O-Glycoside Biomarker of Apolipoprotein C3: Responsiveness to

Dec 5, 2008 - Glyco-isoform ratios were sensitive to liver diseases such as chronic hepatitis C and alcoholic liver cirrhosis. The correlation coeffic...
0 downloads 0 Views 656KB Size
Download PDF
O-Glycoside Biomarker of Apolipoprotein C3: Responsiveness to Obesity, Bariatric Surgery, and Therapy with Metformin, to Chronic or Severe Liver Disease and to Mortality in Severe Sepsis and Graft vs Host Disease Stephen B. Harvey,† Yan Zhang,† Joshua Wilson-Grady,† Teresa Monkkonen,† Gary L. Nelsestuen,*,† Raj S. Kasthuri,‡ Michael R. Verneris,§ Troy C. Lund,§ E. Wesley Ely,|,⊥ Gordon R. Bernard,|,⊥ Harald Zeisler,# Monika Homoncik,∇ Bernd Jilma,O Therese Swan,[ and Todd A. Kellogg[ Department of Biochemistry, Molecular Biology and Biophysics, University of Minnesota, Minneapolis, Minnesota, Department of Medicine, University of Minnesota, Minneapolis, Minnesota, Department of Pediatrics, University of Minnesota, Minneapolis, Minnesota, Department of Pulmonary and Critical Care, Center for Health Services Research, Vanderbilt University Medical Center, Nashville, Tennessee, VA Geriatric Research Education and Clinical Center (GRECC), Vanderbilt University Medical Center, Nashville, Tennessee, Department of Obstetrics and Gynecology, Medical University of Vienna, Vienna, Austria, Department of Internal Medicine IV, Division of Gastroenterology, Medical University of Vienna, Vienna, Austria, Department of Clinical Pharmacology, Medical University of Vienna, Vienna, Austria, and Department of Surgery, University of Minnesota, Minneapolis, Minnesota Received September 8, 2008

The glyco-isoforms of intact apolipoprotein C3 (ApoC3) were used to probe glycomic changes associated with obesity and recovery following bariatric surgery, liver diseases such as chronic hepatitis C (CHC) and alcoholic liver cirrhosis, as well as severe, multiorgan diseases such as sepsis and graft vs host disease (GVHD). ApoC3 glyco-isoform ratios responded to unique stimuli that did not correlate with serum lipids or with other blood components measured in either a control population or a group of extremely obese individuals. However, glycoisoform ratios correlated with obesity with a 1.8-fold change among subjects eligible for bariatric surgery relative to a nonobese control population. Bariatric surgery resulted in rapid change of isoform distribution to that of nonobese individuals, after which the distribution was stable in each individual. Although multiple simultaneous factors complicated effector attribution, the isoform ratios of very obese individuals were nearly normal for diabetic individuals on metformin therapy. Glyco-isoform ratios were sensitive to liver diseases such as chronic hepatitis C and alcoholic liver cirrhosis. The correlation coefficient with fibrosis was superior to that of current assays of serum enzyme levels. Diseases of pregnancy that can result in liver damage, HELLP syndrome and pre-eclampsia, did not alter ApoC3 glyco-isoform ratios. Early after umbilical cord blood transplantation the isoform ratios changed and returned to normal in long-term survivors. Larger changes were observed in persons who died. GVHD had little effect. Persons with severe sepsis showed altered ratios. Similar cut-points for mortality (3.5-fold difference from controls) were found for UCBT and sepsis. Similar values characterized liver cirrhosis. Overall, while changes of glyco-isoform ratios occurred in many situations, individual stability of isoform distribution was evident and large changes were limited to high-level disease. If ratio changes associated with obesity are found to document a risk factor for long-term outcomes, the information provided by glyco-isoform ratio changes may provide important, novel information for diagnostic, prognostic and therapy response to metabolic conditions. Keywords: glycomic isoform biomarkers • bariatric surgery • sepsis • hepatitis C • umbilical cord blood transplant • liver cirrhosis • obesity

The number and heterogeneity of protein products led to early expectations that the plasma proteome would contain many disease-specific protein biomarkers. This was overly optimistic as initially stated and no new clinical assays have arisen from the modern field of proteomics.1 However, much information about * To whom correspondence should be addressed. Professor Gary L. Nelsestuen, 6-155 Jackson Hall, 321 Church St. SE, Minneapolis, MN 55455. E-mail: [email protected]. Fax: (612) 625-2163. † Department of Biochemistry, Molecular Biology and Biophysics, University of Minnesota. ‡ Department of Medicine, University of Minnesota. Current address: Department of Medicine, University of North Carolina, Chapel Hill, North Carolina. § Department of Pediatrics, University of Minnesota. | Center for Health Services Research, Vanderbilt University Medical Center. ⊥ GRECC, Vanderbilt University Medical Center. # Department of Obstetrics and Gynecology, Medical University of Vienna. ∇ Division of Gastroenterology, Medical University of Vienna. O Department of Clinical Pharmacology, Medical University of Vienna. [ Department of Surgery, University of Minnesota. 10.1021/pr800751x CCC: $40.75

 2009 American Chemical Society

the proteome remains unknown. The complexity of the plasma glycoproteome includes the well-known heterogeneity of the carbohydrate moiety. Biosynthesis includes a chain of events that involves carbohydrate processing in the Golgi. Disease states may selectively alter any stage of biosynthesis and export into the plasma, generating an altered glyco-isoform distribution. Current expectations are high for disease biomarkers among plasma glycan structures, either as the carbohydrate chain released from a protein or as an intact glycopeptide. Recent examples include biomarkers for cancer (e.g., refs 2-9) or liver disease.10 There are many challenges to study of glycoproteins and much effort is focused on developing methods for global analysis of the serum glycome (e.g., refs 11-17). An attractive approach to study glycoprotein isoforms is analysis of intact proteins that have been subjected to minimal manipulation. One of the first examples consisted of transferrin and its relationship to inborn errors of glycoprotein biosynthesis or liver disease.18,19 Current technology does not provide for global Journal of Proteome Research 2009, 8, 603–612 603 Published on Web 12/05/2008

research articles analysis of intact proteins in complex mixtures, specific target proteins such as transferrin are pursued with increasing sophistication.20 Of the many glycoproteins in blood plasma, intact apolipoprotein C3 (ApoC3) is unusually accessible. It can be analyzed in the MALDI-TOF mass spectrometer as the intact protein after a one-stage extraction of whole plasma or serum.21 The O-glycan of ApoC3 offers isoforms ranging from the underivatized polypeptide (ApoC3-0) to structures with one (ApoC3-1), two (ApoC3-2) or three sialic acids per chain.22 C-terminal truncation also occurs in vivo, increasing the number of isoforms.21,22 ApoC3 glyco-isoform ratio changes have been observed in uremia23 and kidney disease.24 The absolute concentration of ApoC3 is related to blood lipid content25 and is an independent risk factor for cardiovascular disease.26 In contrast, the glycoisoform ratio did not correlate with fasting triglycerides.22,24,27 Other studies suggested selectivity of ApoC3-2 for small LDL particles.28 Finally, the shared property of ApoC3 glycosylation in the Golgi may link ApoC3 glycoisoform responsiveness to at least some characteristics of a number of other serum glycoproteins. This report presents studies of ApoC3 glyco-isoform distribution in a range of medical conditions, from obesity and bariatric surgery, to liver disease and severe, multiorgan disease states caused by sepsis or graft vs host disease. The results show that changes of glyco-isoform ratios occurred and provided biomarkers for many of these conditions with potential application, especially in metabolic conditions related to obesity.

