Evaluation of a Machine Learning-Based Prognostic Model for Unrelated Hematopoietic Cell Transplantation Donor Selection.

The survival of patients undergoing hematopoietic cell transplantation (HCT) from unrelated donors for acute leukemia exhibits considerable variation, even after stringent genetic matching. To improve the donor selection process, we attempted to create an algorithm to quantify the likelihood of survival to 5 years after unrelated donor HCT for acute leukemia, based on the clinical characteristics of the donor selected. All standard clinical variables were included in the model, which also included average leukocyte telomere length of the donor based on its association with recipient survival in severe aplastic anemia, and links to multiple malignancies. We developed a multivariate classifier that assigned a Preferred or NotPreferred label to each prospective donor based on the survival of the recipient. In a previous analysis using a resampling method, recipients with donors labeled Preferred experienced clinically compelling better survival compared with those labeled NotPreferred by the test. However, in a pivotal validation study in an independent cohort of 522 patients, the overall survival of the Preferred and NotPreferred donor groups was not significantly different. Although machine learning approaches have successfully modeled other biological phenomena and have led to accurate predictive models, our attempt to predict HCT outcomes after unrelated donor transplantation was not successful.

Performance of a multi-biomarker score measuring rheumatoid arthritis disease activity in the CAMERA tight control study.

OBJECTIVES: To evaluate the performance of individual biomarkers and a multi-biomarker disease activity (MBDA) score in the early rheumatoid arthritis (RA) patient population from the computer assisted management in early rheumatoid arthritis (CAMERA) study. METHODS: Twenty biomarkers were measured in the CAMERA cohort, in which patients were treated with either intensive or conventional methotrexate-based treatment strategies. The MBDA score was calculated using the concentrations of 12 biomarkers (SAA, IL-6, TNF-RI, VEGF-A, MMP-1, YKL-40, MMP-3, EGF, VCAM-1, leptin, resistin and CRP) according to a previously trained algorithm. The performance of the scores was evaluated relative to clinical disease activity assessments. Change in MBDA score over time was assessed by paired Wilcoxon rank sum test. Logistic regression was used to evaluate the ability of disease activity measures to predict radiographic progression. RESULTS: The MBDA score had a significant correlation with the disease activity score based on 28 joints-C reactive protein (DAS28-CRP) (r=0.72; p<0.001) and an area under the receiver operating characteristic curve for distinguishing remission/low from moderate/high disease activity of 0.86 (p<0.001) using a DAS28-CRP cut-off of 2.7. In multivariate analysis the MBDA score, but not CRP, was an independent predictor of disease activity measures. Additionally, mean (SD) MBDA score decreased from 53 (18) at baseline to 39 (16) at 6 months in response to study therapy (p<0.0001). Neither MBDA score nor clinical variables were predictive of radiographic progression. CONCLUSIONS: This multi-biomarker test performed well in the assessment of disease activity in RA patients in the CAMERA study. Upon further validation, this test could be used to complement currently available disease activity measures and improve patient care and outcomes.

Rapid, deep and precise profiling of the plasma proteome with multi-nanoparticle protein corona.

Large-scale, unbiased proteomics studies are constrained by the complexity of the plasma proteome. Here we report a highly parallel protein quantitation platform integrating nanoparticle (NP) protein coronas with liquid chromatography-mass spectrometry for efficient proteomic profiling. A protein corona is a protein layer adsorbed onto NPs upon contact with biofluids. Varying the physicochemical properties of engineered NPs translates to distinct protein corona patterns enabling differential and reproducible interrogation of biological samples, including deep sampling of the plasma proteome. Spike experiments confirm a linear signal response. The median coefficient of variation was 22%. We screened 43 NPs and selected a panel of 5, which detect more than 2,000 proteins from 141 plasma samples using a 96-well automated workflow in a pilot non-small cell lung cancer classification study. Our streamlined workflow combines depth of coverage and throughput with precise quantification based on unique interactions between proteins and NPs engineered for deep and scalable quantitative proteomic studies.

