Translating Proteomic Discoveries into Clinical Diagnostics

May 9-10, 2008
Grand Hyatt Hotel
Seattle, WA
AACC thanks all of the scientists who have shared their work and congratulates the winners of our poster conference.
- First Place: Traci Kruer (University of Louisville, Louisville, KY)
- Second Place: Yuan Tian, PhD (Johns Hopkins Medical Institutions, Baltimore, MD)
- Third Place: Sean Barnes, MD, PhD (University of Louisville, Louisville, KY)
Please scroll down to view their award-winning efforts, as well as the work of many other fine scientists.
FIRST PLACE
Expression of novel estrogen-response-element binding proteins in human breast carcinoma predicts clinical behavior
Kruer TL, Kerr DA, Wittliff JL
University of Louisville, Department of Biochemistry & Molecular Biology, Institute for Molecular Diversity and Drug Design, Louisville, KY
While investigating estrogen response element (ERE) binding properties of estrogen receptor-a (hERa) in human breast cancer cytosols, lower molecular weight ERE-binding proteins (ERE-BP) were observed. Recognition properties of ERE-BP were evaluated by electrophoretic mobility shift (EMSA) with ERE sequences present in the 5’-flanking region of the estrogen responsive gene, VitA2.
Cytosols, prepared from human breast cancer samples in 50 mM K2HPO4/KH2PO4 buffer, pH 7.4 with 1.5 mM EDTA, 10 mM Na2MoO4, 1 mM PMSF & 10 μM monothioglycerol, were incubated 16 hr, 4°C with [32P]ERE sequences and separated by EMSA. A method of estimating the ERE-BP level was developed by measuring band intensity from the EMSA profile, expressed in digital light units (DLU)/µg protein and normalized to total DLU. Levels of ERE-BP were correlated with clinical characteristics e.g., overall and disease-free survival, nodal status, tumor maker levels, grade, stage, race, age. No correlation was observed between ERE-binding activities and patient age, race, nodal status, cancer grade or stage. However, when overall or disease-free survival of these groups of patients was analyzed as a function of either above or below mean ERE-binding protein expression, a surprising result was detected from Kaplan-Meier plots. Patients in the above mean group show a lower overall survival than patients in the below mean group (p<.0001). This result was also shown for disease free survival (p=0.0269). Addition of ERE-binding protein expression within a single ER status (i.e., either positive or negative) identified subsets of patients exhibiting decreased overall survival. Strikingly, breast cancer patients with ER negative status and high ERE-binding protein expression exhibited the poorest overall survival of any of the groups. Although these data are preliminary, they clearly support the hypothesis that the presence of these novel ERE-binding proteins in a breast carcinoma is an indicator of poor prognosis for breast cancer patients.
SECOND PLACE
Specific profiling of sialylated glycoproteins from breast carcinoma
Tian Y1, Esteva F J2, Tao SC1, Zhu H1, Song J1, Zhang Z1, Zhang H1
1Department of Pathology and Molecular Sciences, Johns Hopkins University, Baltimore, MD; 2Department of Breast Medical Oncology, University of Texas, M. D. Anderson Cancer Center, Houston, TX
Background: Glycosylation is one of the most important modifications to proteins and protein glycosylation patterns are associated with disease development, such as cancer. The heterogeneity of glycosylation is presented by the modifications of various sugars and a number of potential glycosylation sites on proteins. The question of which specific glycosylation changes are related to a specific disease becomes urgent to be addressed for early-stage diagnosis and therapy of diseases. The objective of this study is to determine the altered glycosylation related to breast tumorigenesis and identify the altered glycoproteins in breast cancer patients.
Methods: To determine the glycosylation changes associated with breast cancer, lectin microarray was employed in this study. The total proteins were extracted from breast tumor and their paired normal tissues, and then used to probe lectin microarray chips. Sialylation was one of glycosylation that showed consistent changes in breast cancer tissues compared to normal tissues. To identify and quantify sialylated glycoproteins in breast cancer tissues, a specific isolation of sialylated glycoproteins was developed and the specificity was determined using human alpha-1 acid glycoprotein, a known sialylated glycoprotein. Briefly, the sialylated glycopeptides were specifically isolated using solid phase capture of sialoglycopeptides using hydrazide chemistry. N-linked sialylated glycopeptides were profiled from paired cancer and normal tissues of three breast cancer patients using liquid chromatography mass spectrometry and tandem mass spectrometry (LC-MS and LC-MS/MS). Results: The expression pattern of N-linked sialylated glycopeptides were generated, and compared to patterns from tryptic peptides and total N-linked glycopeptides isolated from the same patients. N-linked sialoglycopeptides and glycoproteins specifically up-regulated were determined. We further validated the results from the glycoproteomic study using Western blot analyses. We showed that higher expression and sialylation levels of versican in late stage breast cancer patients. Analysis of 25 pairs of normal and cancer breast cancer tissues using immunohistochemistry indicated two expression patterns of versican in breast carcinoma. Statistic analysis of immunohistochemistry results indicated that the epithelial versican expression was highly associated with lymph node metastasis and pathological stages.