Experimental Section All samples used in this study were obtained with approval of the Internal Review Boards of the respective institutions and after receiving signed, informed consent from the patients. Samples of controls, bariatric surgery, and UCBT patients were obtained at the University of Minnesota. Samples from chronic hepatitis C, HELLP, pre-eclampsia, and alcoholic liver cirrhosis were obtained at the University of Vienna and severe sepsis at Vanderbilt University. Samples from Patients with CHC and Liver Cirrhosis. Samples from hepatitis C patients were from a prior study.29 The study protocol was approved by the ethics committee of the Medical University of Vienna. Written informed consent was obtained from 40 subjects before initiating the study. Patients with chronic hepatitis C (HCV-RNA positive by PCR and with raised transaminases; Interferon (IFN)-naive or nonresponders/patients with relapse following previous therapies) considered for antiviral combination therapy were asked to participate. Hepatitis C viral load was determined by Cobas Amplicor HCV Monitor version 2.0 (Roche Diagnostics, Branchburg, NJ). Inclusion criteria were age 19 years or older and chronic hepatitis C infection. Exclusion criteria were Child B or C cirrhosis, a history of thromboembolic or severe cardiovascular disease, hypertension not responding to antihypertensive treatment, history of convulsion, participation in a clinical trial in the 3 weeks preceding the study, INF-alpha therapy within 3 months of inclusion into the study, history of variceal bleeding or evidence of gastrointestinal blood loss, hepatocellular carcinoma or other currently existing malignancy and past history of chemotherapy. A second group of 7 individuals were followed for 48 or 72 weeks. These patients received 180 µg of PEG-IFN-a2a (Pegasys; Roche) once weekly and 1000-1200 mg ribavirin (Copegus; Roche) daily for 48 weeks. Serum alanine amino transferase (ALAT) and aspartate 604

Journal of Proteome Research • Vol. 8, No. 2, 2009

Harvey et al. amino transferase (ASAT) were measured by standard methods (29). The study included 20 females, age 51.3 ( 12.3 and 27 males aged 49.4 ( 10.0. Samples from patients with alcoholic liver cirrhosis were from a different study30 also approved by the ethics committee of the Medical University of Vienna. Samples were obtained with informed patient consent. Samples from patients with liver cirrhosis were characterized for disease stage using the CHILD Pugh Score. Accordingly, 22.7% of patients were classified as grade A, 54.6% as B and 22.7% as the most severe grade C. Of 21 patients 30% were female (age 54.7 ( 8.9) and 70% were male (age 60.0 ( 12.1). Bariatric Surgery Patients. Persons eligible for weight loss surgery by NIH guidelines (BMI g 35 kg/m2 with at least one serious comorbidity, or BMI g 40 kg/m2 without any comorbidity) at the University of Minnesota Medical Center - Fairview Weight Loss Center were asked to participate in a prospective study of proteome changes associated with surgery. Those who provided informed consent were enrolled. Laboratory analyses before and 90 days following surgery included complete lipid panel (high-density lipoproteins [HDL], low-density lipoproteins [LDL-C], triglycerides, total cholesterol), serum creatinine, kidney glomerular filtration rate (GFR), blood urea nitrogen (BUN), fasting blood glucose, and urinary pH. BMI (mean, 45.7 ( 7.2 kg/m2), gender (17 males, 92 females), and age (mean, 41.8 ( 12.2 years) were also collected at each time point. All correlations found between these measures and the glycomic peak ratio are stated in the text. If not stated, no correlation was detected. Samples were taken 1 day before surgery and 30 days and 90 days following surgery. A total of 107 individuals consented for the presurgery sample and 44 completed the study. Of the 44, one patient underwent adjustable gastric banding and one biliopancreatic diversion with duodenal switch. Post surgery statistics were based on patients undergoing Roux-en-Y gastric bypass only. Samples from Patients with HELLP Syndrome and PreEclampsia. Eight patients with HELLP syndrome (Hemolysis, Elevated Liver enzymes, Low Platelet count) were diagnosed using standard criteria,31 consent was obtained and samples collected at the General Hospital of Vienna, Austria. Twentysix additional persons with pre-eclampsia were also diagnosed by standard criteria,32,33 consent obtained and samples collected at the General Hospital of Vienna, Austria. Samples from Patients with Severe Sepsis. Persons meeting the criteria for severe sepsis were diagnosed at the intensive care facility at Vanderbilt University hospitals from from June 1992 to February 1995. Samples were stored at -80 degrees. Informed consent was obtained from the patient or guardian. Two samples from each of 58 patients (22 female, age 48.9 ( 17.6 and 36 males, age 55.6 ( 16.8) were obtained at a 4-day interval. MALDI-TOF profile results are reported for the second sample. APACHE II (Acute Physiology And Chronic Health Evaluation) scores were assigned from patient data and history by standard criteria described by.34 Patients were followed for at least 100 days from the sample date and mortality was recorded. Umbilical Cord Blood Transplantation (UCBT). Samples from patients (53 female, age 35.2 ( 18.3; 62 males age 32.1 ( 20.9) undergoing allogeneic, UCBT were obtained one week before and weekly for the first 100 days after the procedure. Indications for transplantation varied, but included acute lymphocytic and myeloid leukemia. Patients were conditioned with either myeloablative or reduced intensity conditioning.35,36