Development of a multi-biomarker disease activity test for rheumatoid arthritis.

BACKGROUND: Disease activity measurement is a key component of rheumatoid arthritis (RA) management. Biomarkers that capture the complex and heterogeneous biology of RA have the potential to complement clinical disease activity assessment. OBJECTIVES: To develop a multi-biomarker disease activity (MBDA) test for rheumatoid arthritis. METHODS: Candidate serum protein biomarkers were selected from extensive literature screens, bioinformatics databases, mRNA expression and protein microarray data. Quantitative assays were identified and optimized for measuring candidate biomarkers in RA patient sera. Biomarkers with qualifying assays were prioritized in a series of studies based on their correlations to RA clinical disease activity (e.g. the Disease Activity Score 28-C-Reactive Protein [DAS28-CRP], a validated metric commonly used in clinical trials) and their contributions to multivariate models. Prioritized biomarkers were used to train an algorithm to measure disease activity, assessed by correlation to DAS and area under the receiver operating characteristic curve for classification of low vs. moderate/high disease activity. The effect of comorbidities on the MBDA score was evaluated using linear models with adjustment for multiple hypothesis testing. RESULTS: 130 candidate biomarkers were tested in feasibility studies and 25 were selected for algorithm training. Multi-biomarker statistical models outperformed individual biomarkers at estimating disease activity. Biomarker-based scores were significantly correlated with DAS28-CRP and could discriminate patients with low vs. moderate/high clinical disease activity. Such scores were also able to track changes in DAS28-CRP and were significantly associated with both joint inflammation measured by ultrasound and damage progression measured by radiography. The final MBDA algorithm uses 12 biomarkers to generate an MBDA score between 1 and 100. No significant effects on the MBDA score were found for common comorbidities. CONCLUSION: We followed a stepwise approach to develop a quantitative serum-based measure of RA disease activity, based on 12-biomarkers, which was consistently associated with clinical disease activity levels.

Pre-analytical effects of blood sampling and handling in quantitative immunoassays for rheumatoid arthritis.

Variability in pre-analytical blood sampling and handling can significantly impact results obtained in quantitative immunoassays. Understanding the impact of these variables is critical for accurate quantification and validation of biomarker measurements. Particularly, in the design and execution of large clinical trials, even small differences in sample processing and handling can have dramatic effects in analytical reliability, results interpretation, trial management and outcome. The effects of two common blood sampling methods (serum vs. plasma) and two widely-used serum handling methods (on the clot with ambient temperature shipping, "traditional", vs. centrifuged with cold chain shipping, "protocol") on protein and autoantibody concentrations were examined. Matched serum and plasma samples were collected from 32 rheumatoid arthritis (RA) patients representing a wide range of disease activity status. Additionally, a set of matched serum samples with two sample handling methods was collected. One tube was processed per manufacturer's instructions and shipped overnight on cold packs (protocol). The matched tube, without prior centrifugation, was simultaneously shipped overnight at ambient temperatures (traditional). Upon delivery, the traditional tube was centrifuged. All samples were subsequently aliquoted and frozen prior to analysis of protein and autoantibody biomarkers. Median correlation between paired serum and plasma across all autoantibody assays was 0.99 (0.98-1.00) with a median % difference of -3.3 (-7.5 to 6.0). In contrast, observed protein biomarker concentrations were significantly affected by sample types, with median correlation of 0.99 (0.33-1.00) and a median % difference of -10 (-55 to 23). When the two serum collection/handling methods were compared, the median correlation between paired samples for autoantibodies was 0.99 (0.91-1.00) with a median difference of 4%. In contrast, significant increases were observed in protein biomarker concentrations among certain biomarkers in samples processed with the 'traditional' method. Autoantibody quantification appears robust to both sample type (plasma vs. serum) and pre-analytical sample collection/handling methods (protocol vs. traditional). In contrast, for non-antibody protein biomarker concentrations, sample type had a significant impact; plasma samples generally exhibit decreased protein biomarker concentrations relative to serum. Similarly, sample handling significantly impacted the variability of protein biomarker concentrations. When biomarker concentrations are combined algorithmically into a single test score such as a multi-biomarker disease activity test for rheumatoid arthritis (MBDA), changes in protein biomarker concentrations may result in a bias of the score. These results illustrate the importance of characterizing pre-analytical methodology, sample type, sample processing and handling procedures for clinical testing in order to ensure test accuracy.