Conclusion: First, this study identified the sialylation changes associated with breast carcinoma using lectin microarray. Second, an isolation method specific to N-linked sialoglycopeptides was developed and the specificity was evaluated using a known sialylated glycoprotein. Third, the N-linked sialoglycopeptides were profiled from breast cancer tissues and several proteins were identified with up-regulated expression in protein sialylation. One of these proteins, versican, was further validated by Western blot and immunohistochemistry.
THIRD PLACE
C-terminal adiponectin receptor peptide fragment is missing or has low expression in plasma of diabetics
Pugia MJ1, Barnes SL2, Zercher A1, Brock D1, Foltz M1, Valdes, Jr. R2, Jortani SA2
1Siemens Healthcare Diagnostics, Urinalysis Research and Development, Elkhart, IN; 2University of Louisville School of Medicine, Department of Pathology and Laboratory Medicine, Louisville, KY
Background: Patients with metabolic syndrome or diabetes have significantly lower plasma concentrations of adiponectin (Adpn) in plasma compared with nondiabetic controls. This hormone is released by the adipose tissue to slow liver production of glucose and to initiate the oxidation of fatty acids and glucose in muscle. Patients lacking Adpn have difficulty controlling insulin and obesity. Adpn elicits its effects through interaction with its receptors on the cell surfaces in various tissues including muscle (AdipoR1) and liver (AdipoR2). Using an immuno-affinity mass spectrometry approach, we have discovered fragments of the AdipoR1 receptor in plasma. We have also recently developed a monoclonal antibody specific to the C-terminal fragment of AdipoR1. This antibody has been used in the immuno-affinity mass spectrometry studies confirming our original findings. Furthermore, using the same monoclonal antibody, we have developed a quantitative ELISA method to measure the concentrations of the AdipoR1 receptor fragments in plasma.
Hypothesis: The plasma concentrations of AdipoR1 C-terminal fragment are different in diabetics as compared to the controls.
Objective: To use the new ELISA assay to examine the concentrations of C-terminal fragment of AdipoR1 in plasma of diabetic and non-diabetic controls. Methods: A monoclonal antibody made against the last 32 amino acids of the C-terminal fragment of AdipoR1 was used. Blood samples were collected in heparin tubes from patients with diabetes (n = 69) and healthy controls (n = 75). In cases where the measured concentrations of the C-terminal fragment were less than the detection limit or were undetected, we used the 100 pg/mL value for purposes of data analysis. For samples with concentrations greater than the upper limit of quantitation (i.e. 2000 pg/mL), we performed appropriate dilutions using ELISA assay buffer.
Results: The mean ± SEM for concentration of AdipoR1 C-terminal fragment (i.e. 404.3 ± 111.7 pg/mL) was much lower than the mean for controls which was 3127.6 ± 585.3 with p < 0.00002. As expected, C-peptide in the diabetic patients was elevated compared with that of healthy subjects (1913.2 ± 170.6 vs. 946.7 ± 86.5 pmol/L, p < 0.000002). In the diabetic group, 53 individuals (i.e. 77%) had C-terminal AdipoR1 fragment concentrations below the detection limit of 100 pg/mL. Interestingly, 21 of the normal subjects (i.e. 28%) had undetectable AdipoR1 in plasma.
Conclusions: These results suggest that testing for the fragments of this surface membrane-bound receptor protein in blood is potentially a novel approach for diagnosis and or prognosis of diabetes. Longitudinal studies are needed to determine the potential value of this peptide as an early marker for the diagnosis of diabetes.
