Abstract
Objective
Emerging proteomics techniques can be used to establish proteomic outcome signatures and to identify candidate biomarkers for survival following traumatic injury. We applied high-resolution liquid chromatography-mass spectrometry (LC-MS) and multiplex cytokine analysis to profile the plasma proteome of survivors and non-survivors of massive burn injury to determine the proteomic survival signature following a major burn injury.
Design
Proteomic discovery study.
Setting
Five burn hospitals across the U.S.
Patients
Thirty-two burn patients (16 non-survivors and 16 survivors), 19–89 years of age, were admitted within 96 h of injury to the participating hospitals with burns covering >20% of the total body surface area and required at least one surgical intervention.
Interventions
None.
Measurements and Main Results
We found differences in circulating levels of 43 proteins involved in the acute phase response, hepatic signaling, the complement cascade, inflammation, and insulin resistance. Thirty-two of the proteins identified were not previously known to play a role in the response to burn. IL-4, IL-8, GM-CSF, MCP-1, and β2-microglobulin correlated well with survival and may serve as clinical biomarkers.
Conclusions
These results demonstrate the utility of these techniques for establishing proteomic survival signatures and for use as a discovery tool to identify candidate biomarkers for survival. This is the first clinical application of a high-throughput, large-scale LC-MS-based quantitative plasma proteomic approach for biomarker discovery for the prediction of patient outcome following burn, trauma or critical illness.
Keywords: burn, inflammation, proteomic profiling, plasma proteins, LC-MS, biomarker
INTRODUCTION
Burn injury is the most severe form of trauma, accounting for ~330,000 deaths per year worldwide (1); this devastating injury affects nearly every organ system and leads to significant morbidity and mortality (2). The dearth of biomarkers for early assessment of outcome following burn, trauma, or critical illness forces clinicians to rely on parameters such as heart rate, hypotension, and hyperglycemia to monitor the patient's clinical trajectory. These measures do not allow prediction of patient outcome. Discovery of a biomarker or proteomic signature comprised of the expression of a series of proteins would enable early assessment of patient risk for succumbing to burn, trauma, or critical illness. This signature would facilitate early identification of patients requiring aggressive therapy to improve their chances of survival.
The development of new high-throughput proteomic technologies facilitates the identification of protein profiles that can serve as diagnostic biomarkers or expression signatures for early assessment of predicted outcome as a result of disease or injury. Because burns, trauma, and critical illness involve multiple organs, a reliable high-throughput method to assess the systemic response—such as profiling the plasma proteome—would aid in identification of biomarkers, therapeutic targets, proteins with unexpected participation in outcome, and outcome signatures. Measurement of these proteins may be complicated by variable protein abundance due to biological heterogeneity, a broad dynamic range covering more than 10 orders of magnitude, and the complexity of the plasma proteome (3). Despite recent advances in liquid chromatography-mass spectrometry (LC-MS) approaches that have the potential to compensate for these confounders, this platform is often not robust enough to offer the high reproducibility required for large-scale clinical studies (3). Therefore, an effective method for evaluating the reproducibility among the large number of biological analytes is essential. Here, we describe the novel application of quantitative, large-scale proteomic profiling via LC-MS coupled with multiplex cytokine measurement to establish a plasma proteome survival signature and to discover new candidate plasma biomarkers for survival of a severe burn injury. The use of an 18O-labeled “universal” reference sample enables quantitation with concurrent evaluation of reproducibility, while allowing the simultaneous integration of a labeled approach with the label-free approach for quantitation (4). Furthermore, we propose that proteomic outcome signatures may reveal new drug targets in the form of proteins or pathways previously unknown to play a role in the disease or injury process and that a patient's proteomic survival signature and outcome trajectory may be altered via therapeutic intervention.
MATERIALS AND METHODS
Patients
The study, a part of the Inflammation and the Host Response Glue Grant, was approved by the Institutional Review Boards of the University of Texas Medical Branch (Galveston, TX), Loyola University Medical College (Chicago, IL), University of Texas Southwestern (Dallas, TX), University of Washington Seattle (Seattle, WA), and Massachusetts General Hospital (Boston, MA). Seventy-five patients who were enrolled from 2000–2005 and met the following criteria were considered for inclusion in this study: 19–89 years of age, admitted within 96 h after injury to the participating hospitals, and had burns covering more than 20% of the total body surface area requiring at least one surgical intervention. After admission, patients were treated according to the standard of burn care established by this consortium (5). Each subject or family member signed a written informed consent form. Blood was drawn between post-burn days 0 and 19 for plasma isolation. Demographics (age, dates of burn and admission, gender, burn size, and depth of burn) and concomitant injuries such as inhalation injury, infection, morbidity, and mortality were recorded prospectively throughout hospital course. The patients were retrospectively divided into two groups—survivors and non-survivors. Survivors and non-survivors were matched based on age, burn size, burn mechanism, sex, and ethnicity. Patients in both groups arrived at the participating hospitals at similar post-burn times, received similar hospital care, and had a similar time from burn to resuscitation.
Sample Handling
The protocol calls for blood drawn at the time of surgical intervention on an approximate weekly basis. Peripheral blood samples were collected prior to the induction of anesthesia and processed within one hour of being drawn as previously described (6, 7). To minimize the effect of temporal expression, we selected blood samples drawn 7 days or more prior to death and chose a sample from a similar time point (post-burn days 0–19) for the matched surviving patient. One non-surviving patient had a single plasma sample at post-burn day 19, so that sample was used.
Quantitative Proteomic Strategy
An integrated LC-MS-based proteomic strategy that incorporates an 18O-labeled “universal” reference sample for achieving reliable quantitation was used to profile patient plasma (4) (Supplemental Fig. 1 [Supplemental Digital Content 1]). The 18O-labeled “universal” reference sample served as a comprehensive internal standard in each sample, enabling isotope labeling-based quantification of relative protein abundances in patients relative to the reference and facilitating better evaluation of the platform reproducibility for normalization of the label-free abundance data. The label-free and isotope labeling-based dual quantitation approaches can be applied simultaneously to take advantage of their complementariness to achieve robust quantitation.
The overall dynamic measurement range was enhanced by integrating immunoaffinity chromatography for removing the 12 most abundant proteins in human plasma (8) and a cysteinyl-peptide enrichment-based fractionation at the peptide level (9) with an automated high-resolution LC-MS system. The overall throughput of analysis is only slightly compromised since the strategy only produces two fractions per patient sample. Peptide features detected by LC-MS are identified by matching the detected accurate masses and normalized elution times to a pre-established accurate mass and time (AMT) tag database applying the AMT tag strategy (10). The peptide AMT tag database was established based on an extensive LC-MS/MS survey of pooled trauma patients as previously described (11); the database includes the calculated masses and normalized elution times for all identified peptides so that the database serves as a “look-up table” for identifying peptides from LC-MS analyses using the AMT tag approach. Quantitative information for each identified feature was extracted based on either isotope 16O/18O ratios or label-free 16O MS-intensities.
