an analysis consists of a mass spectrum in which the peaks observed in the MALDI mass spectrum correspond to a peptide or protein from the sample (10, 11, 12, 13, 14, 15, 16, 17) . The molecule signals are separated along the X axis according to their m/z ratio. Because, in most cases, MALDI-MS results in a singly charged molecule, the m/z value detected is the molecular weight of the protein. In the case of direct tissue analysis, multiple matrix droplets are deposited on the tissue surface to account for cellular heterogeneity, and the data are collected as a series of mass spectra. Each spectrum contains signals from many hundreds of proteins and peptides specific to that tissue region.
We used MALDI-MS to analyze prospectively a series of 20 human nontumor brain and glial tumor specimens to assess the ability of the technique to catalogue protein expression in these samples and to map these protein profiles. The results demonstrate differential protein expression between nontumor brain and glioma tissues, as well as between low- and high-grade gliomas.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). In no instance did the blinded review contradict the original diagnosis.
|
Tissue Sample Preparation.
Frozen tissues were sectioned (12 µm thick) at -15°C on a Leica cryostat (Leica Microsystems Inc., Bannockburn, IL). The sections were transferred and thawed onto gold-coated stainless-steel MALDI target plates. Consecutive sections (6 µm) were cut, picked up on glass slides, and stained with H&E for pathological classification. For MS profiling, matrix droplets (0.1 µl of saturated sinapinic acid in 50:50 acetonitrile:water containing 0.1% trifluoroacetic acid, v/v) were deposited on the surface of the tissue sample, dried, and additional 0.1 µl droplets were deposited on top of each initial spot. Typically, several droplets (4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20)
are deposited on each tissue section, depending on the section size and morphology, allowing for multiple samples to be taken across the entire tissue surface. In the case of severe morphological changes, i.e., tissue sections containing both tumor and healthy tissue morphology, each region is spotted with multiple droplets to account for tissue heterogeneity during profiling. Sections were dried at least 1 h before MS analysis. Optical images were taken of the spotted sections to align regions of MS analysis with cellular morphology as determined by histopathology.
MALDI MS.
Each matrix droplet on a tissue section was analyzed individually using a MALDI time of flight Voyager DE-STR mass spectrometer (Applied Biosystems, Foster City, CA), equipped with a nitrogen laser (337 nm), used in the linear mode under delayed extraction conditions, as described previously (10, 11, 12, 13, 14, 15, 16, 17)
. The laser spot was approximately circular with a diameter of 
25 µm. Data were collected using an accelerating voltage of 25 kV, 91% grid voltage, 0.05% guide wire voltage, and a delay time of 150 ns. A total of 750 laser shots were averaged for each matrix droplet; spectra were collected for a mass range of 2–50 kDa.
Data Processing.
Spectra were internally calibrated using the singly and doubly charged 
-hemoglobin chain (molecular weight Mr 15,125.8 and 7,563.2, respectively) and thymosin ß4 (Mr 4,964), which we have identified previously in human glioblastoma xenographs. Spectra were baseline corrected and normalized. Data collected from multiple sites within the same pathological region of each biopsy were averaged. The resulting spectrum was deconvoluted and the signals detected between mass/charge 2000 and 21,000 were used for the cluster analysis (see below).
Automated peak detection was performed using DataExplorer software (Applied Biosystems, Framingham, MA), with a signal-to-noise threshold of 3. Using these criteria, between 200 and 400 peaks were detected within a spectrum from 2,000 to 21,000 m/z.
Cluster Analysis.
Unsupervised hierarchical cluster analysis was performed on averaged spectra from 20 biopsies (5 normal, 3 grade II, 3 grade III, 4 grade IV, 1 adrenal pheochromocytoma, 1 embryonal tumor, and 1 dysembryoplastic neuroepithelial tumor) using the Cogid Biomarker Discovery Software (BioAnalyte, Portland, ME). Because MALDI time of flight analysis results in some mass assignment variation (±1–3 Da across the 2000–21000 mass range), a software resolution parameter (set at 700) was used to determine the degree of mass overlap allowed when identifying common peaks between individual spectra. Peak lists input into the analysis program were clustered according to similarity, using a presence/absence criterion. The relationships between samples are demonstrated by a hierarchical tree diagram; for example see Fig. 5
. Tree diagrams are typically defined by connecting branches (lines) linked by nodes (n) where the branches define how the individual samples are related (according, in this case, to their similarity in protein patterns), and the nodes demonstrate which branches are linked and where they intersect. The individual fundamental nodes, reflecting the peak list from each biopsy, are displayed along the Y axis. Peak lists are then related to each other by the higher order nodes constructed across the X axis. Higher order nodes are joined by degree of similarity. The most similar are joined first, with the position of the node illustrating similarity between the respective branches. The degree of similarity measurement calculated for each node is determined from the total number of peaks used in the clustering, not the total number of peaks across all of the samples. Individual ions present in all of the samples were not used in the clustering analysis.