research articles

Apolipoprotein C3: A Prototype Glycome Biomarker GVHD prophylaxis was with mycophenolic acid mofetil (until day 28) and cyclosporine A (CSA).35,36 Correlations between the MALDI-TOF analysis and clinical parameters ((GVH stage, and death) were performed using data collected by the UCBT database at the University of Minnesota. Control Subjects. Results for pre- and postbariatric surgery could be compared with several previous results as well as a control group of this study. The latter consisted of 48 adults (22 males and 26 females, average age 43.9 ( 9.7 years, BMI ) 29.2(5.3 kg/m2) from the Midwestern United States who were randomly selected from a larger cohort to serve as controls for studies related to CHC and liver cirrhosis. These individuals had volunteered for other purposes and gave consent for general proteome analysis. Blood lipids were determined but no medical histories collected or medications recorded. The lipid measurements and averages: serum triglyceride ) 113 ( 76 mg/dL, Cholesterol ) 138 ( 49 mg/dL, and HDL-C 51.8 ( 28.7 mg/dL. A subgroup of the larger control group consisted of individuals with BMI 150 mg/dL (n ) 58, ApoC3-1/ApoC3-2 ratio ) 4.94 ( 2.32) were indistinguishable from all other presurgery subjects (4.94 ( 1.99, p ) 0.99). The same applied to individuals with cholesterol >200 mg/dL (n ) 58, ApoC3-1/ApoC3-2 ) 4.57 ( 1.79) when compared to all others (ratio ) 5.37 ( 2.49, p ) 0.06). Finally, those with any form of dyslipidemia (Triglycerides > 150 mg/dL, LDL > 130 mg/dL, HDL < 40 mg/dL, total cholesterol >200 mg/dL) showed no difference of isoform peak ratios (average ) 4.69 ( 1.67) from the remaining prebariatric surgery subjects (n ) 16, average ) 4.97 ( 2.26, p ) 0.64) who had normal blood lipid levels. In contrast to blood lipids, BMI gave a positive correlation with ApoC3 isoform ratio. A significant correlation (p ) 0.015) was found between the ApoC3-1/ApoC3-2 ratio and BMI for combined controls and prebariatric surgery patients (Figure 2C). The exceptions to the correlation were individuals receiving metformin therapy (solid triangles, Figure 2C). Influence of Bariatric Surgery. Bariatric surgery is an emerging therapy for extreme obesity and its comorbidities.

Apolipoprotein C3: A Prototype Glycome Biomarker

Figure 2. Correlation of ApoC3 concentration and ApoC3-1/ ApoC3-2 ratios with serum triglycerides and BMI. (A) ApoC3 concentration as a function of fasting triglyceride (R ) 0.55, p ) 0.002). (B) ApoC3-1/ApoC3-2 ratio as a function of fasting triglycerides for the control group (solid diamonds, R ) 0.25, p ) 0.48), for prebariatric surgery patients (open diamonds) with correlation coefficients for the combined group calculated (n total ) 149, R ) 0.087, p ) 0.48). (C) ApoC3-1/ApoC3-2 ratio as a function of BMI for the control group (solid diamonds, R ) 0.39, p ) 0.12) and prebariatric surgery subjects who were not on metformin therapy (open diamonds). The correlation coefficient for the combined groups was R ) 0.43 (p ) 0.015). Subjects receiving metformin therapy are shown by solid triangles with red fill. Three diabetic subjects who did not receive metformin (circles with blue fill) are also shown. The circle at BMI ) 43.2 represented an individual on Exenatide therapy alone. The circle at BMI ) 54.3 represented an individual receiving no therapy for diabetes and the circle at BMI ) 37.7 represented an individual on pioglitazone therapy.

This procedure brings about radical change in diet and weight over a short period of time. Peak ratios for 107 subjects before and 44 of the same individuals after bariatric surgery were determined with representative results shown in Figure 3A. Forty-one of 44 subjects showed a decline of 15% to 80% in the ApoC3-1/ApoC3-2 ratio following surgery (5.03 ( 2.26 to 2.56 ( 0.89 at 90 days, Figure 3A, p ) 1*10-10). Only two were unchanged, one received metformin therapy before surgery. Change was nearly complete by 30 days with only 15% decline between 30 and 90 days (not significant). While the range of values for the postsurgery population remained quite large, individual variation over time was small. The average and standard deviation for the peak ratio at 90 days divided by the value at 30 days for each person was 0.84 ( 0.19. The resulting coefficient of variance was virtually identical to that reported for 30 adults (described in Experimental section) who were measured at a 3-year interval (ratio of final to initial value ) 1.02 ( 0.29, see Table 5 of reference 21). Although small, individual variation appeared real; replicate measurements of the same sample gave a coefficient of variance of about 7% for this peak ratio.21 Following surgery, the ratio appeared to be fully adjusted at 90 days despite an average BMI that remained at 38.4 ( 6.7 (Figure 3B). The final average and standard deviation were virtually identical to that reported for nonobese adolescents (data from (21)), for the subset of control subjects from this

research articles

Figure 3. Influence of bariatric surgery on ApoC3-1/ApoC3-2 ratio. (A) Individual responses to bariatric surgery. Results for 11 of the 44 individuals who completed the program are shown. The average and standard deviation for all individuals before surgery (n ) 106, 5.03 ( 2.26, solid square at -5 days) and at 90 days following surgery (n ) 44, solid square at 95 days, average ) 2.64 ( 0.95, p ) 2.3*10-8) are shown. The average and standard deviation for two peak ratios among control groups include the 9 nonobese adolescents at age 13 and 19 (see Experimental section and data from (21), solid diamond) and the nonobese adults from the control population with BMI < 26.0 (average ) 22.8 ( 1.0, solid circle). (B) ApoC3-1/ApoC3-2 of the 44 subjects at 90 days postbariatric surgery.

study who had BMI of less than 26 (compare solid square with triangle and circle on the right side of Figure 3A) or for a group of 18 volunteers for a clinical study of endotoxemia (average ratio ) 2.78 ( 1.00, individual data from Figure 6A of ref 21). In fact, the average and standard deviation was extremely constant across all ages and genders for groups of individuals who were nonobese. The consistency over time suggested that each individual maintains the same isoform ratio as long as health status is constant. No significant correlations were found between the ApoC3-1/ ApoC3-2 ratio and presurgery subjects for the following comorbidities: hypothyoidism (n ) 4), hypertension (n ) 39), obstructive sleep apnea (n ) 55), asthma (n ) 21), GERD (n ) 41). No correlation was observed with serum creatinine, blood urea nitrogen or glomerular filtration rate (not shown). A significant ratio was found for those on statin-type drugs (n ) 9, ApoC3-1/ApoC3-2 ratio ) 3.18 ( 1.43) versus all other presurgery individuals (average ) 5.10 ( 2.67, p ) 0.011). The basis for this effect was not clear however, as 5 of the 9 individuals were also on metformin or pioglitazone therapy (see below). A significant correlation was found between the ApoC3-1/ ApoC3-2 ratio and diabetes (n ) 16, average with diabetes ) 3.65 ( 1.32 and for nondiabetes ) 5.16 ( 2.21, p ) 0.011). Thirteen of the diabetic subjects were receiving metformin therapy. Averages for those on metformin therapy (n ) 14, including one who did not have diabetes) gave slightly better statistics than the diabetic group as a whole (average plus metformin ) 3.43 ( 1.02, all others ) 5.16 ( 2.21, p ) 0.005, also see solid triangles, Figure 3B). BMI did not differ between those on metformin and others (45.0 ( 8.5 vs 45.7 ( 7.1, respectively). Apparent correction of the ApoC3 isoform ratio Journal of Proteome Research • Vol. 8, No. 2, 2009 607

research articles

Harvey et al.