Plasma Protein Processing and Liquid Chromatography-Mass Spectrometric Analysis
The initial protein concentration for each sample was determined by Coomassie protein assay (Pierce). A reference peptide sample was generated by pooling 100 μL of plasma from each patient. The patient samples and the pooled reference sample were individually depleted of the 12 most abundant plasma proteins (albumin, IgG, α1-antitrypsin, IgA, IgM, transferrin, haptoglobin, α1-acid glycoprotein, α2-macroglobulin, apolipoprotein A-I, apolipoprotein A-II, and fibrinogen) using a prepacked ProteomeLab™ IgY-12 affinity LC-10 column (Beckman Coulter) on an Agilent 1100 series HPLC system (Agilent). The flow-through fractions were then individually concentrated using iCON concentrators (Pierce) and digested with trypsin. Immunoaffinity depletion and plasma protein digestion were performed as previously described (8).
The peptides from the reference sample were labeled by stable isotope 18O via trypsin-catalyzed 18O labeling as previously described (10). Identical aliquots of labeled reference peptides were then mixed with an equal amount of each individual patient sample so that each final processed patient sample contained an identical reference sample. Each mixed patient sample was then fractionated as described (11). All final peptide fractions were then analyzed using a fully automated custom-built capillary LC platform coupled on-line using an in-house-manufactured ESI interface to an 11.5 Tesla Fourier transform ion cyclotron resonance (FTICR) mass spectrometer similar to as previously described (10).
Proteomic Data Processing
Analyses of datasets obtained from quantitative LC-Fourier transform ion cyclotron resonance were performed as previously described (10, 12). The AMT tag database was established based on those peptides and proteins previously identified in the trauma patient plasma proteome (11). Peptides were identified in the AMT tag database based on the normalized elution times and accurate mass measurements within a 2.5 ppm mass error and a 2% normalized elution time error. All identified peptides were reported for their 16O/18O isotopic ratios in addition to the label-free 16O-MS intensities. The 16O/18O ratio data and label-free intensity data were processed separately. The median of 16O/18O ratios (log2 scale) was normalized to zero to correct potential errors in mixing the labeled reference and patient samples. The label-free intensity data were also normalized across the patients using the central tendency global normalization technique (13). The data from cysteinyl and non-cysteinyl fractions for each patient were combined, and the corresponding proteins or protein groups determined using the software tool ProteinProphet (14). Two separate datasets were generated for statistical analysis of the proteomic difference. One dataset contained 16O/18O ratios for each detected peptide from each patient, and the other dataset contained normalized label-free intensities.
Orthogonal Validation
β2-Microglobulin, Apolipoprotein A, Antithrombin, and Plasminogen were analyzed using a nephelometric technique on the Plasma Protein Analyzer BN II (Dade Behring) according to the manufacturer's instructions. Complement C8 and C9 were analyzed using a radial immunodiffusion technique (The Binding Site) according to the manufacturer's instructions. Factor X and prothrombin were analyzed by ELISA (Hyphen Biomed) according to the manufacturer's instructions.
Statistical Analysis of Proteomic Data
Patient (label-free 16O intensity) data were filtered to remove peptides that mapped to more than one protein group and peptides that were not observed in more than half of the samples from at least one of the patient groups. The intensities were then scaled to the protein concentration and normalized to the reference intensities. The resulting matrix was then permuted 1,000 times, and the intensity difference between non-survivors and survivors was calculated in each permutation. The number of times the observed ranked difference exceeded the null difference was used to generate a p value. Principal components analysis was performed using peptide features from the significant proteins.
Plasma Cytokine Measurement
Plasma concentrations of 22 cytokines were measured using the Linco Research multiplex bead array and analyzed using MiraiBio software. Unpaired Student's t-tests were used to compare differences in cytokine expression.
Pathway Analysis
Protein annotations were obtained through a review of National Center for Biotechnology Information databases and the Ingenuity Pathways Knowledge Base (www.ingenuity.com). The Ingenuity Pathways Knowledge Base was used to overlay the concentration changes of significant proteins on the genomic and proteomic networks of previously known interactions between human orthologs (6).
RESULTS
Patients
Of the 75 patients meeting the study criteria, 19 were non-survivors. Two who died in nursing homes after discharge and one who died the day of injury were excluded from the analysis to minimize variation in time after injury and prior to death. There were no significant differences between non-survivors (n=16) and matched survivors (n=16) in age, percent third-degree burn, burn mechanism, or sex (Table 1). Although there was a small but significant difference in total burn size (total body surface area burned: 63±21%, non-survivors vs. 49±16%, survivors, p=0.03), the full thickness burn extent was comparable. Non-surviving patients died within 32±19 days of burn. Causes of death are shown in Supplemental Table 1 (Supplemental Digital Content 1). Eight patients died of multiple organ failure, and 8 patients of cardiac dysfunction; sepsis was a cause or co-cause of death in 9 patients.
Table 1.
Patient characteristics
| Characteristica | Non-Survivors (n=16) | Survivors (n=16) | p<0.05c |
|---|---|---|---|
| Age (years) | 49±19 | 39±13 | No |
| Sex (F/M) | 6/10 | 6/10 | No |
| Total body surface area burned (%) | 63±21 | 49±17 | Yes |
| 3rd degree (%) | 48±29 | 34±19 | No |
| Inhalation injury (yes/no) | 6/10 | 8/8 | No |
| Days from injury to admit | 0.4±0.5 | 0.3±0.4 | No |
| Length of hospital stay (days) | 68±48 | ||
| Post-burn days until death | 32±19 | N/Ab | |
| Type burn (flame/scald/other) | 14/1/1 | 14/1/1 | No |
Data are expressed as percentage or mean ± standard deviation.
Not applicable.
Student's t test or Chi-squared test.
Liquid Chromatography-Mass Spectrometry Platform Performance
We performed a total of 64 LC-MS analyses. The assessment of platform reproducibility based on the pair-wise correlation of 18O-labeled reference sample is shown in Supplemental Fig. 2 (Supplemental Digital Content 1). Pair-wise correlation of 18O-labeled reference peptide intensities between any two patient samples had an average correlation of 0.93±0.03, indicating good reproducibility (Supplemental Fig. 2a [Supplemental Digital Content 1]). Significantly higher variations were observed between patient peptide intensities with an average correlation of 0.75±0.11. Comparison of the results from isotope labeling (16O/18O) and the label-free approach revealed good overall correlation (Supplemental Fig. 2b [Supplemental Digital Content 1]).
The estimated dynamic range for this platform is ~105, allowing the detection of concentrations of ~500 ng/mL (3). Low-abundance proteins (1–100 ng/mL) were detected and included platelet factor 4, cathepsin d, metalloproteinase inhibitor 1, insulin-like growth factor binding protein 2, insulin-like growth factor 1, and hepatocyte growth factor-like protein.
Identification of Proteins through Liquid Chromatography-Mass Spectrometry
A total of 4,163 different peptides corresponding to 602 different plasma proteins were identified (strategy outlined in Supplemental Fig. 1 [Supplemental Digital Content 1]). Thirty-nine proteins were differentially detected in survivors and non-survivors (Table 2). Fold changes (non-survivors to survivors) were calculated for the 201 proteins that were detected in samples from both survivors and non-survivors (Table 3).
Table 2.