Characterization of active site on urine trypsin inhibitors
Ma R1, Basu M2, Zmolek KC2, Blatt JR2, Basu SC2, Pugia MJ1,2
1Siemens Healthcare Diagnostics, Urinalysis Research and Development, Elkhart, IN; 2University of Notre Dame, Department of Chemistry and Biochemistry, Notre Dame, IN
Urinary trypsin inhibitors (uTi) are small proteins serving as inhibitors of serine proteases that can be detected in human urine during inflammation. Bikunin is one of the key forms of uTi with two aprotinin-like domains and O-linked and N-linked polysaccharides. Uristatins are small derivatives from Bikunin lacking the O-linked glycoconjugates. Monoclonal antibodies (MAb 421-3G5 and MAb 420 5D11) were raised against uristatin antigen, and they were found to bind an epitope of about 3.86KDa peptide fragment on uristatin without significant cross-reactivity to pro-inhibitors of uTi or aprotinin. Five peptide fragments (A-E from the N-terminus to C-terminus) representing partial sequences of bikunin were chemically synthesized and tested for the ability to inhibit trypsin. The results suggested a secondary structure on the molecular of Bikunin is needed for trypsin inhibition. The MAb 421-3G5 was only able to bind fragment B, which is the region for N-linked polysaccharide. The affinity is much lower than the purified Bikunin standard that indicates this MAb recognize both the fragment B amino acids and N-linked sugars. The affinity of MAb 421-3G5 for purified Bikunin standard was compared to native Bikunin in diabetic urine specimens. The affinity was reduced for many diabetics urine without dilution. Diluted urine eliminated the matrix affect that suggests the existence of associating molecules on the N-linked glycoconjugate region of Bikunin in some patients. MAb 421-3G5 also prevented the inhibition of trypsin from Bikunin, and fragment B partially reduced this effect. All the information suggested the active site of Bikunin is located in the region of fragment B, and the glycoconjugated secondary structure is needed for trypsin inhibition.
Multiplex immunoassay for apolipoproteins: adaptation of a Luminex-based commercial kit for use on dried whole blood spots
Snozek CLH, Hartman SJ, Jaffe AS, Baudhuin LM, McConnell JP
Mayo Clinic, Department of Laboratory Medicine and Pathology, Rochester MN
Familial hypercholesterolemia (FH) is a relatively prevalent hereditary disorder, leading to early-onset atherosclerotic disease, and associated with markedly elevated serum low-density lipoprotein cholesterol and apolipoprotein (Apo) B concentrations. Pediatric screening for FH has been proposed to enhance detection of both children and adults, but questions remain as to the clinical specificity and sensitivity of such an approach. In an attempt to improve diagnostic efficiency, we are developing a tiered assay strategy for analysis of dried whole blood spots collected during routine newborn screening. One tier involves immunoassay for Apo B, AI, and E, using a multiplex commercial kit (LincoPLEX, Millipore) on the Luminex 100 IS platform.
Although the Apo assay is a commercially available kit with adequate analytical sensitivity for this strategy, the protocol had to be modified to suit bloodspot testing. Apo AI, B, and E are eluted from 1/8” discs using 50 uL of the kit Assay Buffer; addition of 0.05% Tween-20 enhances recovery. Ten-fold dilution reduces potential interferences while retaining sufficient Apo concentrations for measurement. The largest departure from the manufacturer’s protocol was in tailoring the calibration curve to suit Apo levels in bloodspot samples, while maintaining enough flexibility to accommodate lot-to-lot variability in the calibration standard. Our modified calibration curve is a 6-point curve that covers a narrower concentration range than the manufacturer’s recommendation, using 3-fold serial dilution rather than the original 5-fold dilutions. Inter- and intra-assay coefficients of variation are <20% for all three Apo analytes. Steps to improve the utility of this assay for newborn screening include using the ratio of Apo B to AI as a control against variations in elution efficiency, and measuring Apo E (a marker of very low-density lipoproteins) to improve specificity for detection of FH.
Clinical validation is planned for summer 2008. This will include assessment of at least 500 newborn bloodspots for levels of total cholesterol and Apo B, AI, and E. Confirmation of high samples will consist of sequencing the low-density lipoprotein receptor gene, mutation of which is responsible for FH. This strategy unites multiple analytical methods to improve the efficiency of screening for FH; early detection of FH probands is expected to enhance diagnosis and treatment of newborns and affected relatives alike.
Targeted quantitation of biomarkers in plasma using peptide MRM-based assays and novel chemistry
Hunter CL, Minkoff M, Nuwaysir LM
Applied Biosystems, Foster City, CA
The utility of multiple reaction monitoring (MRM) on triple quadrupole based MS systems for biomarker verification/validation studies is currently an active area of investigation, driven by the well known sensitivity and selectivity attributes of this type of MS approach. The challenges of assay development and running large scale studies are becoming better understood and strategies for overcoming these challenges are becoming critical. The need for rapid assay development, the need for higher multiplexing and the need for more robust assays are some of the key challenges.