|
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). There was a total of 5 samples of nontumor lateral temporal neocortex, obtained from patients undergoing anterior temporal lobectomy and amygdalohippo-campectomy for histologically confirmed mesial temporal sclerosis, 14 primary brain tumors, and 1 tumor of neural crest origin. Primary brain tumors included those from 13 patients (total of 18 patients for all of the samples), with the following tumors: 3 patients with low-grade (WHO grade II) tumors (1 astrocytoma and 2 oligodendrogliomas); 3 patients with higher-grade (WHO grade III) tumors (1 anaplastic astrocytoma and 3 anaplastic oligoastrocytomas; 1 of these latter patients had a predominantly grade III tumor that had arisen of a grade II oligoastrocytoma, portions of which remained on histological sampling); and 4 patients with glioblastoma multiforme (WHO grade IV). One high-grade (grade III-IV) embryonic tumor in a newborn, 1 adrenal pheochromocytoma, and 1 dysembryoplastic neuroepithelial tumor were also studied (see Table 1
).
|
and 3
). Over the m/z range from 2000–21,000, >200–400 individual signals can be distinguished; each signal is reflective of an individual molecule (primarily a peptide or protein) within the tissue sample. Similarly, each protein profile pattern is distinctive for different cellular morphologies, as shown in Figs. 2
and 3
; representative portions of the spectra are shown. Data collected from a homogeneous anaplastic astrocytoma (Fig. 2)
demonstrate the similarity in spectral patterns, including the specific molecules monitored as well as the relative peak intensities, when comparable cellular compositions from a single biopsy are analyzed. Consequently, analysis of a heterogeneous biopsy (Fig. 3)
, which maintained two distinct morphological regions, white matter and cortex (with infiltrating tumor cells), demonstrates the protein profile similarity within the two white matter regions as well as the pattern distinctions between the white matter and cortex regions. Thus, MALDI-MS is an effective method for the analysis of multiple regions of a tumor, whether from the same region of the neoplasm or from the opposite pole.
|
|
that MALDI-MS can be used to discriminate between different histological grades of gliomas. Low-grade (WHO grade II) glial neoplasms (astrocytoma and oligodendroglioma) and anaplastic astrocytomas (WHO grade III) can be accurately and reproducibly distinguished from glioblastoma multiforme (WHO grade IV) by the presence or absence of peaks over a broad spectrum of analyzable protein signals. Initial bioinformatic investigation, using an unsupervised, hierarchical, cluster analysis, shows that MALDI-MS can be used to distinguish subgroups of tumor tissue from each other, consistent with the WHO classification, as well as from nontumor tissue (Fig. 5)
. Cluster analysis is a simple approach that selects a variable and then divides the matrix set into those objects that do or do not possess that variable (in this case, a selected peak, of which the identity need not be known in advance). This type of analysis results in a dendrogram (Fig. 5)
that defines the signal patterns shared by individual spectra, which permits cluster analysis.
|
) were processed and the resulting peak lists from these data were generated. Each spectrum consisted of between 200 and 400 individual protein signals. In combining these lists, a total of 647 individual protein signals across all of the patient samples were used in the clustering analysis. Analysis of these results demonstrates that each histological subgroup clusters in its own, unique region. Although there are discernable differences in the spectra of grade II and grade III gliomas, initial bioinformatic clustering differentiates most strongly between normal, grade II and III tumors, and glioblastoma multiforme. Tumor samples 10–1 and 10–2 came from opposite poles of a single glioblastoma, 5 cm apart. They were analyzed separately, in a blinded fashion, and not only do both segregate with the other glioblastomas, but more importantly, the closest match of the two samples is one another. This provides additional confirmation of the robust nature of this analysis.
Additional inspection of the cluster analysis (Fig. 5)
identifies two interesting tumors. The first, patient #6, represents a grade II oligoastrocytoma that had recurred after radiation, predominantly as a grade III anaplastic oligoastrocytoma. This tumor still preserved, on histological inspection, discrete regions comprised of grade II tumor cells only, as well as the more numerous areas of grade III tumor, reflected by the presence of spectrum characteristics sharing both grade II tumor and grade III tumor signals (patient biopsy #6–1 and 6–2 on Fig. 5
), when these two tumor regions were analyzed separately. The second tumor (patient #16), represents the one outlier in the grade II/grade III group. This tumor, a cerebral embryonal carcinoma, was detected antenatally a few days before birth and required surgery 3 days postnatally. The entire visible tumor on contrast-enhanced imaging was resected, and the infant received postoperative chemotherapy. Histologically, this tumor possessed aggressive, malignant features and was graded WHO class III-IV (with extensive hypercellularity, nuclear atypia, vascular proliferation, and necrosis, with MIB-1 labeling >10% of cells) by both the original and the blinded study neuropathologist. The patient remains free of disease at 3 years’ time, suggesting that its pathological appearance belied its clinical behavior, a feature suggested by the cluster analysis, which grouped it with the lower-grade tumors. Whereas the presence of pleomorphic, hypercellular features and nuclear atypia in embryonal carcinoma may not portend aggressive behavior, as is the case with pleomorphic xanthoastrocytomas, an alternative explanation, which the MS data would support, may be that the tumor was differentiating into a less aggressive neoplasm. In addition, patient #18, an adrenal pheochromocytoma, which is of neural crest origin, shared essentially no identity with any tumor or with normal brain tissues. This additionally suggests that MALDI-MS may be a robust method to identify tumors of unknown origin or to distinguish between two tumors that share histological features but have different origins, such as neuroblastoma and Ewing’s sarcoma (2)
.
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Early efforts in proteomics date back several decades to the development of two-dimensional gel electrophoresis and the cataloging of individual gel spots to create nascent protein databases (6) . With the explosion of knowledge in the genomic era, new methods have been developed to identify protein targets. Numerous areas have been explored including protein identification, recognition of post-translational modifications, assessment of protein function, and characterization of protein-protein interactions. Many of these methods rely on having rather large quantities of pure protein sample for accurate quantification, and are less useful for rapid screening and quantification of tissues and tumors (6) .
A recent breakthrough in proteomics in clinical samples has been the development of direct analysis of tissues by MS (10, 11, 12, 13, 14, 15, 16, 17) . We have adapted a straightforward mass spectrometric method, MALDI-MS, to the problem of protein and peptide detection from specific cell types in heterogeneous tissue and tumor samples (10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22) . This technique has been used prospectively to analyze a series of human gliomas and normal brain tissues, as well as a tumor of neural crest origin, an adrenal pheochromocytoma. Our results demonstrate that the protein profiles of nontumor brain (temporal lobe), grade II, grade III, and grade IV glial tumors display unique spectra. Simple bioinformatic clustering demonstrates that the data from these spectra can be used to distinguish normal from neoplastic tumor tissues. Even in this small sample of brain tissues and tumors, this method demonstrated a robust ability to discriminate accurately between tumor and nontumor tissue, and to predict reliably the tumor grade in exact conformity with a blinded neuropathologist. Whereas recent data suggest that oligodendrogliomas may have certain genetic alterations, such as loss of portions of chromosomes 1p and 19q, that may be involved in their genesis and do predict sensitivity to certain chemotherapeutic drugs, the sample size of this study was too small to differentiate between histological subtypes of the gliomas; work on this issue is currently in progress (2 , 18) .
Separate peak analysis can be performed using a second round of MS to identify tumor-specific proteins. Our laboratory has recently used this approach to identify two proteins, thymosin ß4 and S100 calcium-binding protein A4, from explants of cultured human glioblastomas (16) . In parallel, by a more laborious method using gene expression analyses followed by screening and in vitro expression of the potential genes, Clark et al. (23) have also shown thymosin ß4 to play an important role in the metastatic potential of melanomas. Methods using tandem MS/MS methods and secondary protein identification through identification in protein databases will prove increasingly important for the identification of not only tumor protein profiles or "signatures," but of also new, individual protein markers and, in time, specific tumor targets (4 , 5 , 7 , 8 , 16 , 19, 20, 21, 22) .
MALDI-MS is a powerful tool for the systematic detection of differential protein profiles of normal and neoplastic tissues (10, 11, 12, 13, 14, 15, 16, 17) . Supplemental modifications of the MALDI MS technique and additional protein identification methods add additional tools to help unravel the proteomic complexity of tumorigenesis. In addition, MALDI-MS can supplement, rather than supplant, other standard or more novel molecular diagnostic techniques, such as gene expression profiling, Northern or Western blotting, and qualitative or quantitative PCR techniques. MALDI technology can also be used to identify nucleic acids or proteins in blood, serum, or other fluids and to characterize single nucleotide polymorphisms in fluids, tissues, or tumors (12 , 24) .