Proteins differentially detected in survivors versus non-survivors
| IPIa | Protein | Reported active in | Direction of change in NS vs. S |
|---|---|---|---|
| IPI00004656.1 | Beta-2-microglobulin | burn not survival | Up |
| IPI00007240.1 | coagulation factor XIII, B polypeptide | burn not survival | Down |
| IPI00029168.1 | Lipoprotein A | burn not survival | Up |
| IPI00166930.4 | CARBOXYPEPTIDASE N SUBUNIT 2 PRECURSOR | burn response | Down |
| IPI00219018.4 | Glyceraldehyde-3-phosphate dehydrogenase | burn response | Down |
| IPI00020996.3 | Insulin-like growth factor binding protein complex acid labile subunit | burn response | Down |
| IPI00550991.1 | serpin peptidase inhibitor, clade A (alpha-1 antiproteinase, antitrypsin), member 3 | burn response | Down |
| IPI00032179.2 | serpin peptidase inhibitor, clade C (antithrombin), member 1 | burn response | Down |
| IPI00019580.1 | plasminogen | burn/survival | Down |
| IPI00006662.1 | Apolipoprotein D | none | Up |
| IPI00451624.1 | cartilage acidic protein 1 | none | Down |
| IPI00297160.2 | CD44 | none | Down |
| IPI00019576.1 | Coagulation factor X precursor | none | Up |
| IPI00015618.1 | coiled-coil domain containing 49 | none | Down |
| IPI00022395.1 | complement component 9 | none | Down |
| IPI00011252.1 | complement component C8 alpha chain | none | Down |
| IPI00294395.1 | Complement component C8 beta chain precursor | none | Down |
| IPI00291867.3 | Complement factor I precursor | none | Down |
| IPI00029751.1 | Contactin 1 | none | Up |
| IPI00013452.7 | glutamyl-prolyl-tRNA synthetase | none | Down |
| IPI00041065.3 | HGF activator like protein | none | Down |
| IPI00018891.5 | HKR1 GLI-Kruppel family member HKR1 | none | Down |
| IPI00328089.2 | human immunodeficiency virus type I enhancer binding protein 3 | none | Down |
| IPI00292530.1 | inter-alpha (globulin) inhibitor H1 | none | Down |
| IPI00021304.1 | Keratin, type II cytoskeletal 2 epidermal | none | Up |
| IPI00011876.1 | methylthioadenosine phosphorylase | none | Up |
| IPI00414037.1 | Myosin tail domain-containing protein /UNCHARACTERIZED PROTEIN C9ORF93 | none | Down |
| IPI00022640.1 | neurogranin | none | Up |
| IPI00218732.2 | paraoxonase 1 | none | Down |
| IPI00419262.1 | PEPTIDYLPROLYL ISOMERASE B / Cyclophilin B | none | Down |
| IPI00219425.3 | Poliovirus receptor beta | none | Down |
| IPI00397137.2 | PREDICTED: hypothetical protein XP_374121 | none | Down |
| IPI00176169.1 | PREDICTED: similar to RIKEN cDNA 2510009E07 | none | Up |
| IPI00022445.1 | Pro-platelet basic protein | none | Down |
| IPI00019568.1 | Prothrombin precursor | none | Down |
| IPI00465275.2 | RAN binding protein 9 | none | Down |
| IPI00010317.3 | Ras-related GTP binding B | none | Down |
| IPI00239405.3 | spectrin repeat containing, nuclear envelope 2 | none | Down |
| IPI00020165.1 | Splice Isoform 1 of Chitotriosidase 1 precursor | none | Down |
International Protein Index ID number
Table 3.
Fold-change in protein between survivors and non-survivors
| IPIa | Protein Name | Fold Change NS:S |
|---|---|---|
| IPI00020165.1 | Splice Isoform 1 of Chitotriosidase 1 precursor | −1.7965 |
| IPI00215894.1 | Splice Isoform LMW of Kininogen precursor | −1.5808 |
| IPI00015618.1 | Hypothetical protein FLJ20291 | −1.0383 |
| IPI00451624.1 | Splice Isoform 1 of Cartilage acidic protein 1 precursor | −0.9083 |
| IPI00013452.7 | Hypothetical protein DKFZp313B047 | −0.8405 |
| IPI00007240.1 | Coagulation factor XIII B chain precursor | −0.7962 |
| IPI00397137.2 | PREDICTED: hypothetical protein XP_374121 | −0.7851 |
| IPI00019580.1 | Plasminogen precursor | −0.7790 |
| IPI00010317.3 | RagB | −0.7482 |
| IPI00419262.1 | Peptidyl-prolyl cis-trans isomerase | −0.7367 |
| IPI00011252.1 | Complement component C8 alpha chain precursor | −0.6699 |
| IPI00022395.1 | Complement component C9 precursor | −0.6519 |
| IPI00021841.1 | Apolipoprotein A-I precursor | −0.6517 |
| IPI00022417.4 | Leucine-rich alpha-2-glycoprotein precursor | −0.6061 |
| IPI00029751.1 | Splice Isoform 1 of Contactin 1 precursor | −0.5999 |
| IPI00022445.1 | Platelet basic protein precursor | −0.5950 |
| IPI00041065.3 | HGF activator like protein | −0.5811 |
| IPI00011261.1 | Complement component C8 gamma chain precursor | −0.5084 |
| IPI00291867.3 | Complement factor I precursor | −0.4991 |
| IPI00032258.4 | Complement C4 precursor | −0.4465 |
| IPI00553177.1 | Alpha-1-antitrypsin precursor | −0.4460 |
| IPI00294395.1 | Complement component C8 beta chain precursor | −0.4404 |
| IPI00022429.3 | Alpha-1-acid glycoprotein 1 precursor | −0.4359 |
| IPI00027848.1 | Macrophage mannose receptor precursor | −0.4007 |
| IPI00013954.3 | PREDICTED: thioredoxin reductase 3 | −0.3957 |
| IPI00022895.5 | Alpha-1B-glycoprotein precursor | −0.3820 |
| IPI00000875.5 | Elongation factor 1-gamma | −0.3777 |
| IPI00022640.1 | Neurogranin | −0.3721 |
| IPI00294193.3 | Splice Isoform 1 of Inter-alpha-trypsin inhibitor heavy chain H4 precursor | −0.3720 |
| IPI00022371.1 | Histidine-rich glycoprotein precursor | −0.3700 |
| IPI00059940.2 | APEM9336 | −0.3684 |
| IPI00032179.2 | Antithrombin III variant | −0.3506 |
| IPI00164623.3 | Complement C3 precursor | −0.3465 |
| IPI00022488.1 | Hemopexin precursor | −0.3450 |
| IPI00018305.1 | Insulin-like growth factor binding protein 3 precursor | −0.3403 |
| IPI00218732.2 | Serum paraoxonase/arylesterase 1 | −0.3401 |
| IPI00022431.1 | Alpha-2-HS-glycoprotein precursor | −0.3303 |
| IPI00294004.1 | Vitamin K-dependent protein S precursor | −0.3168 |