appeared to be independent of these other comorbidities. Correction of glycoisoform ratio appeared independent of other blood properties; fasted triglyceride levels among those receiving metformin were higher than controls (222 ( 124 vs 160 ( 76, respectively, p ) 0.009) as was blood glucose (average ) 142 ( 80 mg/dL vs 93 ( 12 for controls, respectively, p ) 1*10-7). Of the 14 subjects on metformin, seven received other therapies for diabetes. Very few diabetic individuals received single drug therapies other than metformin. One individual on pioglitazone alone had a very low ratio of ApoC3-1/ApoC3-2 (2.1) while another individual on Exenatide alone had a ratio of 5.89 (open circle at BMI ) 43.2, Figure 2C), higher than any of the individuals who received metformin. The individual who received no medical intervention for diabetes had a peak ratio of 5.91 (open circle at BMI ) 54.5, Figure 2C). In fact, those on metformin therapy included some with the lowest ratios among all subjects, including nonobese controls. ApoC3 Glyco-Isoform Ratios in Liver Disease. ApoC3 is synthesized in the liver and its glyco-isoform ratios should be sensitive to diseases that alter any aspect of protein synthesis or export from the liver. In fact, the ApoC3-2/ApoC3-1 ratio was increased by liver disease. The ratio used to demonstrate this property was ApoC3-2/ApoC3-1, the inverse of the relationship used for obesity and triglycerides. This allowed presentation of the results as a positive correlation. Subjects with Chronic Hepatitis C showed reduced levels of ApoC3 as detected by ELISA (Figure 4A, average ) 32 ( 27 µg/ mL vs 65.0 ( 43 for controls, p ) 1.8*10-4). This was expected from lowered serum albumin associated liver cirrhosis that may represent lowered protein synthesis.39 They also had an increased proportion of ApoC3-2 (Figure 4B, average ApoC3-2/ ApoC3-1 ratio for CHC ) 0.79 ( 0.53, controls ) 0.35 ( 0.12, p ) 1.7*10-7). Individually, the level of protein and the peak ratio showed substantial overlap at the 95% confidence limit of the control population. The sum of these terms improved separation from the controls (Figure 4C, p ) 2*10-9). A significant correlation between the ApoC3-2/ApoC3-1 ratio and fibrosis level was found (Figure 4F). This was improved by combination of concentration and peak ratio (data from Figure 4C vs fibrosis, R ) 0.78, p ) 0.002). Separation of all CHC individuals from the 95% confidence limit for controls was not achieved until fibrosis level 4 and more advanced liver cirrhosis (Figure 4F). It was interesting that ratios of 1.25 were not observed until fibrosis level 4, early stage liver cirrhosis. This ratio was a significant cutpoint for mortality in other disease states (see below). Correlation of ApoC3-2/APoC3-1 peak ratio continued among those with liver cirrhosis. As expected, glyco-isoform ratios at CHILD Pugh score A (average peak ratio ) 1.24 ( 0.18) were indistinguishable from those at fibrosis level 4 (Figure 4F). Higher average ratios occurred at CHILD Pugh scores B (1.53 ( 0.48) and C (1.80 ( 0.46). The correlation coefficient for CHILD Pugh score vs ApoC3-2/ApoC3-1 ratio was significant (R ) 0.43, p ) 0.044). Liver damage is commonly determined by serum concentrations of alanine amino transferase (ALT) and aspartate amino transferase (AST). The sum of these enzyme concentrations is shown in Figure 4C. There was 90% separation (26 of 29) from controls at the 95% confidence level (Figure 4C). Despite excellent separation of disease from controls, the correlation coefficient for enzyme level and fibrosis score was not significant (R ) 0.36, p ) 0.49, Figure 4G). 608

Journal of Proteome Research • Vol. 8, No. 2, 2009

Figure 4. Diagnosis of chronic hepatitis C. (A) Relative ApoC3 concentration. The concentration of ApoC3 in CHC patients (open diamonds) is shown relative to the average and 2*SD for the control population described in Figure 1A (Solid Square, control average ) 67.4 µg/mL). A relative concentration of 0.1 was the detection limit of the assay. (B) ApoC3-2/ApoC3-1 Peak ratios. The ApoC3-2/ApoC3-1 ratio for the same individuals (open triangles) is shown relative to the average and 2*SD for the control population (solid square, control average ) 2.83). (C) Sum of ApoC3 concentration and isoform ratio. The quantity (Part B plus 1/Part A)/2 is plotted (open circles) relative to the values for the control population ( 2*SD (solid square). (D) Sum of ALT plus AST serum enzyme activities. The sum is shown for the CHC subjects (X) relative to the average and 2*SD for a control population (solid square, control averages taken from (46), control average ) 44.0). (E) Sum of all measures. The quantity (Part C + Part D)/2 is shown (+) relative to controls (solid square). The error bars represent standard deviations for parts C and D summed by standard methods. (F) Correlation of ApoC3-2/ ApoC3-1 ratio with fibrosis (R ) 0.55, p ) 0.002). The value shown at 0 is the average and standard deviation for the control population. The value shown at 5 is the average and standard deviation of 21 subjects with alcoholic liver cirrhosis. (G) Enzyme activities for CHC subjects (R ) 0.36, p ) 0.49). The value at 0 fibrosis is the average for the control population ((SD).

Combination of all measurements: ApoC3 concentration, ApoC3-2/ApoC3-1 ratio and enzyme levels resulted in an average of the different assays (Figure 4E). The combination did not separate CHC from controls better than enzyme levels and the correlation with fibrosis score was slightly less good than that of ApoC3-2/ApoC3-1 peak ratio alone (R ) 0.69, p ) 0.0047). ApoC3 Isoform Ratios in Pre-Eclampsia and HELLP Syndrome. The HELLP syndrome presents the potential for early detection of liver disease. Diagnosis of HELLP can be difficult, leading to liver damage before proper intervention is initiated.31 Eight subjects with HELLP syndrome gave ApoC3-2/ApoC3-1 ratios of 0.32 ( 0.07, virtually identical to the average for all control populations. The peak ratio among twenty-six persons with pre-eclampsia (0.36 ( 0.19) was also indistinguishable from controls. ApoC3 Isoform Ratios in Severe Sepsis. A prior study showed a correlation between ApoC3 isoform ratios and mortality in severe sepsis.40 These results are shown in Figure

research articles

Apolipoprotein C3: A Prototype Glycome Biomarker

Figure 5. Peak ratios versus APACHE II scores for severe sepsis. ApoC3-2/ApoC3-1 (open diamonds, A) and APACHE II scores times 0.1 (open squares, B) are shown along with the value: (part A + Part B)/2 (open triangles, C). The multiplier served to equalize the impact of peak ratio and APACHE II score on the overall value. The lines drawn represent the optimum cut-point for separation of mortality within 30 days versus long-term survival.

5 along with the current prognostic scale for patients with severe sepsis, the APACHE II score. The optimum cut-point for prognosis of survival/death from the ApoC3-2/ApoC3-1 ratio was 1.25 (Figure 5A). At this ratio, false negatives (4 of 19) gave a sensitivity of 79% while false positives (13 of 39) gave a specificity of 67%. The difference was highly significant (p ) 0.0056). The APACHE II score (times 0.1) cutpoint of 1.6 gave 7 false negatives (out of 19, 63% specificity) and 6 false positives (out of 39, 85% specificity) and had higher confidence than the apoC3 isoform ratio (p ) 9*10-5, Figure 5B). Combination of ApoC3 peak ratios with APACHE II scores gave the same sensitivity and specificity as APACHE II scores alone (Figure 5C, p ) 4.6*10-5). ApoC3 in UCBT. Allogeneic hematopoietic cell transplantation is challenged by graft versus host disease (GVHD) where the donor immune system recognizes the host as foreign and mounts an immune-mediated attack toward the host. A small retrospective group of patients who developed GVHD showed altered glyco-isoforms of ApoC3 as previously described.41 Therefore, a larger prospective study was established to discover the potential value of this biomarker for prediction or prognosis of GVHD and/or patient survival. Figure 6 shows analysis of 118 patients from whom blood was obtained on a weekly schedule from days -7 to +100 relative to UCB infusion. The average number of samples per person was 7.2 with a range from 3 to 15. The samples were distributed approximately equally along the time intervals shown in Figures 6A and B. The average and standard deviation for the ApoC3-2/ApoC3-1 ratios obtained in the individual time windows are shown at the midpoint of each time interval (10-day intervals from -10 to +30 days, 20 day interval from 30 to 50 and 25 day intervals thereafter). Average ratios of ApoC3-2/ApoC3-1 increased following UCBT. This was expected due to the adverse impact of the conditioning chemotherapy and radiation therapy use prior to infusion of UCB. Results for long-term survivors (n ) 29, Table 1) who did not experience GVHD showed an increase from 0.46 ( 0.22 to 0.76 ( 0.48 at 10-20 days (Figure 6A, significantly different from baseline only at days 10-20 (p ) 0.009) and 20-30 (p ) 0.006)). Only one instance was recorded of a ratio above the optimum cutpoint for mortality in severe sepsis (1.25, Table 1). Over time, averages returned to the pretransplantation levels (0.44 ( 0.11, Figure 6A). Long-term survivors who experienced GVHD gave slightly higher averages but none with significant difference from long-term survivors without GVHD. Six individuals in this group experienced peak ratios greater than 1.25 (Table 1, not significant versus survivors without GVHD).

Figure 6. ApoC3-2/ApoC3-1 ratios following UCBT. (A) Average of long-term survivors (>365 days) followed before and after UCBT at zero time. The average ratio for survivors with no evidence of GVHD is shown (solid diamonds) as well as the average and standard deviation for values from subjects who developed GVHD at some time after transplant (open squares). (B) Peak ratios for those who died within 1 year of UCBT. Average and standard deviations are shown for nonsurvivors who did not develop GVHD (solid diamonds) and those who did develop GVHD (open squares). (C) ApoC3-2/ApoC3-1 ratios within 50 days of death. Each line represents a different individual. The dashed line designates a value of 1.25. Table 1. ApoC3-2/ApoC3-1 Ratios As a Function of UCBT, GVHD, and Death

category

A.GVHD Non-Survivorsb B. GVHD Non-Survivorsb C. GVHD Non-Survivorsb D. Non-Survivors NO GVHDb E. All nonsurvivorsb F. GVHD Survivorsb G. GVHD Survivorsb H. GVHD Survivorsb I. Survivors NO GVHDd

n for n ratio > grade total 1.25a

All 0-1 2-4

All 0-1 2-4

10 2 8 10 20 34 6 28 29

6 1 5 6 12 6 1 5 1

p (categories compared)c

0.008 (A vs F) 0.009 (C vs F) 0.008 (D vs F) 1.25 within the time window defined by footnote b. b Includes those with samples taken (30 days of GVHD and/or 30 days before death. c Chi squared value for the proportion of individuals with ratio >1.25 within the time window defined in footnotes a and b. d Includes samples taken (30 days of transplant.

Pretransplant values for persons who died of any cause (0.46 ( 0.18) did not differ from those of long-term survivors (Figure 6B). However, isoform ratios increased earlier and for longer duration after transplant. They also reached higher values. The averages in the first three intervals following UCBT were 0.83 Journal of Proteome Research • Vol. 8, No. 2, 2009 609

research articles -4

( 0.58 (p ) 6*10 relative to long-term survivors without GVHD), 1.31 ( 0.94 (p ) 0.017 relative to long-term survivors without GVHD) and 1.58 ( 1.49 (p ) 1.4*10-4 relative to longterm survivors without GVHD). Of those who died, 5 of the 8 who experienced GVHD grades 2-4 and who had samples taken within 30 days of death showed ratios above 1.25 (not significant relative to deaths without GVHD, Table 1). Identical proportions of persons with and without GVHD who died also developed ratios above 1.25 (Table 1). In fact, GVHD did not produce a significant impact on the proportion of individuals with ratios greater than 1.25, except when associated with long-term survival vs death. Of 19 individuals who displayed a peak ratio greater than 1.25 at some time during the study, 7 survived and 12 died. These proportions were similar to those for sepsis where a ratio above 1.25 corresponded to a similar death rate (Figure 5). Often, elevated ratios of ApoC3-2/ApoC3-1 occurred only at short times before death (Figure 6C). For individuals who died long after the study was completed, samples were not available within 30 days of death. Of 21 individuals who fit this description, 6 displayed ratios greater than 1.25 long before death. The high ratios occurred at death minus 81, 97, 113, 122, 190 and 248 days. This suggested some ability for long-term prognosis.

Discussion Glycoprotein biosynthesis depends on polypeptide synthesis and trafficking through the Golgi apparatus. The biological function of the heterogeneous glycan product is not well understood. Incomplete structures occur at all levels from addition of the first sugar(s) to the addition of “extra” sialic acids (see Figure 1A and ref 18 for examples). Glycomic structures and distributions may provide insight into disease mechanisms that impact on cellular events leading to protein export. This study found a novel correlation of ApoC3 isoform distribution with obesity and showed rapid correction following bariatric surgery. Few other correlations were found with the exception of nearly normal isoform ratios in presurgery subjects who received metformin therapy. The corrective agent was not clear however, as metformin therapy included nearly all of the individuals with diagnosed diabetes. Thus, significant correction toward normal isoform distribution could arise from diabetes or from the therapeutic agent. The same matter arose for apparent glycoisoform correction by statin drugs where the largest effect occurred in those also receiving metformin therapy. We suggest that the most likely effective agent was metformin or other drugs rather than diabetes itself. Metformin has a cellular effect, increasing responsiveness to insulin. It is sometimes applied to prediabetic individuals to aid in weight loss and lower the risk of developing diabetes (reviewed in 42). It is possible that the ApoC3 isoform ratio reported a less wellunderstood effect of this drug on general cellular metabolism. If correct, this ratio may offer insight into mechanism of metformin or other drug action and may constitute a useful marker to determine responsiveness of an individual to therapy. The numbers of individuals on monotherapy with other drugs was small. One subject on exenatide therapy alone did not have corrected isoform ratios while another on pioglitazone therapy alone gave a fully corrected ratio. These drugs have divergent mechanisms of action. Future studies should be designed to determine whether the correction of the ApoC3-1/ ApoC3-2 ratio is drug-specific. 610