In this work, the unique combination of triple quadrupole and ion trapping capabilities of the hybrid triple quadrupole - linear ion trap mass spectrometer (QTRAP® System) has been utilized to create >1000 high quality, specific MRM transitions for multiple peptides to many plasma proteins. Iterative analysis provided rapid refinement of MRM parameters without requiring synthetic peptides. Intelligent use of retention time using new acquisition software enables many more MRM transitions to be included in a single acquisition method, while maintaining good peak area reproducibility. Finally, a chemical labelling strategy has been employed to create global reference standards to enable quantitative comparisons between many clinical samples.
The analytical reproducibility of the MRM method developed was found to be extremely high in depleted plasma, with the majority of peptides being measured with %CV<10. This method is currently being applied to a clinical sample set to assess the feasibility of performing a large scale study using this methodology.
A real time interaction approach to immunoassay development and evaluation of novel biomarker parameters
Houle JF, Smith PT
Axela, Inc., Toronto, Ontario
A novel approach to employing real time interaction information to develop and implement biomarker assays directly in complex media has been implemented using diffractive optics technology (dotTM). Studies were performed to demonstrate relative kinetic measurements, antibody ranking and pairing, buffer optimization, cross reactivity studies and routine immunoassay directly in complex matrices. Use of an open format design allowed multiplex measurement across broad dynamic range. (>6 logs) This is critical in development of assays where both high and low abundance markers are required to provide adequate discrimination. The simultaneous measurement of the cardiac markers CRP and NT pro-BNP is presented as an example. Data on development of a rapid routine assay for Neuron Specific Enolase, a marker for traumatic brain injury is used to illustrate quantitative immunoassay.
Real time binding data also allows access to potential new diagnostic targets including protein complexes and low affinity antibodies to tumor antigens. The direct measurement of the cardiac Troponin complex and its subcomponents is illustrated in acute myocardial infarction patient samples. The ability to detect autoantibodies to traditional cancer markers directly in patients’ sera enables the characterization of relative affinity and may offer a novel evaluation parameter in the management of disease. The measurement of autoantibodies against PSA was performed in serum samples from patients with and without prostate cancer as evidenced by corresponding biopsies. This capability to monitor binding in real-time may provide an advantage over end-point assays where intervening wash steps can effectively dissociate low affinity interactions. These types of studies were further extended to monitoring and characterization of immune responses to pathogens, vaccines and therapeutic biologicals. In conclusion, the dot platform enabled assess to analyte characteristics that are inaccessible on traditional immunoassay platforms.
A novel approach for selective purification of transcriptional factor proteins
Steele PS, Linder MW, Valdes Jr. R, Jortani SA
University of Louisville Department of Pathology and Laboratory Medicine, Louisville, KY
Introduction: We have previously used affinity capture mass spectrometry to study the interferon-gamma (INF-γ) mediated binding of nuclear proteins to oligonucleotides within the proximal promoter region of cytochrome P4502E1 (CYP2E1) gene. Our preliminary results using INF-γ stimulated human hepatocellular liver carcinoma (HepG2) cells have indicated that a 42 KDa protein captured by the CYP2E1 annealed DNA sequence was interferon-stimulated gene factor 3 gamma (ISGF3γ). We have also shown that in the presence of competitor sequences during affinity capture, the signal to noise ratio for ISGF3γ was reduced by 2-fold. The term interferon regulatory factor 9 (IRF9) is often used interchangeably with ISGF3γ. However, the latter protein represents a complex which also includes IRF9 and Signal Transducers and Activator of Transcription (STAT) proteins.
Objective: To further support the preliminary identification of the captured protein by affinity mass spectrometry using a novel approach supported by standard techniques.
Methods: We initially isolated biotinylated oligonucleotides and their captured proteins using Dynabeads® MyOne Strepavidin C1 magnetic beads (Invitrogen, Oslo Norway). Subsequently, we removed sequence specific proteins bound to the biotinylated oligonucleotide/strepavidin complex by restriction digest (BST II, New England Biolabs, Ipswich, MA). Finally, the isolated protein was identified by western blot using a 1:250 dilution of an ISGF3 γ antibody (BD Biosciences Pharminogen, Rockville, MD).
Results: