Discovery and Identification of Potential ... - ACS Publications

Discovery and Identification of Potential Biomarkers in a Prospective Study of Chronic Lymphoid Malignancies Using SELDI-TOF-MS Laurent Miguet,† Ralf Bogumil,‡ Philippe Decloquement,† Raoul Herbrecht,§ Noelle Potier,† Laurent Mauvieux,*,| and Alain Van Dorsselaer*,† Laboratory of Bio-Organic Mass Spectrometry, ECPM, UMR 7178, Hubert CURIEN Multidisciplinary Institute, Louis Pasteur University, Strasbourg, France, Ciphergen Biosystems GmbH, Berlin-Hennigsdorf, Germany, Department of Hematology and Oncology, Hautepierre Hospital, Strasbourg, France, and Institute of Hematology and Immunology, Medical School, Louis Pasteur University and Laboratory of Hematology, Hautepierre Hospital, Strasbourg, France Received February 22, 2006

The accurate diagnosis of the different forms of chronic mature B-cell lymphocytic malignancies is of primary importance to determine an appropriate and efficient treatment. Usually, the diagnosis is achieved by morphology and immunophenotyping. Nevertheless, the diagnostic tools available are not able to discriminate pathologies with variable evolution, or to classify some of them. To discover new biomarkers, we used peptide and protein profiling SELDI-TOF-MS, to analyze 39 chronic B-cell malignancies and 20 control serum samples. Markers of interest were subsequently identified and characterized. In the obtained SELDI-MS profiles, most of the differences were observed in three mass ranges (m/z ) 13 000; m/z ) 9000; m/z < 2000). Identification of these biomarkers was achieved either by direct enrichment on the ProteinChip arrays followed by on-chip-MS/MS or by chromatographic fractionation, 1D-gel followed by nanoLC-MS/MS analysis. An increase of a sulfite form of transthyretin (13 841 Da) was observed in the patient group. A second set of markers at 8.6 and 8.9 kDa was identified as complement related fragment proteins, the C3a and C4a anaphylatoxins. In the low mass range, several peptides originating from N-terminal and C-terminal processing of the C3 alpha and C4 alpha chains were specifically observed in 38% of the patient sera, but in none of the control sera. This study emphasizes the usefulness of mass spectrometry studies in such malignancies. Keywords: SELDI • biomarker • chronic B-cell lymphoid malignancy • CLL • mass spectrometry • proteomic

Introduction Chronic mature B-cell lymphoid malignancies are frequent in adults. They represent 4% of the cancers in western countries where the augmentation is 5 to 10% per year.1 The diagnosis of these different neoplasms is based on morphological and immunophenotypical criteria, using for instance the Matutes Score,2 but also on cytogenetic or molecular biology studies. The World Health Organization (WHO) has settled a worldwide recognized classification of B-cell neoplasms. Nevertheless, for part of these pathologies involving mature B-lymphoid cells, a * To whom correspondence should be addressed. For biological aspect: Dr. Laurent Mauvieux, Institut d’He´matologie et d’Immunologie, Faculte´ de Me´decine, Universite´ Louis Pasteur, 4 rue Kirschleger, 67085 Strasbourg Cedex, France. E-mail: [email protected]. Fax: 133390-244016. For proteomic aspect: Dr. Alain Van Dorsselaer Laboratoire de Spectrome´trie de Masse Bio-Organique, ECPM, UMR 7178, Institut Pluridisciplinaire Hubert CURIEN, Universite´ Louis Pasteur, 25 rue Becquerel 67087 Strasbourg Cedex 2 France. E-mail: [email protected]. Fax: 133-390-242781. † CURIEN Multidisciplinary Institute, Louis Pasteur University. ‡ Ciphergen Biosystems GmbH. § Department of Hematology and Oncology, Hautepierre Hospital. | Institute of Hematology and Immunology, Medical School, Louis Pasteur University and Laboratory of Hematology, Hautepierre Hospital.

2258

Journal of Proteome Research 2006, 5, 2258-2269

Published on Web 08/05/2006

precise diagnostic is still difficult because morphological data are ambiguous, and no specific antigen or cytogenetic marker is available,3,4 thus leading for instance to the arbitrary definition of a subgroup called “atypical chronic lymphoid leukemia”. Also, even for well classified entities, there is a marked and unexplained heterogeneity concerning their natural histories and responses to treatment. Although, new markers as CD38,5 Zap70,6 or mutational status of immunoglobulins7 are currently under investigation, there is still a need for tools that will allow a more accurate classification, critical for the treatment decision. Many studies have shown that biological fluids (serum, urine, cerebrospinal fluid), contain an important number of biomarkers associated with various pathologies,8-16 but only few studies deal with the profiling of lymphoma diseases.17 The presence of such biomarkers may open the possibility to generate a mass spectral diagnosis that could allow a better discrimination between various lymphoproliferations pathologies.17 It may also detect diagnostic targets that could be assessed with other methods, mainly antibody-based tests. In this study, we applied SELDI-MS ProteinChip technology16,18,19 to generate specific serum protein profiles searching 10.1021/pr060058y CCC: $33.50

 2006 American Chemical Society

Potential Biomarkers in Chronic Lymphoid Malignancies

research articles

Figure 1. Schematic diagram of the strategy used for the enrichment and the purification of the proteins of interest. Numbers in bracket refer to the title numbers of the Materials and Methods section.

for potential biomarkers of chronic B-cell malignancies. The ProteinChip strategy consisted in the use of arrays on which a chromatographic phase is grafted in order to bind the different serum proteins, according to their corresponding biochemical properties. The proteins selectively retained on the surface were analyzed using a SELDI-TOF mass spectrometer. The comparison of such protein profiles from patients vs controls allowed us to find differences in protein abundance levels. This comparison was followed by a statistical analysis to confirm and validate those differences. Then the proteins or peptides of interest were further purified from the crude biological fluids, identified and characterized by different mass spectrometric methods. As the aim was to detect early markers of disease, we analyzed crude sera collected prospectively at the time of diagnosis, prior to any antitumoral treatment. Four different chromatographic arrays: anionic and cationic surfaces (Q10 and CM10), hydrophobic (H50) and an IMAC copper surface were applied three times to 39 patients vs 20 normal controls. After the protein profile analyses, different protein purification strategies were applied to identify and characterize the proteins: liquid chromatography, 1D-gel and nanoLC-MS/MS. For biomarkers in the peptide range, identification was performed directly on the chromatographic array by coupling a SELDI source to a Q-TOF-MS/MS instrument, which allows direct MS/MS analysis of the peptides.

Materials and Methods Patients. This prospective study included 39 untreated patients with chronic B-cell lymphoid malignancies. Lymphoid malignancies diagnoses were assigned on peripheral blood lymphocytes according to standard World Health Organization (WHO) criteria using classical analysis: morphology, immunophenotyping, cytogenetics, and molecular biology.

Blood samples were obtained in 7 mL BD vacutainer glass red top tubes (BD 366431) following venous blood sample procedures of Strasbourg University Hospital Laboratories (EXAM/PREL/ANNX/001). The clotted samples were centrifuged (10 min, 4 °C, 3500 × g), aliquoted in eppendorf tubes and frozen at -80 °C within 2 h after venous puncture. All patients’ sera were collected at the moment of diagnosis with informed consent, following ethics committee approval. Control serum was obtained with similar procedure from healthy people undergoing systematic check-up, using part of the material that was not necessary for clinical analyses. Controls and patients were more than 18 years old. SELDI-TOF-MS Analysis. The ProteinChip array contains 8 spots coated with a specific chromatographic surface. To determine the best profiling conditions, we have used a set of different chromatographic surfaces constituted by weak cation exchange (CM10), anion exchange (Q10), immobilized metal ion affinity capture with copper (IMAC-Cu) and hydrophobic array (H50). The normal phase (NP20) was used to bind the peptides without any selectivity. All spectra were recorded on a PBS IIC ProteinChip Array reader (Ciphergen Biosystems Inc. Fremont, USA). A first series of studies was performed to determine the optimal quantity of serum to use. In this aim, a series of serum dilution (from 1/10 to 1/500) was performed and analyzed. The best signal-to-noise and resolution obtained on two of the most intense peaks of the spectra (m/z ) 6442 and m/z ) 6640) were obtained for the 1/100 dilution (see Supporting Information Figure 1). For this study, 39 patient’s sera were tested. Sera were denatured for half an hour at room temperature by adding a 9 M urea 1% CHAPS, phosphate buffer pH 7.4 (U9 buffer) (to 10 µL of serum 90 µL of denaturing buffer was added). The arrays were washed and pre-equilibrated twice with the surface Journal of Proteome Research • Vol. 5, No. 9, 2006 2259

research articles specific binding buffers (CM10: 100 mM sodium acetate pH 4 and pH 6; Q10: 100 mM Tris pH 9 and pH 7; H50: 10% CH3CN, 0.1% TFA, 150 mM NaCl; IMAC Cu: 100 mM sodium phosphate, 0.5 M NaCl pH 7). The denatured samples were diluted 10 times in these specific binding buffers. 100 µL of each processed serum were incubated in a 96 well Bioprocessor (Ciphergen Biosystems, USA) for 1 h at room temperature. The spots were washed three times by 100 µL of the corresponding binding buffers for 5 min each, before a final rinse with deionized water. After the spots were dried, 0.5 µL of saturated solution of sinapinic acid in water/acetonitrile (1:1) with 0.5% TFA, used as matrix was deposed twice. For data acquisition three different settings were used. The array spot was divided into three mass ranges including a low mass range with a focus mass at 2000 Da and a laser intensity of 160, a middle mass range with a focus mass at 8000 Da and a laser intensity of 180 and a high mass range with a focus mass at 13000 Da and a laser intensity of 200. The SELDI-TOF-MS instrument was calibrated externally using selected peaks from an All-in-One protein mixture (Ciphergen Biosystems) (IgG; Albumin; Enolase; Carbonic Anhydrase; Myoglobin; CytochromeC; Hirudine). Software and Statistical Analysis. Analyses of the generated spectra were performed using ProteinChip software version 3.2.1 and CiphergenExpress software version 2.1. All recorded spectra were focused on a specific m/z area. The spectra were treated with an automatic baseline subtraction and normalized using total ion current normalization as implemented in the software. The peak clusters were generated taking only peaks with signal-to-noise ratio greater than 5 and the mass tolerance for cluster labeling was set to 0.2%. For p-value calculation, normalized spectra were analyzed with the Mann-Whitney U test for nonparametric data sets. All the spectra were also inspected manually in all mass ranges in order to validate the different marker candidates. Manual Integration of the Spectra. Manual integration was necessary in order to discriminate the cluster at m/z ) 13 849. As the software was not able to generate a good clustering of this peak due to a too low resolution, manual integration was undertaken in order to attribute an area to the cluster of this peak. The integration was manually realized by calculating the peak area on all the spectra and on two different acquisitions. Purification and Identification of Proteins Detected by SELDI-TOF-MS. Figure 1 shows the different strategies used for the enrichment and purification of the marker proteins. The different steps depicted in the figure appear in brackets, refer to the numbers in the materials and methods titles. Protein Enrichment by Anion Exchange Column (1). Serum fractionation employing anion exchange Q-hyper D spin columns (Ciphergen Biosystems, USA) was used as a first purification step. 100 µL of the serum was diluted in 150 µL of U9 and incubated for half an hour at room temperature. The samples were diluted twice in binding buffer, and loaded on the spin column and incubated under shacking for half an hour at room temperature. The column was centrifuged at low speed (1000 rpm for 1 min), to obtain the flow through fractions. Then, a sequential fractionation of the column with a decreasing buffer pH (pH 9; pH 7; pH 5; pH 4; pH 3; organic solvent) was performed. Each solution was incubated for 20 min on the column. All the fractions obtained during this phase were spotted on both CM10 and NP20 arrays. This spotting allows the visualization of the different proteins in each fraction in order to localize the proteins of interest. 2260

Journal of Proteome Research • Vol. 5, No. 9, 2006

Miguet et al.

Enrichment by Reverse Phase Beads (2). Fractions containing the proteins of interest were further purified using reverse phase beads RPC Poly Beads, (Biosepra, Cergy France) in a batch method. The fractions were adjusted to 10% acetonitrile 0.1% TFA and mixed with 50 µL of RPC beads equilibrated in 10% acetonitrile, 0.1% TFA and incubated for 30 min at room temperature. The fractionation and elution were performed by iterative washes with increasing amounts of acetonitrile in 0.1% TFA. A 2-µL portion of the different fractions was spotted on NP20 in order to control the elution of the biomarkers. The fractions of interest were concentrated by vacuum centrifugation and dissolved in sample buffer for SDS-PAGE. SDS-PAGE (3). Fractions containing a high ratio of proteins of interest were combined and loaded on a 16% Tris Tricine polyacrylamide gel (Invitrogen, UK) and stained overnight using colloidal coomassie blue (Safe-stain, Invitrogen). Passive Elution of Proteins from SDS-PAGE (4). Gel bands of interest were excised and washed 3 times for 15 min with 50% acetonitrile/50 mM ammonium bicarbonate. The gel pieces were then dehydrated in 100% acetonitrile (15 min), covered with 30 µL of 50% formic acid, 25% acetonitrile, 15% 2-propanol, 10% H2O and incubated for 3 h at room temperature under vigorous shaking. 2 µL of the samples were analyzed directly on a NP20 array. This step allows the comparison of the molecular mass of the passively eluted protein with the original SELDI-TOF MS spectra to confirm that the correct marker protein would be digested.20 Purification by HPLC (5). LC was performed on a Macherey Nagel Nucleosil 300-5C4 column (1 × 125 mm; particle size 5 µm) with the Waters 2690 Alliance HPLC system. The solvent system consisted of 0.1% trifluoroacetic acid in water (solvent A) and 0.1% trifluoroacetic acid in acetonitrile (solvent B). Elution was performed at a flow rate of 50 µL/min with a 2-30% gradient (solvent B) over 15 first minutes followed by a 30-90% gradient (solvent B) over 75 min before the reconditioning of the column at 98% of solvent A. The HPLC was coupled with an UV reader (Waters 996 photodiode array detector), and an ESI-TOF (LCT, Waters, MA). The elution was followed by ESI-MS and the fractions of interest were collected. Protein Digestion and LC-MS/MS Analysis (6). The gel bands or the collected fraction were digested by sequencing grade trypsin (Promega). The digests were analyzed by nanoLCMS/MS on a Bruker ESI ion-Trap HCT+. Raw data were processed and converted into a “.mgf” using Data Analysis software (Bruker). Databank searches were performed against the Swiss-Prot database using the Mascot search algorithm. Searches were done with a mass tolerance of 0.2 Da in MS mode and 0.2 Da in MS/MS mode. One missed cleavage per peptide was allowed and oxidation of methionine as variable modification was taken into account. Searches were performed without any taxonomic restriction and without constraining protein molecular mass or isoelectric point. To have a good confidence in the different protein/peptide identifications, only high quality MS/MS spectra were accepted (MASCOT score better than 35 for each spectrum), and all the MS/MS spectra were manually interpreted in order to determine the partial or complete amino acid sequences. Enrichment of Peptides below 2000 Da (7). For the enrichment of the small peptides, the serum was fractionated using Q-Hyper D anion exchange spin columns and the peptides were localized in the flow through fraction. To further deplete the samples of larger proteins like IgG and albumin, the fractions were passed through a YM50 (Millipore) membrane


filtration. The flow through of the YM50 membranes was diluted 1:4 with 100 mM sodium acetate pH 3.2 and loaded in a bioprocessor (50 µL per spot) on several spots of a CM10 array equilibrated with the same buffer. After 1 h incubation, the CM10 array was washed three times for 5 min with binding buffer (150 µL) and subsequently washed 30 s in 8 mL of ultrapure water. After drying, 0.5 µL of a 20% saturated solution of CHCA in 0.1% TFA 50% acetonitrile was added twice as matrix. The CM10 was loaded in the PCI 1000 Tandem MS interface and the peptides of interest were identified using CID-MS/MS directly from the ProteinChip array. SELDI Quadrupole Time-of-Flight (Q-TOF) Tandem MS (8). Some of the tryptic digests and the small peptide markers were analyzed on a Q-TOF Tandem mass spectrometer (QStar Pulsar i, ABI Applied Biosystems, Darmstadt, Germany) equipped with a SELDI source: PCI 1000 ProteinChip Interface (Ciphergen Biosystems). The system was externally calibrated using human Angiotensin I (m/z 1296.6853) and porcine Dynorphin A (m/z 2147.1990). Raw data were analyzed using the instruments Analyst Software (ABI). Databank searches were performed against the Swiss-Prot database using the Mascot search algorithm.

research articles

Results

Figure 2. Example of SELDI-MS spectra of two patients and two control samples in the range from 4 to 20 kDa, obtained on CM10 array at pH 4. The zones between the lines are the zones presenting interest at 8 and 13 kDa.

Protein Profiling. In this study, we performed the profiling analysis of the sera of 39 patients at diagnosis vs 20 controls using four different chromatographic surfaces (CM10, IMAC Cu, Q10, and H50). All data were reproduced at least three times, ensuring a good confidence in the results. The different coefficients of variation (CV ) σ/mean X100) were calculated on two of the most intense and invariable peaks of the untreated spectra (no normalization). The average CV on repeated experiment using the same sample was of 25%. The average CV between the different samples of the patient group was 20%, the control group was 28%, and the average CV of the all spectra (patient and controls) was of 30%. This has to be compared with the 200% threshold used for the selection of protein of interest. Profiles generated on CM10 arrays allowed the best discrimination between the control and the patient groups yielding a series of proposed biomarkers. Profiles generated on IMAC Cu also yield the same biomarkers. H50 and Q10 arrays resulted in a worse discrimination of the two groups. Figure 2 shows representative SELDI-MS spectra of two patients and two controls obtained on a CM10 array at pH 4, using sinapinic acid as matrix. Variation of intensity between patients and controls are visible at 13-14 kDa and 8-9 kDa. The spectra on Figure 2 do not show the area below 2000 Da, which presents also some differences, because specific optimized acquisition parameters are required to observe low molecular peaks (see materials and methods). An identification of the different biomarkers was then undertaken. Identification of the Peaks Observed at 13 kDa. SELDI Analysis. Two major peaks with an average mass at 13 766 and 13 876 Da are clearly detected in the control serum profiles. In the spectra obtained from the patient’s sera, these two peaks are sharply reduced. An intense peak is observed in the patient sera but not in the control with an average mass of 13 848 Da. Figure 3A shows the superposition of the spectra obtained from a patient (in bold line) and a control serum (in regular line). The scatter plot representation (Figure 3B) shows the variations of the relative intensities of these peaks observed between patients and controls: peaks at 13 766 and 13 876 Da are

consistently higher in controls than in patients. The p values (Mann-Whitney U test) of these differences were respectively 8.7 × 10-10 and 1.5 × 10-7. The peak at 13 848 Da could not be correctly clustered automatically by the software, although the overexpression of this protein is clearly visible on the patients spectra. The applied software cluster tolerance (0.2%) corresponds to an interval of more than 28 Da, which is equal to the difference between the peak at 13 848 and the 13 876 Da, preventing a good discrimination between the two peaks. Manual inspection and labeling of the peaks was then necessary. We then performed for this peak manual integration on all spectra. A scatter plot representation obtained from this manual integration is shown on Figure 3C. The statistical analysis was performed using a Mann-Whitney U test. The calculated p value of this cluster was below 1 × 10-4. To identify the different biomarkers, we have undertaken their purifications followed by MS analysis. Isolation of the Protein at 13 848 Da in the Patient Sera. A first step of purification from patient serum was performed on an anion exchange column followed by HPLC in a second step. As it is not possible to discriminate small molecular mass differences with a 1D-gel (data not shown), we have decided to use HPLC coupled with an ESI-TOF mass spectrometer in order to follow the different protein elution. The different fractions were collected, spotted on NP20 and analyzed by SELDI-MS to identify the fraction containing a protein with a mass close to the average value of 13 848 Da measured in the patient sera. One of the fraction collected yielded a major peak at 13 841.19 ( 0.87 Da observed on ESI-TOF mass spectrometer (Figure 4A), and 13 842.5 Da on the SELDI-MS. This major peak was assumed to correspond to the peak cluster at 13 848 Da. This mass measured on the purified protein is about 7 Da lower than the cluster mass measured directly on the serum by SELDI-MS. With the external calibration used for the profiling study, the mass accuracy of the SELDI-TOF is about 0.1% so a deviation of 7 Da on the average mass is possible. Also in this region beside this dominant peak at 13 848 Da several other peaks badly resolved were observed in the different sera spectra and a shift of the peak top is expected. Journal of Proteome Research • Vol. 5, No. 9, 2006 2261

research articles

Miguet et al.

Figure 3. (A) SELDI-MS spectra of the peaks in the 13 kDa area. In bold is the spectrum obtained from patient serum and in thin from control. (B) Scatter plot representation of the clusters at 13 767 and 13 877 Da respectively. (C) Scatter plot representation of the manually integrated area of the cluster at 13 848 Da.

Therefore, we assume that the mass of the protein observed with an average mass of 13 848 Da in the SELDI analyze of crude sera is in fact of 13 841.19 ( 0.87 Da (Figure 4). Identification of the Protein at 13 848 Da Detected in the Patient Group. The purified protein fraction was then digested using trypsin and the digest was analyzed by MALDI-TOFMS and ESI-Trap-MS/MS. Data obtained from peptide mass fingerprinting (MALDI experiment) and MS/MS fragmentation (nanoLC-MS/MS) was submitted to MASCOT algorithm which allowed the identification of transthyretin (TTR) as the major compound of this fraction (7 peptides, 71% coverage see Table 1). As only very minor peaks were attributed to albumin tryptic peptides, we concluded that the major protein contained in this fraction was TTR, or a slightly modified form. Purified with this TTR related protein with a mass of 13 841 Da, a minor protein (less than 20%) with a mass of 13 761.14 Da was clearly detected in the ESI spectrum (Figure 4A). This minor protein display a mass very close to the mass of the unmodified form of the TTR at 13 761.4 Da. Since in the digest peptides mixture of the TTR fraction no other protein except albumin was detected, it is likely that this protein also produced TTR digestion peptides. Therefore, we conclude that this second protein 13 761.14 Da is the unmodified form of TTR. Since the SELDI analysis of this fraction also displays a 2262


corresponding peak, we conclude that the peak with an average mass of 13 766 Da on the SELDI profiling, and measured with an exact mass of 13 761.14 Da in the ESI experiment, does correspond to the unmodified form of TTR in the patient sera with an expected mass of 13 761.4 Da. The mass difference between the two species present in the ESI-MS spectra is of 80.05 Da. Modification of 80 Da Present in the Patient Group for the Protein with an Exact Mass of 13 840 Da. Phosphorylation (OPO3H2) or sulfatation of tyrosine (O-SO3H) or thiol sulfitation (S-SO3H), could result in a 80 Da mass shift. A sulfatation is not likely, since this modification is known to be very labile, and that during the purification process very acid solutions (pH about 2) were used, which is expected to induce a loss of the sulfatation. To differentiate between the two other possibilities (O-phosphorylation and S-sulfitation), we treated the protein purified from the patient serum either with 25 mM DTT or alkaline phosphatase as described by Heintz et al.,21 and spotted it on a NP20 array. The results are shown in Figure 5. While the protein mass remained unmodified with alkaline phosphatase treatment, the reduction with DTT revealed a mass shift of about 80 Da. Such a result supports the idea that in the patient the cystein 10 (which is the only cystein of the TTR) contains a sulfite adduct.22-23

research articles


Figure 4. (A) ESI spectrum showing the collected fraction containing the TTR obtained by HPLC purification from a patient serum. The ESI spectrum shows a major form at 13 840 Da (sulfite form of the TTR) and a minor form at 13 760 Da (unmodified form of the TTR). (B) SELDI-MS spectrum of the collected HPLC purification. Table 1. Summary of the Identified Biomarker and Peptides Obtained by MS/MS Analysisa SELDI corresponding peaks (Da)

13851

name of the protein

transthyretin precursor (Prealbumin)

accession no. (Swiss-Prot)

no. of matching peptides

P02766

7

8611 8937

complement C4-A precursor complement C3 precursor

P0C0L4 P01024

1 2

1898 1867

complement C-4A precursor complement C3 precursor

P0C0L4 P01024

1 1

sequence of matching peptides

VLDAVR VEIDTK GSPAINVAVHYER AADDTWEPFASGK YTIAALLSPYSYSTTAVVTNPK ALGISPFHEHAEVVFTANDSGPB TSESGELHGLTTEEEFVEGIYK LGOYASPTAK SVOLTEK VFLDCCNYITELR NGFKSHALQLNNRQIR SSKITHRIHWESASLL

a Peptides identified by nano-LC-MS/MS are underlined. Peptides identified by direct SELDI-Q-TOF-MS/MS are in italics. Peptides identified by both techniques are underlined and in italics.

To show that the 80 Da modification affects cystein in position 10, we tried to isolate the corresponding tryptic peptide and to measure its mass. This digested protein was analyzed on both MALDI and ESI mass spectrometers, but neither the modified nor the unmodified peptides (CPLMVK) were detected in repeated experiments. When tryptic digestion of pure commercial TTR was performed and analyzed by nanoLC-MS/MS, the unmodified peptide was also not detected. The fact that this peptide was not recovered is very surprising but this was

already described by The´berge et al.23 We conclude then that the protein observed in the patient sera and not in the control with an exact mass of 13 840 Da could only be the sulfite form of the transthyretin protein.24 Identification for Proteins at 13 766 and 13 876 Da Detected in the Control Sera. Two consecutive chromatographic steps were performed on a control serum as described upper for the patient sera. A fraction containing masses at 13 760.93 ( 0.69 and 13 879.92 ( 0.37 Da observed on ESI-TOF mass Journal of Proteome Research • Vol. 5, No. 9, 2006 2263

research articles

Miguet et al.

Figure 5. Treatment of the purified TTR with 25 mM DTT and alkaline phosphatase on NP20. The purified TTR is shown on the top panel. A mass shift of about 80 Da is observed when the TTR is treated with 25 mM DTT (bottom spectrum) but no shift after treatment with alkaline phosphatase (middle panel).

spectrometer (close from the average value of 13 766 and 13 876 Da observed on the SELDI spectra), was collected and spotted on NP20 (13 759 and 13 880 Da). The retention time of these proteins was similar to that of the protein identified as TTR in the patient sera. This collected fraction was digested using trypsin and analyzed by nanoLC-MS/MS. Data obtained by MS/MS fragmentation was submitted to MASCOT algorithm. As all the peptides were attributed to TTR (5 peptides 37% coverage), and only few contamination by albumin and immunoglobulin, we conclude therefore that the peak at 13 760.93 Da in the control sera is the unmodified form of TTR as already found in the patient’s sera, and that the peak at 13 879.92 Da could only be the cysteinylated form of TTR with an expected mass of 13 880.4 Da.21,23 To confirm this hypothesis, the protein was treated with DTT as described before and the protein mass revealed the expected mass shift of about 120 Da (see Supporting Information Figure 2). TTR Dosage by Immunonephelometry. TTR levels were assessed in 10 sera of each group by an independent laboratory on a clinical validated routine immunophelometry analyzer (BN2, DadeBehring, UK). This technique measures all forms of TTR and is independent of the different post-translational modifications of the protein. Normal values for TTR in serum are comprised between 0.2 and 0.4 g/L. The average concentration measured in the patient group was respectively 0.282 g/L [0.097-0.563] (SD ) 0.14) and 0.283 g/L [0.241-0.368] (SD ) 0.05) for controls, very close to the normal values. No significant difference variation in the TTR concentration was observed between the two groups (Mann-Whitney test p ) 0.6842). Its abundance correlated with m/z area measured in SELDI-MS spectra showing that the 13 800 m/z peaks corresponded to TTR. The results were highly correlated (R2 > 0.90 for the patient group and R2 > 0.92 for the control group) validating the identification of the protein (see Figure 6). Peaks between 8 and 9 kDa. SELDI Analysis. The second series of peaks of interest was observed between 8 and 9 kDa. In this zone, two peaks (at 8.6 kDa and 8.9 kDa) were observed as being overexpressed in a significant number of patient sera. Figure 7A shows two typical spectra obtained from a patient 2264


Figure 6. Correlation between the TTR dosage in g/L by immunoassays, and the calculated area of the different forms of the TTR measured on the SELDI-MS spectra.

and a control. The scatter plots summarize the experiment and show the increase of these two peaks in the patient group (Figure 7B). The p value for the 8936 and 8611 Da peaks were 9.7 × 10-8 and 1 × 10-3, respectively. To identify these two markers, purification was undertaken, using classical chromatographic steps on sera samples. Purification of the Proteins. Enrichment of these proteins was achieved using serum from one patient displaying a large peak at 8.6 and 8.9 kDa. A first step of enrichment was


research articles

Figure 7. (A) SELDI-MS spectra of patient and control sample in the 8 kDa area. (B) Scatter plot representation of the cluster at 8937 and 8612 Da.

performed using an anion exchange fractionation on Q-Hyper D spin columns. Then, a reverse phase step using beads and a batch method was used. This second step allowed a better enrichment than HPLC, and both proteins of interest were found in the fraction containing 40% acetonitrile. After this second enrichment step, the small proteins were purified using a SDS-PAGE polyacrylamide gel and the bands of interest were cut-out from the gel. Half of each gel band was used for passive elution of the protein. The second half of the band was digested using trypsin. Proteins contained in the passive elution fraction were analyzed on a NP20 array. This step validated that the bands cut-out for tryptic digestions did contain the marker of interest at m/z ) 8612 and m/z ) 8937 found in the SELDIMS profiles. The gel bands were therefore submitted to tryptic digestion followed by MS analysis. Identification of the Protein at m/z ) 8937 and m/z ) 8612. The tryptic digest mixtures of proteins at 8936 and 8611 Da were analyzed by both nanoLC-MS/MS and by SELDI-QTOFMS/MS (Table 1). From the tryptic digest of the 8936 Da peak, analyzed by nanoLC-MS/MS, one peptide was sequenced, allowing the identification of the C3 complement protein. Subsequent SELDI-Q-TOF-MS/MS have confirmed this identification by sequencing a second peptide. The combination of both techniques has allowed a better identification and a better sequence coverage on the identified protein. The location of the identified peptide on the C3 complement protein corresponds to a known processed fragment of this protein: the C3a anaphylatoxin fragment. The calculated mass of this C3a anaphylotoxin protein without the C-terminal arginine is of 8938 Da, fitting very well with the experimental SELDI mass which is of 8936 Da. In the same way, the digest of the gel band containing the peak at m/z ) 8612, one single peptide was sequenced (by both techniques), with a sequence corresponding to a peptide originating from the C4 complement protein. As for the protein

at 8936 Da identified above, the sequence corresponds to a part of the C4a anaphylotoxin processed protein, having a calculated mass without C-terminal arginine of 8608 Da, fitting very well with the experimental SELDI mass at 8611 Da. These two biomarker candidates correspond unambiguously to the C3a and C4a anaphylotoxin protein. Peaks in the Region below 2000 Da. SELDI Analysis. In 38% of the patient sera (15 patients sera from the 39 tested) a pattern of several peptide signals were observed (m/z ) 1897; 1867; 1759; 1740, and 1690 Da) that were not observed in any samples of the control group making them highly specific for the disease group. Figure 8A shows the different peaks observed in the small peptide range and the scatter plots are represented in Figure 8B. To identify these peptides different purification steps were performed. Purification of the Peptides. To identify the different peptides, enrichment was performed in the same way than for the other proteins by a first step of anion exchange fractionation followed by a cutoff membrane step in order to remove large proteins like albumin and immunoglobulins. The flow-through was then enriched on the chromatographic cation exchange (CM10) array as described in the materials section. The array could then be directly analyzed in the tandem MS instrument equipped with a SELDI source. Identification of the Peptides. Figure 9A shows the MS spectrum in the low mass region using the SELDI tandem MS source. As the acquisitions were performed on a SELDI-Q-TOF mass spectrometer having a higher resolution than the standard SELDI-TOF (Figure 8A), the observed mass of the different peaks are much more precise. An example of the CID-MS/ MS spectra directly from the CM10 array is shown for the peptide at 1738.96 Da in Figure 9B. Even if the MS/MS spectrum is noisy, there are a lot of y and b fragments allowing the identification of the sequence with a relatively good confidence. This technique was applied to all the peptides and allowed clear sequence identification by MS/MS of all peptides Journal of Proteome Research • Vol. 5, No. 9, 2006 2265

research articles

Miguet et al.

Figure 8. (A) SELDI-MS spectra of a patient and control sample in the low mass region. (B) Scatter plot representation of the peaks at 1898 and 1867 Da.

Figure 9. (A) MS spectrum on the ProteinChip Tandem MS interface. The region of the peptide markers below 2000 Da is shown. The peptides were identified by CID-MS/MS from the CM10 array. (B) MS/MS spectrum of the precursor ion with m/z 1739.96. The fragment spectra shows predominantly y and b ions as indicated allowing a clear identification.

that are listed in Figure 9A and in Table 1. The peptide at 1864.05 Da with a sequence of: SSKITHRIHWESASLL (position 1304-1319) was identified as the complement cleavage fragment C3f desArg originating from the alpha chain of the complement C3. In addition, 4 additional peptides were also assigned to C3f fragments, which are progressively truncated at the N-terminal (m/z ) 1865.04 (1305-1319); m/z ) 1778.02 (1306-1319); m/z ) 1690.96 (1307-1319); m/z ) 1449.78 (1308-1319); m/z ) 1348.71 (1309-1319)). The peptide at m/z ) 1896.06 Da, with a sequence of: NGFKSHALQLNNRQIR (position 1337-1352) was identified as a C4d fragment originating from a cleavage of the complement C4b protein. In addition, a truncated C4d peptide was found, missing one amino acid at the C terminus (m/z ) 1739.96 (1337-1351)). This different pattern with successive single amino acid truncations indicate that aminopeptidases and carboxypeptidases seem to be involved in the generation of the low Mr peptide fragments. 2266


Discussion Analysis of serum and other body fluid by proteomics techniques like SELDI-MS is increasingly used to discover and to identify new biomarkers. These techniques have been applied to enhance the diagnosis of numerous diseases, ranging from neurodegenerative disease like Alzheimer disease to congestive heart failure and different neoplasia,25-28 for reviews see.16,29 We evidence here several proteins that are differently expressed in sera of chronic B cell lymphoid proliferations samples vs control subjects. These differences were observed at the moment of the diagnosis thus avoiding the biases induced by the chemotherapy or immunotherapy, or a long disease progression. They were independent of the degree of hyperlymphocytosis. Overexpression of the sulfite form of TTR at cystein 10 was identified for the first time in cancer and a series of peptides below 2 kDa originating from C3 and C4 complement protein was also identified. Unmodified and cysteinilated forms of the TTR were down-regulated in patients,


and two intense peaks at 8936 and 8611 Da, not detected in controls, were identified as C3a and C4a anaphylatoxins in patients. The use of unfractionated serum may potentially mask low abundant proteins, but in our hands, the study of fractionated samples did not show additional biomarker. As a consequence, most of these candidate biomarkers are not synthesized by the tumor itself contrary to more classical tumor markers that are released into the bloodstream. Nevertheless, such modified forms of common serum proteins or fragments of these proteins might be indirectly linked to the disease state and may be useful for diagnosis of variant diseases. It can be assumed that the disease will result in the differential secretion, release, or activation of enzymes such as proteases or other processing enzymes from the affected cells or as a response to cell damage. If differences in modifying enzymes are manifest in the disease compared to the control, then it can be further hypothesized that high abundant serum proteins acts as substrates for these enzymes and as a result each disease state is associated by different processing and modification of major serum proteins, which can be visualized with SELDI-MS experiments. Such an explanation has been given in recent publications to discuss the results obtained by mass spectrometric protein profiling experiments.26,27 Marshall et al. detected in sera from myocardial infract patients a specific pattern of peptides consisting of the complement released peptide C3f and fibrinogen A peptides and truncated forms from these peptides. In a different work, Fung et al. (2005) demonstrated that two host proteins, identified as candidate marker for ovarian cancer by SELDIMS, transthyretin and inter-alpha trypsin inhibitor are posttranslationally modified and these modifications occur to different extend in different cancer types. Using these modified forms in multivariant analysis should improve classification of different cancer types.26 After the analysis of the SELDI protein profiles, we focused in identifying all the proteins and peptides which showed significant differences between patients and controls. Optimized enrichment techniques (HPLC for the different forms of TTR, 1D-gel for peaks at 8 kDa, and on-Chip enrichment for small peptides) combined with MS/MS analysis revealed an unambiguous identification of the potential biomarkers. To date, mass spectrometry study is the only technique able to see the sulfite modification on the TTR. Effectively, classical 2DE approach, using reducing conditions (about 1% DTT), will remove the thiol modification and prevent its detection, and no specific antibody is available. In the peptide range, the direct coupling of the chromatographic arrays to a QTOF instrument proved to be a fast route to identification of markers, since only partial enrichment of the peptides is needed and can be achieved on the cation exchange arrays as shown in tear fluid.30 TTR is a 55 kDa homotetramer, formerly known as prealbumin for its electrophoretic mobility anodal to albumin. TTR binds with thyroid hormones and also form a complex with retinal binding proteins, playing a role for the transport of vitamin A. TTR variants have been described and associated with amyloid deposits in cardiac or neural tissues.21,31-32 Synthesized by the liver and choroids plexus of the brain, it is used as a marker for denutrition.33 Several physiological posttranslational modifications have been identified for this protein, mainly sulfonation, cysteinylation, conjugation with glutathion or cysteinyl glycine and amino acid substitution.21,23,24,34 Here, in all patients, the distribution of different physiologically forms of TTR was modified: the sulfite form was up-regulated,

research articles contrary to the unmodified and cysteinylated forms that were down regulated in controls. Down regulation of unmodified TTR has been already described in cancers11,35 and denutrition.36 Also, variations of the level of posttranslationally modified TTR have been already described in ovarian and to a lesser extent in colon cancers, but only down regulation was reported.26 In all of the patients studied here, we observed an up-regulation of the S-sulfonated TTR, which has to our knowledge not been described previously in cancer. It is known that S-sulfonated TTR is increased in two hereditary diseases: molybdenum cofactor deficiency37 and isolated sulfite oxidase deficiency.38 In these diseases, defect of sulfite oxidase is responsible for the increase of the ratio of S-sulfonated vs unmodified TTR, which can be considered to be a diagnostic tool.39 It can be hypothetized that a possible modification of the activity of sulfite oxidase may occur in chronic B-cell malignancies, that is not observed in other studies cancers (breast, colon, ovarian, prostate cancers).11,26 The molecular origin of such an effect is presently not understood. Even if measured in limited groups, TTR levels were not statistically different in patients and controls that are both in normal range. This suggests that there is no modification of TTR synthesis in patients, and the sulfited TTR was independent of the total quantity of TTR in serum. Chronic lymphoid malignancies are accumulative disease, due notably to a lack of apoptosis, which differs from most of other cancers that are more proliferative. The health status of chronic lymphoroliferative disease is often conserved, at least at the time of diagnosis. In this context, a normal value of TTR concentration is not surprising. Then, the increase of the sulfite form of the TTR observed in the patient group is probably linked to a posttranslational modification induced by the disease. This observation is different from the decrease of TTR described in previous studies that are possibly associated with the alterated status of cancer patients. The variation in the ratio of the different post-translational forms of the TTR demonstrated in this study and only observable by mass spectrometry techniques, and not by immunoassays which detect all forms of TTR. The second family of differentially expressed proteins was complement proteins of the innate immune system, which activate antibodies to eliminate pathogens. The complement system is composed of more than 25 different proteins produced by different tissues and cells including hepatocytes, macrophages, and gut epithelial cells. These proteins are activated by a variety of agents and their activation proceeds in a cascade fashion leading to lysis. In this study, two cleaved forms of the C3 and C4 complement (C3a and C4a) known as anaphylatoxins were substantially elevated in patients, suggesting the activation of the complement system via antigenantibody complexes (the classical pathway, opposite to pathogen activation using alternate pathway). Accordingly, in independent serum samples, some patients harbored an activation of the complement system based on the measurement of the hemolytic activity of complement and on C3/C4 ELISA dosage (data not shown). Low C3 and C4 levels have been previously described in CLL.40 Impaired classical pathway of the complement has been already observed in such patients41 that could be associated with short survival.42 In the same way, we found elevated levels of C3a and C4a using SELDI-TOF that correspond to an activation of the classical pathway. As previously studied by Ohishi et al.,43 this activation could be induced by cross-linked membrane IgM of pathologic lymphoid Journal of Proteome Research • Vol. 5, No. 9, 2006 2267

research articles cells, that triggers the classical pathway of the complement. Clinical relevance of these observations has to be studied in larger groups. In a recent SELDI study, C3 anaphylatoxin and a truncated form have been detected and validated as candidate markers up-regulated in breast cancer.44 Small peptides (under 2000 Da) specifically detected in patients, were also originating from complement proteins. They corresponded to the C3f and C4d fragment, that also originate from the cleavage of C3 and C4 following activation. After the cleavage of C3 into C3a and C3b, serine esterase factor I cleaves C3b into a inactive product iC3b, releasing a small peptide: C3f. In mouse45 and humans,46 C3f is further digested, resulting in progressive loss of N or C-terminal amino acids,26,45 thought to participate to its elimination from serum.39 We observed in patients (and not in controls) the same C3f peptide (SSKITHRIHWESASLLR) that was identified in myocardial infarction patients resulting from complement activation.27 Furthermore, in the work of Marschall et al. and in our study, N-terminal truncations of the peptide were detected most likely resulting from the activity of a N-terminal aminopeptidases. Interestingly, we also detected a peptide derived from complement C4: C4d and one N-terminal truncated form of this peptide. Villanueva et al. studied the serum peptidome pattern and compared control with samples from breast, prostate, and bladder cancer have also observed differential expression of C3f and C4d peptides.47 In conclusion, we describe in this prospective study the first analysis of serum originating from chronic B-lymphocyte proliferations using SELDI-TOF. Reproducible results were obtained, showing the differential expression of posttranslationally modified forms of transthyretin, proteins and small peptides resulting from the activation of complement. Although the C3 and the C4d peptides have been previously detected in cancer and noncancer diseases, the elevated sulfited TTR have not been previously described in cancer. Moreover, these posttranslational modifications and protein processing are only detectable using mass spectrometry analysis. We have applied decision tree algorithm (classification and regression tree [CART] software (See Supporting Information Figure 3) on the 59 samples in this study and have obtained a pattern model using only two peaks: the unmodified form of transthyretin at 13767 Da and the anaphylatoxin C3a peak at 8937 Da. In this model, all the samples were correctly classified in corresponding control and patient group (sensitivity of 100% and selectivity of 100%), so such an approach deserves additional studies. The interest of these findings should be confirmed using a larger set of patient samples that is possible using automatized SELDITOF procedures. In larger studies, attention should also be paid to peaks that were sporadically detected in only a few patients and thus not described here, but could be additional markers of disease subgroups. Abbreviations. SELDI-TOF, surface enhanced laser desorption ionization time of flight; TTR, transthyretin; CID, collision induce dissociation; ESI, electrospray ionization; CLL, chronic lymphoid leukemia.

Acknowledgment. Laurent MIGUET gratefully acknowledges the ARC foundation for financial support. The authors thank Dr J. Goetz for help with complement dosage and helpful discussion. The authors thank Dr. S. Vorderwu ¨ lbecke for the biomarker pattern analysis, and the Laboratoire de Biochimie Generale et Specialise of the Hopital Civil de Strasbourg for TTR dosage. The authors thank also T. Rabilloud for his helpful 2268


Miguet et al.

discussions. This work was partly supported by the Program Hospitalier de Recherche Clinique National No. 2674, Hoˆpitaux Universitaires de Strasbourg, and the Association The´rapie Ge´nique et Cancer (ATGC).

Supporting Information Available: Supporting Figures 1-3. This material is available free of charge via the Internet at http://pubs.acs.org. References (1) Harris, N. L. In World Health Organization Classification of Tumours; Jaffe, E. S., Harris, N. L., Stein, H., Vardiman, J. W., Eds.; IARC Press: Lyon, France 2001; 121-130. (2) Matutes, E.; Owusu-Ankomah, K.; Morilla, R.; Garcia Marco, J.; Houlihan, A.; Que, T. H.; Catovsky, D. Leukemia 1994, 8, 16401645. (3) Urbano-Ispizua, A.; Matutes, E.; Villamor, N.; Sierra, J.; Pujades, A.; Reverter, J. C.; Feliu, E.; Cervantes, F.; Vives-Corrons, J. L.; Montserrat, E. Br. J. Hamaetol. 1992, 81, 178-183. (4) Matutes, E. J. Clin. Pathol. 1995, 48, 194-197. (5) Domingo-Domenech, E.; Domingo-Claros, A.; Gonzalez-Barca, E.; Beneitez, D.; Alonso, E.; Romagosa, V.; De Sanjos, S.; Petit, J.; Granena, A.; Fernandez de Sevilla, A. Haematologica 2002, 87, 1021-1027. (6) Durig, J.; Nuckel, H.; Cremer, M.; Fuhrer, A.; Halfmeyer, K.; Fandrey, J.; Moroy, T.; Klein-Hitpass, L.; Duhrsen, U. Leukemia 2003, 17, 2426-2434. (7) Hamblin, T. J.; Davis, Z.; Gardiner, A.; Oscier, D. G.; Stevenson, F. K. Blood 1999, 94, 1848-1854. (8) Coombes, K. R.; Morris, J. S.; Hu, J.; Edmonson, S. R.; Baggerly, K. A. Nat. Biotechnol. 2005, 23, 291-292. (9) Anderson, N. L.; Anderson, N. G. Mol. Cell. Proteomics 2002, 1, 845-867. (10) Anderson, N. L.; Polanski, M.; Pieper, R.; Gatlin, T.; Tirumalai, R. S.; Conrads, T. P.; Veenstra, T. D.; Adkins, J. N.; Pounds, J. G.; Fagan, R.; Lobley, A. Mol. Cell. Proteomics 2004, 3, 311-326. (11) Zhang, Z.; Bast, R. C., Jr.; Yu, Y.; Li, J.; Sokoll, L. J.; Rai, A. J.; Rosenzweig, J. M.; Cameron, B.; Wang, Y. Y.; Meng, X. Y.; Berchuck, A.; Van Haaften-Day, C.; Hacker, N. F.; de Bruijn, H. W.; van der Zee, A. G.; Jacobs, I. J.; Fung, E. T.; Chan, D. W. Cancer Res. 2004, 64, 5882-5890. (12) Voshol, H.; Brendlen, N.; Muller, D.; Inverardi, B.; Augustin, A.; Pally, C.; Wieczorek, G.; Morris, R. E.; Raulf, F.; van Oostrum, J. J. Proteome Res. 2005, 4, 1192-1199. (13) Vranckx, R.; Savu, L.; Lambert, N.; de Conchard, G. V.; Grosse, R.; Mourey, M. S.; Corman, B. Am. J. Physiol. 1995, 268, 536548. (14) Hu, Y.; Zhang, S.; Yu, J.; Liu, J.; Zheng, S. Breast 2005, 14, 250255. (15) Paradis, V.; Degos, F.; Dargere, D.; Pham, N.; Belghiti, J.; Degott, C.; Janeau, J. L.; Bezeaud, A.; Delforge, D.; Cubizolles, M.; Laurendeau, I.; Bedossa, P. Hepatology 2005, 41, 40-47. (16) Wiesner, A. Curr. Pharm. Biotechno. 2004, 5, 45-67. (17) Lin, Z.; Jenson, S. D.; Lim, M. S.; Elenitoba-Johnson, K. S. Mod. Pathol. 2004, 17, 670-678. (18) Carlson, S. M.; Najmi, A.; Whitin, J. C.; Cohen, H. J. Proteomics 2005, 5, 2778, 2788. (19) Issaq, H. J.; Conrads, T. P.; Prieto, D. A.; Tirumalai, R.; Veenstra, T. D. Anal. Chem. 2003, 75, 148A-155A. (20) Le Bihan, M. C.; Tarelli, E.; Coulton, G. R. Proteomics 2004, 4, 2739-2753. (21) Heintz, D.; Wurtz, V.; High, A. A.; Van Dorsselaer, A.; Reski, R.; Sarnighausen, E. Electrophoresis 2004, 25, 1149-1159. (22) Lim, A.; Prokaeva, T.; McComb, M. E.; O’Connor, P. B.; Theberge, R.; Connors, L. H.; Skinner, M.; Costello, C. E. Anal. Chem. 2002, 74, 741-751. (23) The´berge, R.; Connors, L.; Skinner, M.; Skare, J.; Costello, C. E. Anal. Chem. 1999, 71, 452-459. (24) Lim, A.; Prokaeva, T.; McComb, M. E.; Connors, L. H.; Skinner, M.; Costello, C. E. Protein Sci. 2003, 12, 1775-1785. (25) Issaq, H. J.; Veenstra, T. D.; Conrads, T. P.; Felschow, D. Biochem. Biophys. Res. Commun. 2002, 292, 587-592. (26) Fung, E. T.; Yip, T. T.; Lomas, L.; Wang, Z.; Yip, C.; Meng, X. Y.; Lin, S.; Zhang, F.; Zhang, Z.; Chan, D. W.; Weinberger, S. R. Int. J. Cancer 2005, 115, 783-789. (27) Marshall, J.; Kupchak, P.; Zhu, W.; Yantha, J.; Vrees, T.; Furesz, S.; Jacks, K.; Smith, C.; Kireeva, I.; Zhang, R.; Takahashi, M.; Stanton, E.; Jackowski, G. J. Proteome Res. 2003, 2, 361-372.

research articles

Potential Biomarkers in Chronic Lymphoid Malignancies (28) Tolson, J.; Bogumil, R.; Brunst, E.; Beck, H.; Elsner, R.; Humeny, A.; Kratzin, H.; Deeg, M.; Kuczyk, M.; Mueller, G. A.; Mueller, C. A.; Flad, T. Lab. Invest. 2004, 84, 845-856. (29) Xiao, Z.; Prieto, D.; Conrads, T. P.; Veenstra, T. D.; Issaq, H. J. Mol. Cell. Endocrinol. 2005, 230, 95-106. (30) Grus, F. H.; Podust, V. N.; Bruns, K.; Lackner, K.; Fu, S.; Dalmasso, E. A.; Wirthlin, A.; Pfeiffer, N. Invest. Ophthalmol. Vis. Sci. 2005, 46, 863-876. (31) Nishi, H.; Kimura, A.; Harada, H.; Hayashi, Y.; Nakamura, M.; Sasazuki, T. Biochem. Biophys. Res. Commun. 1992, 187, 460466. (32) Ueno, S.; Fujimura, H.; Yorifuji, S.; Nakamura, Y.; Takahashi, M.; Tarui, S.; Yanagihara, T. Brain 1992, 115, 1275-1289. (33) Kuszajewski, M. L.; Clontz, A. S. Nursing 2005, 35, 70-71. (34) Schweigert, F. J.; Wirth, K.; Raila, J. Proteome Sci. 2004, 2, 5. (35) Mahlck, C. G.; Grankvist, K. Gynecol. Obstet. Invest. 1994, 37, 135140. (36) Fuhrman, M. P.; Charney, P.; Mueller, C. M. J. Am. Diet. Assoc. 2004, 104, 1258-1264. (37) Kishikawa, M.; Nakanishi, T.; Shimizu, A.; Yoshino, M. Pediatr. Res. 2000, 47, 492-494. (38) Sass, J. O.; Nakanishi, T.; Sato, T.; Shimizu, A. Ann. Clin. Biochem. 2004, 41, 157-159. (39) Kishikawa, M.; Sass, J. O.; Sakura, N.; Nakanishi, T.; Shimizu, A.; Yoshioka, M. Biochim. Biophys. Acta 2002, 1588, 135-138.

(40) Schlesinger, M.; Broman, I.; Lugassy, G. Leukemia 1996, 10, 15091513. (41) Hidvegi, T.; Ermolin, G. A.; Efremov, E. E.; Dikov, M. M.; Kurmanova, L. V.; Vnashenkova, G. V.; Merkulova, M. V.; Kokai, M.; Panya, A.; Fust, G. Immunol. Lett. 1989, 22, 1-6. (42) Varga, L.; Czink, E.; Miszlai, Z.; Paloczi, K.; Banyai, A.; Szegedi, G.; Fust, G. Clin. Exp. Immunol. 1995, 99, 112-116. (43) Ohishi, K.; Kanoh, M.; Shinomiya, H.; Hitsumoto, Y.; Utsumi, S. J. Immunol. 1995, 154, 3173-3179. (44) Li, J.; Orlandi, R.; White, C. N.; Rosenzweig, J.; Zhao, J.; Seregni, E.; Morelli, D.; Yu, Y.; Meng, X. Y.; Zhang, Z.; Davidson, N. E.; Fung, E. T.; Chan, D. W. Clin. Chem. 2005, 51, 2229-2235. (45) Yeo, S.; Roh, G. S.; Kim, D. H.; Lee, J. M.; Seo, S. W.; Cho, J. W.; Kim, C. W.; Kwack, K. Proteomics 2004, 4, 3308-3317. (46) Harrison, R. A.; Farries, T. C.; orthrop, F. D.; Lachmann, P. J.; Davis, A. E. Complement 1988, 5, 27-32. (47) Villanueva, J.; Shaffer, D. R.; Philip, J.; Chaparro C. A.; ErdjumentBromage, H.; Olshen A. B.; Fleisher, M.; Lilja, H.; Brogi, E.; Boyd, J.; Sanchez-Carbayo, M.; Holland E. C.; Cordon-Cardo, C.; Scher H. I.; Tempst, P. J. Clin. Invest. 2006, 116, 271-284.

PR060058Y

Journal of Proteome Research • Vol. 5, No. 9, 2006 2269

Discovery and Identification of Potential ... - ACS Publications

Recommend Documents