Journal of Proteome Research • Vol. 8, No. 2, 2009

Harvey et al. Aberrant ApoC3-1/ApoC3-2 ratios were corrected by bariatric surgery long before BMI reached the desired level. For example, at 90 days postsurgery the average BMI was still 38.4 ( 6.7 kg/m2 but the ApoC3 isoform ratios were characteristic of nonobese controls. This rapid correction of the isoform ratio mimicked the corrective effect that bariatric surgery often has on type 2 diabetes (reviewed in ref 43). In the future, it would be of interest to determine whether extreme ApoC3 isoform ratios can be corrected by means such as diet and exercise. ApoC3-1/ApoC3-2 ratios among nonobese adults and adolescents as well as postbariatric surgery patients were seldom above 4.0 (see Figures 2B and 3C). It is possible that persons with ratios greater than 4.0 had some level of pathology that was not detected by other means. Of course, application of a specific cut-point for diagnosis overlooks individual differences. That is, persons who are normally at one extreme of the control range may experience large changes before they exceed the values of other healthy individuals. Intraperson comparisons over time should provide optimum detection of health change or challenge. In this regard, intraperson comparison before and after bariatric surgery showed 95% detection of a significant change in status. In contrast, attempts to discriminate persons eligible for surgery from a large population that included controls would result in substantial false positives and negatives from some high values among controls and low values among prebariatric surgery patients. Detection of change from longitudinal study of one individual appeared much more specific than categorical reference to a single cutpoint. An important property of the postbariatric surgery subjects was the stability of the ApoC3-1/ApoC3-2 ratio in each individual and an average and standard deviation that equaled that of several control groups of nonobese individuals. This occurred despite enormous differences in diet of postbariatric surgery subjects relative to the nonobese controls. This may suggest that the ratios at 90 days postsurgery represented the genetically- or developmentally determined ratio for each individual that was maintained under a wide array of circumstances. High stability of the glyco-isoform ratios and resistance to change was consistent with observations associated with mortality in severe sepsis and UCBT. Except in cases where these complex multiorgan diseases involved liver damage, the isoform ratio represented a surrogate biomarker of general health. The changes were not large. An ApoC3-2/ApoC3-1 ratio greater than 1.25 (3.5 times the average for the control group) was associated with over 50% mortality in both conditions. Furthermore, a ratio change of 3.5-fold represented only a 2.1-fold change in the percentage of the ApoC3-2 isoform (from 25.5 to 55% of total full length ApoC3-1 plus ApoC3-2). There was almost no response to GVHD unless associated with death. There was also no detected change associated with HELLP or pre-eclmapsia. One conclusion from these collective observations was that each individual had substantial resistance to alteration of his or her proteome. Observation that a two -fold change in a glyco-isoform ratio characterized a neardeath experience may help explain the difficulty in development of new protein biomarker assays. In the case of ApoC3, disruption of biosynthetic function by 2-fold appeared to represent a calamitous event. This challenge may apply to common proteins. It would not apply to proteins from new tissue forms such as cancer cells. Disease type may impact the expectations and approaches to biomarker discovery. These

research articles

Apolipoprotein C3: A Prototype Glycome Biomarker findings also make the 1.8-fold ratio change associated with extreme obesity appear more significant. As expected, ApoC3 isoform ratios changed with degenerative liver diseases such as chronic hepatitis C and alcoholic liver cirrhosis. Isoform ratios attained values greater than 1.25, the cut-point for more than 50% mortality in severe sepsis and UCBT. However, this only occurred at fibrosis score 4 or at more advanced stages of liver cirrhosis. An impact of liver disease on the ApoC3 glyco-isoform distribution agreed qualitatively with plasma glycome changes found in a study of CCl4induced liver damage in the rat.10 Results for the animal model suggested that glycomic changes were more sensitive than serum ALT or AST enzyme levels for detection of liver fibrosis. While the current study did not find greater sensitivity for ApoC3 glyco-isoform changes (Figure 4), the results may be biased by the methods used for selection of the human subjects. That is, serum ALT and AST are important contributors to subject selection, thereby introducing a bias toward individuals with high AST and ALT. Superior correlation coefficients between fibrosis stage and ApoC3 isoform distribution (Figure 4F) suggested that glycoisomer ratios may prove superior to enzyme levels (Figure 4G) in some experimental designs. A critical need for chronic liver disease is prognosis of those who advance to either cirrhosis or liver cancer (reviewed in ref 44). It is possible that factors influencing glyco-isoform ratios will prove superior to other tests in this respect. For example, although average isoform ratios were significantly altered by fibrosis stage 2 (Figure 4B, average for subjects ) 0.59 ( 0.27, average for controls ) 0.353 ( 0.12, p ) 1.4*10-5), 7 of 13 ratios remained within two standard deviations of controls. In fact, complete separation of controls from disease on the basis of glyco-isoform ratios was not achieved until fibrosis stage 4, early stage cirrhosis. It is possible that an aberrant glyco-isoform ratio at fibrosis stage 2 best defined degraded cellular function that was poised for rapid progression to cirrhosis while a normal ratio defined healthier tissue that will progress more slowly. This possibility must be tested by a large prospective study. Overall, this study examined a number of possible roles that a specific glycan isoform may play in diagnosis or prognosis of several disease states. Although a single glycan was studied, several findings may apply to other plasma glycoprotein biomarkers that are synthesized in the liver. One of the strongest conclusions of this study and previous work 21,40,41 was the existence of a personal glyco-isomer distribution for each individual, its stability over time and resistance to change. Substantial liver disease or near-death status was needed to generate ratio changes of 3- or 4-fold from the average for the control population. This agreed with the relative change of ApoC3-2/ApoC3-1 ratio reported in chronic renal failure (1.7fold24). Three-fold represented the approximate range found among the control population. Consequently, individuals with a personal isoform ratio at one extreme of the normal range could experience a 3-fold change before diagnosis by an aberrant ratio. Thus, longitudinal assay of each individual and diagnosis by personal change will provide the most sensitive approach to analysis. Findings related to ApoC3 isoforms and obesity with possible correction by drug therapy or surgery may be of greatest interest. This glycomic biomarker may assist in understanding novel types of pathology that are associated with obesity, the mechanism of certain drug actions and/or mechanisms that might be employed to correct some of the adverse

impacts of extreme obesity. Much further work is needed to determine the types of risk factors or pathology documented by the ApoC3 isoform ratio and the importance of these to longterm health.

Acknowledgment. This work was supported by funds from the endowment to the Samuel Kirkwood Professorship (G.L.N.), a Brainstorm Award from the Cancer Center, University of Minnesota (M.R.V. and G.L.N.). We are indebted to Ms. Lynn Frydrych for assistance in analysis of the MALDI-TOF profiles. References (1) Zolg, W. The proteomic search for diagnostic biomarkers: lost in translation? Mol. Cell. Proteomics 2006, 5 (10), 1720–6. (2) Qiu, Y.; Patwa, T. H.; Xu, L.; Shedden, K.; Misek, D. E.; Tuck, M.; Jin, G.; Ruffin, M. T.; Turgeon, D. K.; Synal, S.; Bresalier, R.; Marcon, N.; Brenner, D. E.; Lubman, D. M. Plasma glycoprotein profiling for colorectal cancer biomarker identification by lectin glycoarray and lectin blot. J. Proteome Res. 2008, 7 (4), 1693–703. (3) Zhao, J.; Patwa, T. H.; Qiu, W.; Shedden, K.; Hinderer, R.; Misek, D. E.; Anderson, M. A.; Simeone, D. M.; Lubman, D. M. Glycoprotein microarrays with multi-lectin detection: unique lectin binding patterns as a tool for classifying normal, chronic pancreatitis and pancreatic cancer sera. J. Proteome Res. 2007, 6 (5), 1864– 74. (4) Zhao, J.; Qiu, W.; Simeone, D. M.; Lubman, D. M. N-linked glycosylation profiling of pancreatic cancer serum using capillary liquid phase separation coupled with mass spectrometric analysis. J. Proteome Res. 2007, 6 (3), 1126–38. (5) Zhao, J.; Simeone, D. M.; Heidt, D.; Anderson, M. A.; Lubman, D. M. Comparative serum glycoproteomics using lectin selected sialic acid glycoproteins with mass spectrometric analysis: application to pancreatic cancer serum. J. Proteome Res. 2006, 5 (7), 1792–802. (6) Kirmiz, C.; Li, B.; An, H. J.; Clowers, B. H.; Chew, H. K.; Lam, K. S.; Ferrige, A.; Alecio, R.; Borowsky, A. D.; Sulaimon, S.; Lebrilla, C. B.; Miyamoto, S. A serum glycomics approach to breast cancer biomarkers. Mol. Cell. Proteomics 2007, 6 (1), 43–55. (7) Kyselova, Z.; Mechref, Y.; Al Bataineh, M. M.; Dobrolecki, L. E.; Hickey, R. J.; Vinson, J.; Sweeney, C. J.; Novotny, M. V. Alterations in the serum glycome due to metastatic prostate cancer. J. Proteome Res. 2007, 6 (5), 1822–32. (8) Leiserowitz, G. S.; Lebrilla, C.; Miyamoto, S.; An, H. J.; Duong, H.; Kirmiz, C.; Li, B.; Liu, H.; Lam, K. S. Glycomics analysis of serum: a potential new biomarker for ovarian cancer? Int. J. Gynecol. Cancer 2008, 18 (3), 470–5. (9) Saldova, R.; Royle, L.; Radcliffe, C. M.; Abd Hamid, U. M.; Evans, R.; Arnold, J. N.; Banks, R. E.; Hutson, R.; Harvey, D. J.; Antrobus, R.; Petrescu, S. M.; Dwek, R. A.; Rudd, P. M. Ovarian cancer is associated with changes in glycosylation in both acute-phase proteins and IgG. Glycobiology 2007, 1712, 1344–56. (10) Desmyter, L.; Fan, Y. D.; Praet, M.; Jaworski, T.; Vervecken, W.; De Hemptinne, B.; Contreras, R.; Chen, C. Rating of CCl(4)-induced rat liver fibrosis by blood serum glycomics. J. Gastroenterol. Hepatol. 2007, 22 (7), 1148–54. (11) Patwa, T. H.; Zhao, J.; Anderson, M. A.; Simeone, D. M.; Lubman, D. M. Screening of glycosylation patterns in serum using natural glycoprotein microarrays and multi-lectin fluorescence detection. Anal. Chem. 2006, 78 (18), 6411–21. (12) Atwood, J. A., 3rd; Cheng, L.; Alvarez-Manilla, G.; Warren, N. L.; York, W. S.; Orlando, R. Quantitation by isobaric labeling: applications to glycomics. J. Proteome Res. 2008, 7 (1), 367–74. (13) Kim, Y. G.; Jang, K. S.; Joo, H. S.; Kim, H. K.; Lee, C. S.; Kim, B. G. Simultaneous profiling of N-glycans and proteins from human serum using a parallel-column system directly coupled to mass spectrometry. J. Chromatogr., B 2007, 850 (1-2), 109–19. (14) Kita, Y.; Miura, Y.; Furukawa, J.; Nakano, M.; Shinohara, Y.; Ohno, M.; Takimoto, A.; Nishimura, S. Quantitative glycomics of human whole serum glycoproteins based on the standardized protocol for liberating N-glycans. Mol. Cell. Proteomics 2007, 6 (8), 1437– 45. (15) Miura, Y.; Hato, M.; Shinohara, Y.; Kuramoto, H.; Furukawa, J.; Kurogochi, M.; Shimaoka, H.; Tada, M.; Nakanishi, K.; Ozaki, M.; Todo, S.; Nishimura, S. BlotGlycoABCTM, an integrated glycoblotting technique for rapid and large scale clinical glycomics. Mol. Cell. Proteomics 2008, 7 (2), 370–7.

Journal of Proteome Research • Vol. 8, No. 2, 2009 611

research articles (16) Zhuang, Z.; Starkey, J. A.; Mechref, Y.; Novotny, M. V.; Jacobson, S. C. Electrophoretic analysis of N-glycans on microfluidic devices. Anal. Chem. 2007, 79 (18), 7170–5. (17) Wada, Y.; Azadi, P.; Costello, C. E.; Dell, A.; Dwek, R. A.; Geyer, H.; Geyer, R.; Kakehi, K.; Karlsson, N. G.; Kato, K.; Kawasaki, N.; Khoo, K. H.; Kim, S.; Kondo, A.; Lattova, E.; Mechref, Y.; Miyoshi, E.; Nakamura, K.; Narimatsu, H.; Novotny, M. V.; Packer, N. H.; Perreault, H.; Peter-Katalinic, J.; Pohlentz, G.; Reinhold, V. N.; Rudd, P. M.; Suzuki, A.; Taniguchi, N. Comparison of the methods for profiling glycoprotein glycans--HUPO Human Disease Glycomics/Proteome Initiative multi-institutional study. Glycobiology 2007, 17 (4), 411–22. (18) Bergen, H. R.; Lacey, J. M.; O’brien, J. F.; Naylor, S. Online singlestep analysis of blood proteins: the transferrin story. Anal. Biochem. 2001, 296 (1), 122–9. (19) Lacey, J. M.; Bergen, H. R.; Magera, M. J.; Naylor, S.; O’brien, J. F. Rapid determination of transferrin isoforms by immunoaffinity liquid chromatography and electrospray mass spectrometry. Clin. Chem. 2001, 47 (3), 513–8. (20) Sanz-Nebot, V.; Balaguer, E.; Benavente, F.; Neususs, C.; Barbosa, J. Characterization of transferrin glycoforms in human serum by CE-UV and CE-ESI-MS. Electrophoresis 2007, 28 (12), 1949–57. (21) Nelsestuen, G. L.; Zhang, Y.; Martinez, M. B.; Key, N. S.; Jilma, B.; Verneris, M.; Sinaiko, A.; Kasthuri, R. S. Plasma protein profiling: unique and stable features of individuals. Proteomics 2005, 5 (15), 4012–24. (22) Bondarenko, P. V.; Cockrill, S. L.; Watkins, L. K.; Cruzado, I. D.; Macfarlane, R. D. Mass spectral study of polymorphism of the apolipoproteins of very low density lipoprotein. J. Lipid Res. 1999, 40 (3), 543–55. (23) Atger, V.; Beyne, P.; Frommherz, K.; Roullet, J. B.; Drueke, T. Presence of Apo B48, and relative Apo CII deficiency and Apo CIII enrichment in uremic very-low density lipoproteins. Ann. Biol. Clin. (Paris) 1989, 47 (8), 497–501. (24) Harake, B.; Caines, P. S.; Thibert, R. J.; Cheung, R. M. A simple micromethod for rapid assessment of the distribution of apolipoprotein C isoforms in very-low-density lipoprotein. Clin. Biochem. 1991, 24 (3), 255–60. (25) Alborn, W. E.; Prince, M. J.; Konrad, R. J. Relationship of apolipoprotein A5 and apolipoprotein C3 levels to serum triglycerides in patients with type 2 diabetes. Clin. Chim. Acta 2007, 378 (1-2), 154–8. (26) Gervaise, N.; Garrigue, M. A.; Lasfargues, G.; Lecomte, P. Triglycerides, apo C3 and Lp B:C3 and cardiovascular risk in type II diabetes. Diabetologia 2000, 43 (6), 703–8. (27) Huff, M. W.; Nestel, P. J. Metabolism of apolipoproteins CII, CIII1, CIII2 and VLDL-B in human subjects consuming high carbohydrate diets. Metabolism 1982, 31 (5), 493–8. (28) Mauger, J. F.; Couture, P.; Bergeron, N.; Lamarche, B. Apolipoprotein C-III isoforms: kinetics and relative implication in lipid metabolism. J. Lipid Res. 2006, 47 (6), 1212–8. (29) Homoncik, M.; Sieghart, W.; Formann, E.; Schmid, M.; Ferenci, P.; Gangl, A.; Jilma, B.; Peck-Radosavljevic, M. Erythropoietin treatment is associated with more severe thrombocytopenia in patients with chronic hepatitis C undergoing antiviral therapy. Am. J. Gastroenterol. 2006, 101 (10), 2275–82. (30) Homoncik, M.; Jilma-Stohlawetz, P.; Schmid, M.; Ferlitsch, A.; Peck-Radosavljevic, M. Erythropoietin increases platelet reactivity and platelet counts in patients with alcoholic liver cirrhosis: a

612

Journal of Proteome Research • Vol. 8, No. 2, 2009

Harvey et al.

(31) (32) (33)

(34) (35)

(36)

(37) (38)

(39) (40)

(41) (42) (43)

(44) (45)

(46)

randomized, double-blind, placebo-controlled study. Aliment Pharmacol Ther 2004, 20 (4), 437–43. Stone, J. H. HELLP syndrome: hemolysis, elevated liver enzymes, and low platelets. JAMA 1998, 280 (6), 559–62. Report of the National High Blood Pressure Education Program Working Group on High Blood Pressure in Pregnancy. Am. J. Obstet. Gynecol. 2000, 183 (1), S1-S22. ACOG practice bulletin. Diagnosis and management of preeclampsia and eclampsia. Number 33, January 2002. American College of Obstetricians and Gynecologists. Int. J. Gynaecol. Obstet., 2002, 77 (1), 67-75. Knaus, W. A.; Draper, E. A.; Wagner, D. P.; Zimmerman, J. E. APACHE II: a severity of disease classification system. Crit. Care Med. 1985, 13 (10), 818–29. Barker, J. N.; Weisdorf, D. J.; Defor, T. E.; Blazar, B. R.; Mcglave, P. B.; Miller, J. S.; Verfaillie, C. M.; Wagner, J. E. Transplantation of 2 partially HLA-matched umbilical cord blood units to enhance engraftment in adults with hematologic malignancy. Blood 2005, 105 (3), 1343–7. Brunstein, C. G.; Barker, J. N.; Weisdorf, D. J.; Defor, T. E.; Miller, J. S.; Blazar, B. R.; Mcglave, P. B.; Wagner, J. E. Umbilical cord blood transplantation after nonmyeloablative conditioning: impact on transplantation outcomes in 110 adults with hematologic disease. Blood 2007, 110 (8), 3064–70. Chan, D. C.; Watts, G. F. Apolipoproteins as markers and managers of coronary risk. QJM 2006, 99 (5), 277–87. Chan, D. C.; Watts, G. F.; Nguyen, M. N.; Barrett, P. H. Apolipoproteins C-III and A-V as predictors of very-low-density lipoprotein triglyceride and apolipoprotein B-100 kinetics. Arterioscler. Thromb. Vasc. Biol. 2006, 26 (3), 590–6. Pemberton, P. W.; Smith, A.; Warnes, T. W. Non-invasive monitoring of oxidant stress in alcoholic liver disease. Scand. J. Gastroenterol. 2005, 40 (9), 1102–8. Nelsestuen, G. L.; Harvey, S. B.; Zhang, Y.; Kasthuri, R. S.; Sinaiko, A. R.; Ely, E. W.; Bernard, G. R.; Homoncik, M.; Jilma, B. Top-down proteomic analysis by MALDI-TOF profiling: concentrationindependent biomarkers Proteomics - Clinical. Applications 2008, 2, 158–166. Kasthuri, R. S.; Verneris, M. R.; Ibrahim, H. N.; Jilma, B.; Nelsestuen, G. L. Studying multiple protein profiles over time to assess biomarker validity. Expert Rev Proteomics 2006, 3 (4), 455–64. Salpeter, S. R. A Systematic review of metformin treatment in persons at risk for diabetes mellitus. Diabetes, 2008 (July), www.cardiologyreviewonline.com. Guidone, C.; Manco, M.; Valera-Mora, E.; Iaconelli, A.; Gniuli, D.; Mari, A.; Nanni, G.; Castagneto, M.; Calvani, M.; Mingrone, G. Mechanisms of recovery from type 2 diabetes after malabsorptive bariatric surgery. Diabetes 2006, 55 (7), 2025–31. Parkes, J.; Guha, I. N.; Roderick, P.; Rosenberg, W. Performance of serum marker panels for liver fibrosis in chronic hepatitis C. J. Hepatol. 2006, 44 (3), 462–74. Vaith, P.; Assmann, G.; Uhlenbruck, G. Characterization of the oligosaccharide side chain of apolipoprotein C-III from human plasma very low density lipoproteins. Biochim. Biophys. Acta 1978, 541 (2), 234–40. Levent, G.; Ali, A.; Ahmet, A.; Polat, E. C.; Aytac, C.; Ayse, E.; Ahmet, S. Oxidative stress and antioxidant defense in patients with chronic hepatitis C patients before and after pegylated interferon alfa-2b plus ribavirin therapy. J. Transl. Med. 2006, 4, 25.

PR800751X