| IPI00292950.3 | Heparin cofactor II precursor | −0.3107 |
| IPI00550991.1 | Alpha-1-antichymotrypsin precursor | −0.3032 |
| IPI00032291.1 | Complement C5 precursor | −0.2979 |
| IPI00414037.1 | Myosin tail domain-containing protein | −0.2909 |
| IPI00159652.10 | PREDICTED: KIAA0826 protein | −0.2885 |
| IPI00019568.1 | Prothrombin precursor | −0.2879 |
| IPI00009244.3 | Adapter-related protein complex 1 sigma 1B subunit | −0.2858 |
| IPI00298853.5 | Vitamin D-binding protein precursor | −0.2847 |
| IPI00402142.2 | PREDICTED: similar to Hypothetical protein CBG23155 | −0.2840 |
| IPI00550315.1 | Ig kappa chain C region | −0.2805 |
| IPI00029863.3 | Alpha-2-antiplasmin precursor | −0.2765 |
| IPI00021727.1 | C4b-binding protein alpha chain precursor | −0.2729 |
| IPI00022389.1 | Splice Isoform 1 of C-reactive protein precursor | −0.2690 |
| IPI00019591.1 | Splice Isoform 1 of Complement factor B precursor | −0.2577 |
| IPI00019576.1 | Coagulation factor X precursor | −0.2546 |
| IPI00009793.1 | Complement C1r-like proteinase | −0.2394 |
| IPI00021304.1 | Keratin, type II cytoskeletal 2 epidermal | −0.2349 |
| IPI00298497.3 | Fibrinogen beta chain precursor | −0.2334 |
| IPI00021439.1 | Actin, cytoplasmic 1 | −0.2253 |
| IPI00296176.1 | Coagulation factor IX precursor | −0.2130 |
| IPI00025276.1 | Splice Isoform XB of Tenascin-X precursor | −0.2125 |
| IPI00021891.5 | Splice Isoform Gamma-B of Fibrinogen gamma chain precursor | −0.2077 |
| IPI00239405.3 | Splice Isoform 1 of Nesprin 2 | −0.2063 |
| IPI00043201.2 | OTTHUMP00000042330 | −0.2061 |
| IPI00026199.1 | Plasma glutathione peroxidase precursor | −0.2026 |
| IPI00303602.1 | Hypothetical protein FLJ12806 | −0.1977 |
| IPI00009920.1 | Complement component C6 precursor | −0.1915 |
| IPI00292530.1 | Inter-alpha-trypsin inhibitor heavy chain H1 precursor | −0.1912 |
| IPI00299026.3 | Tissue alpha-L-fucosidase precursor | −0.1891 |
| IPI00218746.1 | Complement Component 1, q subcomponent, beta polypeptide precursor | −0.1842 |
| IPI00220644.6 | Pyruvate kinase 3 isoform 2 | −0.1829 |
| IPI00292946.1 | Thyroxine-binding globulin precursor | −0.1827 |
| IPI00029168.1 | Apolipoprotein | −0.1815 |
| IPI00007118.1 | Plasminogen activator inhibitor-1 precursor | −0.1733 |
| IPI00303963.1 | Complement C2 precursor | −0.1716 |
| IPI00032220.3 | Angiotensinogen precursor | −0.1711 |
| IPI00011694.1 | Trypsin I precursor | −0.1622 |
| IPI00242956.2 | Fc fragment of IgG binding protein | −0.1614 |
| IPI00456831.1 | PREDICTED: similar to hemicentin | −0.1613 |
| IPI00166930.4 | 62 kDa protein | −0.1523 |
| IPI00000494.4 | 60S ribosomal protein L5 | −0.1500 |
| IPI00023014.1 | Von Willebrand factor precursor | −0.1451 |
| IPI00216691.4 | Profilin-1 | −0.1441 |
| IPI00010295.1 | Carboxypeptidase N catalytic chain precursor | −0.1422 |
| IPI00028413.3 | Inter-alpha (globulin) Inhibitor H3 | −0.1403 |
| IPI00032311.3 | Lipopolysaccharide-binding protein precursor | −0.1245 |
| IPI00009028.1 | Tetranectin precursor | −0.1228 |
| IPI00022391.1 | Serum amyloid P-component precursor | −0.1145 |
| IPI00020996.3 | Insulin-like growth factor binding protein complex acid labile chain precursor | −0.1123 |
| IPI00329775.4 | PCPB protein | −0.1070 |
| IPI00455197.1 | PREDICTED: similar to RIKEN cDNA 2610034N24 | −0.1043 |
| IPI00291262.3 | Clusterin precursor | −0.1015 |
| IPI00006662.1 | Apolipoprotein D precursor | −0.0996 |
| IPI00289334.1 | Splice Isoform 1 of Filamin B | −0.0917 |
| IPI00007199.3 | Protein Z-dependent protease inhibitor precursor | −0.0910 |
| IPI00017601.1 | Ceruloplasmin precursor | −0.0896 |
| IPI00022331.1 | Phosphatidylcholine-sterol acyltransferase precursor | −0.0879 |
| IPI00292043.3 | Sidekick 2 | −0.0833 |
| IPI00027230.3 | Endoplasmin precursor | −0.0817 |
| IPI00026314.1 | Gelsolin precursor | −0.0813 |
| IPI00296099.3 | Thrombospondin-1 precursor | −0.0773 |
| IPI00165421.3 | SERPINC1 protein | −0.0760 |
| IPI00296053.3 | Fumarate hydratase, mitochondrial precursor | −0.0756 |
| IPI00007425.1 | Desmocollin 1b | −0.0750 |
| IPI00219217.2 | L-lactate dehydrogenase B chain | −0.0690 |
| IPI00304273.1 | Apolipoprotein A-IV precursor | −0.0655 |
| IPI00293925.1 | Splice Isoform 1 of Ficolin 3 precursor | −0.0650 |
| IPI00011876.1 | S-methyl-5-thioadenosine phosphorylase | −0.0645 |
| IPI00017696.1 | Complement C1s subcomponent precursor | −0.0619 |
| IPI00290903.5 | Hypothetical protein LOC130951 | −0.0549 |
| IPI00216699.1 | Splice Isoform 2 of Unc-112 related protein 2 | −0.0543 |
| IPI00291866.3 | Plasma protease C1 inhibitor precursor | −0.0482 |
| IPI00234095.3 | PREDICTED: similar to alanyl tRNA synthetase | −0.0482 |
| IPI00219018.4 | Glyceraldehyde-3-phosphate dehydrogenase, liver | −0.0476 |
| IPI00021857.1 | Apolipoprotein C-III precursor | −0.0431 |
| IPI00022937.2 | Coagulation factor V | −0.0421 |
| IPI00291175.3 | Vinculin isoform VCL | −0.0418 |
| IPI00021263.3 | 14-3-3 protein zeta/delta | −0.0377 |
| IPI00029739.3 | Splice Isoform 1 of Complement factor H precursor | −0.0362 |
| IPI00396348.1 | Serine (or cysteine) proteinase inhibitor, clade A (alpha-1 antiproteinase, antitrypsin), member 4 | −0.0263 |
| IPI00010402.2 | Hypothetical protein | −0.0263 |
| IPI00011062.1 | Splice Isoform 1 of Carbamoyl-phosphate synthase [ammonia], mitochondrial precursor | −0.0229 |
| IPI00465439.4 | Fructose-bisphosphate aldolase A | −0.0148 |
| IPI00018903.1 | Ventral anterior homeobox 2 | −0.0136 |
| IPI00297160.2 | Hypothetical protein DKFZp451K1918 | −0.0125 |
| IPI00008558.1 | Plasma kallikrein precursor | −0.0099 |
| IPI00022434.1 | Serum albumin precursor | −0.0026 |
| IPI00298971.1 | Vitronectin precursor | −0.0025 |
| IPI00305461.2 | Inter-alpha-trypsin inhibitor heavy chain H2 precursor | −0.0023 |
| IPI00021817.1 | Vitamin K-dependent protein C precursor | 0.0006 |
| IPI00470919.1 | Hypothetical protein DKFZp686K08164 | 0.0062 |
| IPI00029061.1 | Selenoprotein P precursor | 0.0120 |
| IPI00001701.2 | Doublecortin domain containing protein | 0.0222 |
| IPI00395488.2 | Vasorin | 0.0238 |
| IPI00550349.1 | Fetuin B | 0.0302 |
| IPI00180240.2 | Thymosin-like 3 | 0.0328 |
| IPI00022426.1 | AMBP protein precursor | 0.0333 |
| IPI00010970.3 | Scm-like with four mbt domains 2 | 0.0365 |
| IPI00182739.3 | Hypothetical protein FLJ37870 | 0.0409 |
| IPI00020091.1 | Alpha-1-acid glycoprotein 2 precursor | 0.0410 |
| IPI00171473.2 | Spondin-1 precursor | 0.0456 |
| IPI00023728.1 | Gamma-glutamyl hydrolase precursor | 0.0502 |
| IPI00022229.1 | Apolipoprotein B-100 precursor | 0.0530 |
| IPI00010779.3 | Tropomyosin 4 | 0.0547 |
| IPI00296165.5 | Complement C1r subcomponent precursor | 0.0585 |
| IPI00004373.1 | Mannose-binding protein C precursor | 0.0594 |
| IPI00008603.1 | Actin, aortic smooth muscle | 0.0638 |
| IPI00019943.1 | Afamin precursor | 0.0709 |
| IPI00024057.2 | Transgelin 2 | 0.0811 |
| IPI00019581.1 | Coagulation factor XII precursor | 0.0821 |
| IPI00163207.1 | Splice Isoform 1 of N-acetylmuramoyl-L-alanine amidase precursor | 0.0834 |
| IPI00298828.1 | Beta-2-glycoprotein I precursor | 0.0848 |
| IPI00333541.3 | Filamin A | 0.0854 |
| IPI00027487.3 | Creatine kinase, M chain | 0.0856 |
| IPI00064667.2 | Glutamate carboxypeptidase-like protein 2 precursor | 0.0882 |
| IPI00017672.2 | Hypothetical protein FLJ25678 | 0.0925 |
| IPI00073772.4 | Fructose-1,6-bisphosphatase | 0.0966 |
| IPI00002547.1 | Calpain 6 | 0.0985 |
| IPI00027235.1 | Splice Isoform 1 of Attractin precursor | 0.1013 |
| IPI00220327.2 | Keratin, type II cytoskeletal 1 | 0.1080 |
| IPI00298994.3 | Talin 1 | 0.1211 |
| IPI00006114.4 | Pigment epithelium-derived factor precursor | 0.1235 |
| IPI00013508.3 | Alpha-actinin 1 | 0.1277 |
| IPI00297550.7 | Coagulation factor XIII A chain precursor | 0.1313 |
| IPI00419453.1 | Ig kappa chain V-III region VG precursor | 0.1322 |
| IPI 00021885.1 | Splice Isoform Alpha-E of Fibrinogen alpha/alpha-E chain precursor | 0.1376 |
| IPI00012011.3 | Cofilin, non-muscle isoform | 0.1469 |
| IPI00025252.1 | Protein disulfide-isomerase A3 precursor | 0.1566 |
| IPI00025864.1 | Cholinesterase precursor | 0.1617 |
| IPI00003351.2 | Extracellular matrix protein 1 precursor | 0.1694 |
| IPI00006146.3 | Serum amyloid A2 | 0.1727 |
| IPI00306311.7 | Pleckstrin | 0.1801 |
| IPI00027223.2 | Isocitrate dehydrogenase [NADP] cytoplasmic | 0.1803 |
| IPI00010180.3 | Liver carboxylesterase 1 precursor | 0.1886 |
| IPI00023673.1 | Galectin-3 binding protein precursor | 0.1916 |
| IPI00007257.4 | Alcadein alpha-1 | 0.1928 |
| IPI00022418.1 | Splice Isoform 1 of Fibronectin precursor | 0.1965 |
| IPI00465436.3 | Catalase | 0.2045 |
| IPI00296608.6 | Complement component C7 precursor | 0.2053 |
| IPI00027462.1 | Calgranulin B | 0.2090 |
| IPI00294542.1 | WUGSC:H_2G3A.1 protein | 0.2366 |
| IPI00022394.2 | Complement C1q subcomponent, C chain precursor | 0.2374 |
| IPI00218413.1 | Biotinidase precursor | 0.2507 |
| IPI00216298.5 | Thioredoxin | 0.2646 |
| IPI00216589.1 | Splice Isoform 2 of ATP-binding cassette, sub-family B, member 9 precursor | 0.2771 |
| IPI00011651.1 | Receptor-type tyrosine-protein phosphatase gamma precursor | 0.2858 |
| IPI00478493.1 | Haptoglobin precursor | 0.2885 |
| IPI00003590.1 | Quiescin Q6 | 0.2937 |
| IPI00007739.1 | HemK protein homolog | 0.3321 |
| IPI00398997.2 | PREDICTED: similar to ribosomal protein L31 | 0.3609 |
| IPI00018136.1 | Splice Isoform 1 of Vascular cell adhesion protein 1 precursor | 0.3809 |
| IPI00029260.2 | Monocyte differentiation antigen CD14 precursor | 0.3941 |
| IPI00020986.2 | Lumican precursor | 0.4339 |
| IPI00019579.1 | Complement factor D precursor | 0.4561 |
| IPI00235407.1 | Serine/threonine kinase 36 | 0.5271 |
| IPI00218816.5 | Hemoglobin beta chain | 0.5327 |
| IPI00022420.3 | Plasma retinol-binding protein precursor | 0.5611 |
| IPI00021842.1 | Apolipoprotein E precursor | 0.5732 |
| IPI00298281.3 | Laminin gamma-1 chain precursor | 0.5847 |
| IPI00410714.1 | Hemoglobin alpha-1 globin chain | 0.6605 |
| IPI00478003.1 | Alpha-2-macroglobulin precursor | 0.7246 |
| IPI00168178.1 | Hypothetical protein FLJ34521 | 0.7644 |
| IPI00004656.1 | Beta-2-microglobulin precursor | 0.9139 |
International Protein Index ID number
Unsupervised methods (e.g., hierarchical clustering or principal components analyses) did not demonstrate separation according to survival, indicating that other factors aside from survival determine the majority of differences in protein abundance in the dataset. The 39 proteins that were significantly different between the cohorts clustered according to survival.
Cytokine Differences between Non-survivors and Survivors Detected by Multiplex Technology
Cytokine measurement revealed that IL-4, IL-8, GM-CSF, and MCP-1 were significantly higher in non-survivors than in survivors (p<0.05 for all, Fig. 1a–d). There were no differences in circulating levels of the 18 additional cytokines (Supplemental Table 2 [Supplemental Digital Content 1]).
Figure 1.
Comparison of serum levels of GM-CSF (a), IL-8 (b), IL-4 (c), and MCP-1 (d) between surviving and non-surviving patients. Data are expressed as the mean ± SEM. *p<0.05.
Biological Implications of the Proteomic Survival Signature
The 43 proteins (39 identified by mass spectrometry and 4 by multiplex cytokine analysis) were used to create a proteomic survival signature. To determine which biological pathways are implicated by the proteins expressed differentially between the two patient populations, we examined the 43 proteins using the Ingenuity Pathway Knowledge Base. The proteins were associated with the following biological functions (n): inflammatory disease (13), cell movement (15), hematological system development and function (17), immune response (16), cell death (13), immune and lymphatic system function and development (12), tissue development (14), molecular transport (15), and cell signaling (16). We further examined signaling pathways to identify the tissues involved in the systemic response to burn. The main canonical pathways that were affected were the coagulation cascade, the complement response, hepatic acute-phase response signaling, and inflammatory cytokine pathways. Proteins in the coagulation cascade were down-regulated (factor 2, factor 13B, plasminogen, SERPINC1), while factor 10 was up-regulated in non-survivors. Proteins active in the complement cascade were down-regulated in non-survivors (Complement components 9, 8A, 8B, and complement factor 1). Proteins active in acute phase response signaling and/or inflammatory cytokine signaling were detected at higher levels in non-survivors (IL-4, IL-8, CCL2), while SERPINA3, factor 2, plasminogen, and complement component 9 were more abundant in survivors. In the glucocorticoid receptor signaling pathways, CCL2, CSF2, IL-4, and IL-8 were up-regulated in non-survivors. The activity of the proteins in the complement cascade, coagulation cascade, acute phase response pathway, and cytokine signaling suggest that the response of the liver may be crucial to survival following a burn injury. Use of a networking tool that permits visualization of interactomes revealed that 13 of the 43 proteins interact directly with each other; IL-8, CCL-2, and F2 are the central molecules through which IL-4, LPA, CD44, plasminogen, SERPINC1, PPBP, HABP2, F10, CSF2, and B2M are linked (Supplemental Fig. 3 [Supplemental Digital Content 1]), suggesting that therapies designed to modulate the expression or interactions of these proteins might affect patient outcome.
Orthogonal Candidate Biomarker Measurement
We measured a subset of the LS-MS-identified proteins to determine whether standard clinical laboratory techniques could be used to detect differences in protein abundance between surviving and non-surviving patients. β2-microglobulin, apolipoprotein A, factor X, prothrombin, complement C8 and C9, anti-thrombin 3, and plasminogen were selected as candidates for orthogonal validation based on the p values and peptide-fold change data. Beta-2 microglobulin was found to be decreased in non-survivors using nephelometry (p<0.05), confirming the LC-MS data. Given the small sample sizes and sample preprocessing steps in the LC-MS not used in the clinical laboratory, confirmation of one of the proteins to a statistically significant degree was notable.
DISCUSSION
After severe burn, the major causes of death, particularly in adult and elderly patients, are infectious and septic complications resulting from organ dysfunction. The post-burn hypermetabolic response frequently leads to multi-organ failure and septic complications, increasing morbidity and mortality in the burn patient population. The hypermetabolic response is driven by systemic inflammation mediated by cytokines and acute-phase proteins. Although the burn size differed between the patient groups, the severity of the burn was the same. Several studies have demonstrated that the third-degree component is more important for the prediction of outcome than is the total body surface area burned (1, 16–18). Our data confirm this and strongly support the central role of the liver in survival from a severe burn injury.
Forty-three proteins were significantly altered in the plasma proteome of non-survivors compared to survivors, including coagulation proteins, adhesion molecules, inflammatory markers, metabolic markers, and hepatic acute-phase proteins. Although this is the first time that many of these proteins have been associated with a burn injury, the physiological processes that they represent are required for recovery from any disease or traumatic injury. The response to a burn injury requires an intact coagulation cascade for survival. Hypercoagulation with subsequent increased consumption of coagulation factors leads to disseminated intravascular coagulation, which is characterized by decreased levels of plasminogen and coagulation factor XIII in non-survivors. Therapeutic modulation of hypercoagulability may improve burn survival, as suggested by the clinical trial of drotrecogin alfa, recombinant activated C in patients with septic shock (19, 20). Hypercoagulation and disseminated intravascular coagulation occur alongside increased inflammation and abundance of adhesion molecules. Vascular cell adhesion molecule 1, a major cell adhesion molecule, is markedly increased in non-survivors immediately post-burn. White blood cells adhere to the endothelium and subsequently diapedese through the blood vessel wall, releasing inflammatory molecules such as metalloproteinases and cytokines into the tissue. Inflammatory markers such as TIMP-1, GM-CSF, IL-4, IL-8, and MCP-1 are up-regulated in non-survivors, indicating a greater inflammatory response in non-survivors and confirming our results from a genomic analysis of white blood cells (Lancet, under review). These markers are also associated with insulin resistance and metabolic derangements. Markers of insulin resistance such as retinol binding protein 4, sex hormone binding globulin, CD-14, glycosylphosphatidylinositol specific phospholipase D1, and pro-platelet basic protein were increased in non-survivors. The complement system is also deranged in non-survivors. Significant decreases in complement component 1, complement factor D, complement 8, and complement 9 in non-survivors further support the notion that hyperinflammatory and hyperimmune responses are modulated differently in non-survivors, ultimately contributing to immune compromise, subsequent infection, and septic complications.
The majority of the 43 proteins associated with burn survival are produced in the liver, supporting the hypothesis that the liver plays a central role in the post-burn response as well as in determining survival. The liver's metabolic, inflammatory, immune, and acute phase functions play a pivotal role in recovery from injury by modulating multiple pleiotropic pathways. The hepatic acute-phase response is characterized by an increase in production of acute phase proteins coupled with a failure to produce constitutive proteins. This shift in hepatic protein synthesis leads to physiologic alterations of transport, metabolic, and immune functions, which may prove to be useful for predicting survival. A study of 31,338 burn patients demonstrated that healthy liver function is essential for survival following traumatic injury (21). A sub-group analysis of 180 patients with liver disease prior to the burn injury revealed a 6–8-fold increase in mortality with pre-existing liver dysfunction was. Our study supports this conclusion and further suggests that the secretion of liver-derived proteins can be used as predictive markers for survival or as part of the proteomic survival signature in severely burned patients. Furthermore, the differentially expressed proteins indicate that hepatic insulin resistance accompanies post-burn peripheral insulin resistance. That insulin therapy may be acting to overcome hepatic insulin resistance may explain why insulin therapy improves morbidity and mortality in severely burned as well as critically ill patients (22–24).
Our study demonstrates that the integrated LC-MS-based quantitative proteomic discovery approach can be used to successfully identify candidate proteins for classification of patient outcomes, whether for use in establishing proteomic outcomes signatures or to distinguish proteins that will serve as potential biomarkers or targets for therapeutic intervention. The plasma proteomics-based discovery efforts have traditionally been challenged by lack of reproducibility of quantitative results, low dynamic range of detection, and low sample throughput. The present large-scale quantitative strategy was enabled by the incorporation of a stable isotope labeled “universal” reference sample into each patient sample, which provided an effective means for assessing the reproducibility of large-scale sample analyses (4). The labeled reference sample also enabled the application of isotope-labeling-based quantitation across an unlimited number of samples, distinguishing this technique from traditional labeling approaches, which are always limited to pair-wise comparisons or comparison between a few conditions. The simultaneous integration of the two quantitative approaches into a single strategy also enhanced the overall confidence of the quantitative data by providing cross-validation between the two approaches. The current quantitative strategy can also be effectively integrated with the immunoaffinity chromatography and cysteinyl-peptide enrichment-based limited fractionation with high resolution LC-MS to offer relatively high dynamic range measurements (>105) for achieving good proteome coverage and the throughput necessary for analyzing a large number of samples in a relatively short time. Low molecular weight and low abundance proteins such as cytokines cannot be effectively detected via LC-MS methods. By augmenting the LCMS technique with multiplex cytokine analysis, we can reliably detect the plasma proteome alterations that accompany the response to injury. However, because our cytokine assay only measures a select number of known cytokines, the likelihood of missing expression differences for other non-measured cytokines, or additional low molecular weight or low abundance proteins, is highly likely. Development of techniques to adequately assess these classes of proteins would allow full interrogation of the post-burn plasma proteome.
Because a huge immuno-inflammatory response accompanies a burn, trauma, or critical illness, high-throughput multiplex cytokine measurement was performed to determine whether complementary information regarding protein abundance could be used to build the proteomic survival signature. Pro-inflammatory cytokines have been suggested as potential biomarkers for outcome and as targets for interventional trials. Multiplex cytokine measurement has been used to successfully predict patient outcomes as well (25, 26), although attempts to alter cytokine expression to date have failed to improve patient outcomes. This combination of quantitative proteomic strategies has allowed us to quantitatively measure more than 600 proteins in samples from 32 patients; 43 proteins were identified as significantly differing between the patient cohorts, demonstrating a role for discovery proteomics in treating critical illness.
One drawback to this study was the small patient numbers. This was due to the lack of non-survivors enrolled in the program, as the intent to treat the patient is one of the enrollment criteria. Additional studies with larger patient numbers will be necessary to confirm the results reported here and to further develop this LC-MS approach. Future efforts may focus on determining whether the plasma protein profiles vary based on the cause of death. At this point, however, the numbers of patients are too small. Moreover, in light of the potential demonstrated in our small patient cohort, studies are being currently conducted with larger patient numbers to investigate outcomes that are not as clearly determined such as multi-organ failure. A further concern about this study is the difficulty in orthogonally validating the proteins identified by LC-MS. This may be due to the following: large biological variation in a small data set; removal of binding proteins (e.g., albumin) prior to LC-MS, which leads to the removal of additional proteins that can still be detected via clinical assays; the inability of antibody-based technologies to recognize the protein identified by LC-MS due to epitope masking; or the inability of the peptide database to distinguish between the presence of multiple isoforms of a particular protein. That β2-microglobulin was confirmed as a potential biomarker using traditional techniques demonstrated that the LC-MS platform can be used to identify targets for confirmation in further studies with larger patient numbers.
Conclusions
This study supports the utility of the quantitative LC-MS technique used here in discovering proteins relevant to critical illness, defining proteome signatures, and identifying potential biomarkers. High-throughput proteomic screening techniques such as the quantitative LC-MS technique may provide mechanistic insight into the clinical phenomenon that stress, injury, or acute illness can induce insulin resistance. Considering that post injury loss of more than 30–40% of lean body mass is associated with 100% mortality (27), the proteomic survival signature may also yield important clues regarding the mechanisms of survival. Finally, the finding that the majority of the 43 proteins with differential abundances between non-survivors and survivors had not been previously associated with the burn response let alone with survival illustrates the applicability of LC-MS and multiplex cytokine analysis for detecting potential biomarkers and concurrently defining a proteomic outcome signature. This provides for the possibility of improving clinical management and enabling development of successful interventions that alter outcome trajectories and decrease mortality after a severe burn.
Supplementary Material
ACKNOWLEDGMENTS
The authors thank Deborah Benjamin, Wes Benjamin, Maria Cantu, Mario Celis, Tabatha Elliot, Kathryn Epperson, Eric Henry, Holly Goode, Kara Hougen, Joanna Huddleston, Mary Kelly, Xuyang Liang, Maria Magno, Liz Montemayor, Marc Nicolai, Sylvia Ojeda, Maricela Pantoja, Cathy Reed, Lisa Richardson, Lucile Robles, Pam Stevens, Sierra Tinney, Judith Underbrink, Becky Whitlock, the nutrition department, and the respiratory therapy team for their assistance in obtaining the study measurements. The authors are grateful to Colleen D. Campbell and Dr. Kasie Cole-Edwards for their critical review of the manuscript and to Eileen Figueroa and Steve Schuenke for their technical expertise and support. The magnitude of the clinical and proteomic data reported here required the efforts of many individuals at participating institutions. In particular, we wish to acknowledge the supportive research environment created and sustained by the participants in the Glue Grant Program: Henry V. Baker, Ph.D., Ulysses G.J. Balis, M.D., Paul E. Bankey, M.D., Ph.D., Timothy R. Billiar, M.D., Bernard H. Brownstein, Ph.D., Steven E. Calvano, Ph.D., Irshad H. Chaudry, Ph.D., J. Perren Cobb, M.D., Joseph Cuschieri, M.D., Ronald W. Davis, Ph.D., Asit K. De, Ph.D., Brian G. Harbrecht, M.D., Douglas L. Hayden, M.A., Laura Hennessy, R.N., Jeffrey L. Johnson, M.D., James A. Lederer, Ph.D., Stephen F. Lowry, M.D., Ronald V. Maier, M.D., John A. Mannick, M.D., Philip H. Mason, Ph.D., Grace P. McDonald-Smith, M.Ed., Carol L. Miller-Graziano, Ph.D., Michael N. Mindrinos, Ph.D., Joseph P. Minei, M.D., Ernest E. Moore, M.D., Avery B. Nathens, M.D., Ph.D., M.P.H., Grant E. O'Keefe, M.D., M.P.H., Laurence G. Rahme, Ph.D., Michael B. Shapiro, M.D., Jason Sperry, M.D., Ph.D., John D. Storey, Ph.D., Robert Tibshirani, Ph.D., Mehmet Toner, Ph.D., H. Shaw Warren, M.D., Michael A. West, M.D., PhD., and Bram Wispelwey, M.S.
Financial Support: This study was supported in part by a Large Scale Collaborative Research Grant from the National Institute of General Medical Sciences (U54 GM-62119-04) awarded to Ronald G. Tompkins at the Massachusetts General Hospital, Boston, MA, in part by research grants awarded to David N. Herndon at the University of Texas Medical Branch, Galveston, TX by the National Institute of General Medical Sciences (P50 GM-60338, R01 GM-56687, T32 GM-008256), and in part by grants from Shriners Hospitals for Children to Celeste C Finnerty (8740, 8507). Portions of the research were supported by a grant from the NIH National Center for Research Resources (RR018522 and 5P41RR018522-10) and National Institute of General Medical Sciences (8 P41 GM103493-10). LC-MS proteomic analyses were performed in the Environmental Molecular Sciences Laboratory, a U.S. Department of Energy national scientific user facility located at the Pacific Northwest National Laboratory in Richland, Washington. The Pacific Northwest National Laboratory is a multi-program national laboratory operated by Battelle Memorial Institute for the U.S. Department of Energy under Contract DE-AC05-76RL01830.
Footnotes
The authors have not disclosed any potential conflicts of interest
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
REFERENCES
- 1.Peden M, McGee K, Sharma G. The Injury Chart Book: A graphical overview of the global burden of injuries. The World Health Organization; Geneva: 2002. [Google Scholar]
- 2.Herndon DN, Tompkins RG. Support of the metabolic response to burn injury. Lancet. 2004;363(9424):1895–1902. doi: 10.1016/S0140-6736(04)16360-5. [DOI] [PubMed] [Google Scholar]
- 3.Qian WJ, Jacobs JM, Liu T, et al. Advances and challenges in liquid chromatography-mass spectrometry-based proteomics profiling for clinical applications. Molecular & Cellular Proteomics. 2006;5(10):1727–1744. doi: 10.1074/mcp.M600162-MCP200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Qian WJ, Liu T, Petyuk VA, et al. Large-scale multiplexed quantitative discovery proteomics enabled by the use of an (18)O-labeled “universal” reference sample. Journal of Proteome Research. 2009;8(1):290–299. doi: 10.1021/pr800467r. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Silver GM, Klein MB, Herndon DN, et al. Standard operating procedures for the clinical management of patients enrolled in a prospective study of Inflammation and the Host Response to Thermal Injury. Journal of Burn Care & Research. 2007;28(2):222–230. doi: 10.1097/BCR.0B013E318031AA44. [DOI] [PubMed] [Google Scholar]
- 6.Calvano SE, Xiao W, Richards DR, et al. A network-based analysis of systemic inflammation in humans. Nature. 2005;437(7061):1032–1037. doi: 10.1038/nature03985. [DOI] [PubMed] [Google Scholar]
- 7.Finnerty CC, Jeschke MG, Herndon DN, et al. Temporal cytokine profiles in severely burned patients: a comparison of adults and children. Mol Med. 2008;14(9–10):553–560. doi: 10.2119/2007-00132.Finnerty. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Liu T, Qian WJ, Mottaz HM, et al. Evaluation of multiprotein immunoaffinity subtraction for plasma proteomics and candidate biomarker discovery using mass spectrometry. Molecular & Cellular Proteomics. 2006;5(11):2167–2174. doi: 10.1074/mcp.T600039-MCP200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Liu T, Qian WJ, Strittmatter EF, et al. High-throughput comparative proteome analysis using a quantitative cysteinyl-peptide enrichment technology. Anal Chem. 2004;76(18):5345–5353. doi: 10.1021/ac049485q. [DOI] [PubMed] [Google Scholar]
- 10.Qian WJ, Monroe ME, Liu T, et al. Quantitative proteome analysis of human plasma following in vivo lipopolysaccharide administration using 16O/18O labeling and the accurate mass and time tag approach. Molecular & Cellular Proteomics. 2005;4(5):700–709. doi: 10.1074/mcp.M500045-MCP200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Liu T, Qian WJ, Gritsenko MA, et al. High dynamic range characterization of the trauma patient plasma proteome. Molecular & Cellular Proteomics. 2006;5(10):1899–1913. doi: 10.1074/mcp.M600068-MCP200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Monroe ME, Tolic N, Jaitly N, et al. VIPER: an advanced software package to support high-throughput LC-MS peptide identification. Bioinformatics. 2007;23(15):2021–2023. doi: 10.1093/bioinformatics/btm281. [DOI] [PubMed] [Google Scholar]
- 13.Callister SJ, Barry RC, Adkins JN, et al. Normalization approaches for removing systematic biases associated with mass spectrometry and label-free proteomics. Journal of proteome research. 2006;5(2):277–286. doi: 10.1021/pr050300l. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Nesvizhskii AI, Keller A, Kolker E, et al. A statistical model for identifying proteins by tandem mass spectrometry. Anal Chem. 2003;75(17):4646–4658. doi: 10.1021/ac0341261. [DOI] [PubMed] [Google Scholar]
- 15.Kersey PJ, Duarte J, Williams A, et al. The International Protein Index: an integrated database for proteomics experiments. Proteomics. 2004;4(7):1985–1988. doi: 10.1002/pmic.200300721. [DOI] [PubMed] [Google Scholar]
- 16.Hart DW, Wolf SE, Beauford RB, et al. Determinants of blood loss during primary burn excision. Surgery. 2001;130(2):396–402. doi: 10.1067/msy.2001.116916. [DOI] [PubMed] [Google Scholar]
- 17.Jeschke MG, Chinkes DL, Finnerty CC, et al. Pathophysiologic response to severe burn injury. Ann Surg. 2008;248(3):387–401. doi: 10.1097/SLA.0b013e3181856241. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Sheridan RL, Schnitzer JJ. Management of the high-risk pediatric burn patient. J Pediatr Surg. 2001;36(8):1308–1312. doi: 10.1053/jpsu.2001.25805. [DOI] [PubMed] [Google Scholar]
- 19.Bernard GR, Vincent JL, Laterre PF, et al. Efficacy and safety of recombinant human activated protein C for severe sepsis. New Engl J Med. 2001;344(10):699–709. doi: 10.1056/NEJM200103083441001. [DOI] [PubMed] [Google Scholar]
- 20.Weintraub J, Satterwhite T, Allo M. The use of drotrecogin alfa recombinant activated protein C for severe sepsis in the critically burned patient: a new treatment approach. Burns. 2006;32(6):783–787. doi: 10.1016/j.burns.2006.02.021. [DOI] [PubMed] [Google Scholar]
- 21.Price LA, Thombs B, Chen CL, et al. Liver disease in burn injury: evidence from a national sample of 31,338 adult patients. Journal of Burns and Wounds. 2007;7:e1. [PMC free article] [PubMed] [Google Scholar]
- 22.Jeschke MG, Klein D, Herndon DN. Insulin treatment improves the systemic inflammatory reaction to severe trauma. Ann Surg. 2004;239(4):553–560. doi: 10.1097/01.sla.0000118569.10289.ad. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Langouche L, Van den Berghe G. Glucose metabolism and insulin therapy. Critical Care Clin. 2006;22(1):119–129. vii. doi: 10.1016/j.ccc.2005.09.005. [DOI] [PubMed] [Google Scholar]
- 24.van den Berghe G, Wouters P, Weekers F, et al. Intensive insulin therapy in critically ill patients. New Engl J Med. 2001;345(19):1359–1367. doi: 10.1056/NEJMoa011300. [DOI] [PubMed] [Google Scholar]
- 25.Finnerty CC, Herndon DN, Przkora R, et al. Cytokine expression profile over time in severely burned pediatric patients. Shock. 2006;26(1):13–19. doi: 10.1097/01.shk.0000223120.26394.7d. [DOI] [PubMed] [Google Scholar]
- 26.Remick DG. Cytokine therapeutics for the treatment of sepsis: why has nothing worked? Current Pharmaceutical Design. 2003;9(1):75–82. doi: 10.2174/1381612033392567. [DOI] [PubMed] [Google Scholar]
- 27.Chang DW, DeSanti L, Demling RH. Anticatabolic and anabolic strategies in critical illness: a review of current treatment modalities. Shock. 1998;10(3):155–160. doi: 10.1097/00024382-199809000-00001. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.