The work reported here represents the first stage in the proteomic analysis of gliomas, which is the diagnostic and prognostic value of the protein patterns obtained directly from the patient biopsies. In addition to increasing the number of patients to be studied, the next phase will focus on identifying the proteins most differentially expressed among disease states. Work analyzing pure tumor populations of microdissected infiltrative tumor cells in normal white matter is also in progress (21) . Bioinformatic analysis of proteomic spectra from large numbers of tumors, performed using computational tools adapted from gene profiling, will enhance the utility of direct proteomic tissue profiling. Additional studies to identify and characterize the suspected peptide/protein biomarkers and to characterize their function and biological role are under way. Identification of novel proteins involved in gliomagenesis may permit their use as diagnostic and prognostic markers as well as therapeutic targets.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Notes: S. A. S. and R. J. W. contributed equally to this work. Supplemental data relating to this article may be found at www.aacr.org.
Requests for reprints: Robert J. Weil, Unit on Tumor Physiology and Surgical Therapeutics, Surgical Neurology Branch, Building 10, Room 5D37, Surgical Neurology Branch, National Institutes of Neurological Disorders and Stroke, NIH, 10 Center Drive/9000 Rockville Pike, Bethesda, MD 20892-1414. Phone: (301) 496-2921; Fax: (301) 402-0380; E-mail: weilr{at}ninds.nih.gov
Received 6/16/03; revised 10/ 7/03; accepted 10/12/03.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Goncalves, E. Charafe-Jauffret, F. Bertucci, S. Audebert, Y. Toiron, B. Esterni, F. Monville, C. Tarpin, J. Jacquemier, G. Houvenaeghel, et al. Protein Profiling of Human Breast Tumor Cells Identifies Novel Biomarkers Associated with Molecular Subtypes Mol. Cell. Proteomics, August 1, 2008; 7(8): 1420 - 1433. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Bouamrani, J. Ternier, D. Ratel, A.-L. Benabid, J.-P. Issartel, E. Brambilla, and F. Berger Direct-Tissue SELDI-TOF Mass Spectrometry Analysis: A New Application for Clinical Proteomics Clin. Chem., November 1, 2006; 52(11): 2103 - 2106. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. S. Cornett, J. A. Mobley, E. C. Dias, M. Andersson, C. L. Arteaga, M. E. Sanders, and R. M. Caprioli A Novel Histology-directed Strategy for MALDI-MS Tissue Profiling That Improves Throughput and Cellular Specificity in Human Breast Cancer Mol. Cell. Proteomics, October 1, 2006; 5(10): 1975 - 1983. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Ernst, C. Melle, B. Schimmel, A. Bleul, and F. von Eggeling Proteohistography--Direct Analysis of Tissue with High Sensitivity and High Spatial Resolution Using ProteinChip Technology J. Histochem. Cytochem., January 1, 2006; 54(1): 13 - 17. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. M. Jamshedur Rahman, Y. Shyr, P. B. Yildiz, A. L. Gonzalez, H. Li, X. Zhang, P. Chaurand, K. Yanagisawa, B. S. Slovis, R. F. Miller, et al. Proteomic Patterns of Preinvasive Bronchial Lesions Am. J. Respir. Crit. Care Med., December 15, 2005; 172(12): 1556 - 1562. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. A. Schwartz, R. J. Weil, R. C. Thompson, Y. Shyr, J. H. Moore, S. A. Toms, M. D. Johnson, and R. M. Caprioli Proteomic-Based Prognosis of Brain Tumor Patients Using Direct-Tissue Matrix-Assisted Laser Desorption Ionization Mass Spectrometry Cancer Res., September 1, 2005; 65(17): 7674 - 7681. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Shankar, N. Gude, F. Cullinane, S. Brennecke, A. W Purcell, and E. K Moses An emerging role for comprehensive proteome analysis in human pregnancy research Reproduction, June 1, 2005; 129(6): 685 - 696. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Chaurand, S. A. Schwartz, M. L. Reyzer, and R. M. Caprioli Imaging Mass Spectrometry: Principles and Potentials Toxicol Pathol, January 1, 2005; 33(1): 92 - 101. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Chaurand, M. E. Sanders, R. A. Jensen, and R. M. Caprioli Proteomics in Diagnostic Pathology: Profiling and Imaging Proteins Directly in Tissue Sections Am. J. Pathol., October 1, 2004; 165(4): 1057 - 1068. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. H. Geho, E. F. Petricoin, and L. A. Liotta Blasting into the Microworld of Tissue Proteomics: A New Window on Cancer: Commentary re S. A. Schwartz et al., Protein Profiling in Brain Tumors Using Mass Spectrometry: Feasibility of a New Technique for the Analysis of Protein Expression. Clin. Cancer Res., 10: 981-987, 2004. Clin. Cancer Res., February 1, 2004; 10(3): 825 - 827. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