Characterization of a multiplex, 12-biomarker test for rheumatoid arthritis.

Rheumatoid arthritis (RA) is a chronic inflammatory disorder that primarily involves the joints. Accurate and frequent assessment of RA disease activity is critical to optimal treatment planning. A novel algorithm has been developed to determine a multi-biomarker disease activity (MBDA) score based upon measurement of the concentrations of 12 serum biomarkers in multiplex format. Biomarker assays from several different platforms were used in feasibility studies to identify biomarkers of potential significance. These assays were adapted to a multiplex platform for training and validation of the algorithm. In this study, the analytical performance of the underlying biomarker assays and the MBDA score was evaluated. Quantification of 12 biomarkers was performed with multiplexed sandwich immunoassays in three panels. Biomarker-specific capture antibodies were bound to specific locations in each well; detection antibodies were labeled with electrochemiluminescent tags. Data were acquired with a Sector Imager 6000, and analyte concentrations were determined. Parallelism, dynamic range, cross-reactivity, and precision were established for each biomarker as well as for the MBDA score. Interference by serum proteins, heterophilic antibodies, and common RA therapies was also assessed. The individual biomarker assays had 3-4 orders of magnitude dynamic ranges, with good reproducibility across time, operators, and reagent lots; the MBDA score had a median coefficient of variation of <2% across the score range. Cross-reactivity as well as interference by serum rheumatoid factor (RF), human anti-mouse antibodies (HAMA), or common RA therapies, including disease-modifying antirheumatic drugs and biologics, was minimal. The same MBDA score was observed in different subjects despite having different biomarker profiles, supporting prior literature reports that multiple pathways contribute to RA.

Validation of a novel multibiomarker test to assess rheumatoid arthritis disease activity.

OBJECTIVE: Quantitative assessment of disease activity in rheumatoid arthritis (RA) is important for patient management, and additional objective information may aid rheumatologists in clinical decision making. We validated a recently developed multibiomarker disease activity (MBDA) test relative to clinical disease activity in diverse RA cohorts. METHODS: Serum samples were obtained from the Index for Rheumatoid Arthritis Measurement, Brigham and Women's Hospital Rheumatoid Arthritis Sequential Study, and Leiden Early Arthritis Clinic cohorts. Levels of 12 biomarkers were measured and combined according to a prespecified algorithm to generate the composite MBDA score. The relationship of the MBDA score to clinical disease activity was characterized separately in seropositive and seronegative patients using Pearson's correlations and the area under the receiver operating characteristic curve (AUROC) to discriminate between patients with low and moderate/high disease activity. Associations between changes in MBDA score and clinical responses 6-12 weeks after initiation of anti-tumor necrosis factor or methotrexate treatment were evaluated by the AUROC. RESULTS: The MBDA score was significantly associated with the Disease Activity Score in 28 joints using the C-reactive protein level (DAS28-CRP) in both seropositive (AUROC 0.77, P < 0.001) and seronegative (AUROC 0.70, P < 0.001) patients. In subgroups based on age, sex, body mass index, and treatment, the MBDA score was associated with the DAS28-CRP (P < 0.05) in all seropositive and most seronegative subgroups. Changes in the MBDA score at 6-12 weeks could discriminate both American College of Rheumatology criteria for 50% improvement responses (P = 0.03) and DAS28-CRP improvement (P = 0.002). Changes in the MBDA score at 2 weeks were also associated with subsequent DAS28-CRP response (P = 0.02). CONCLUSION: Our findings establish the criterion and discriminant validity of a novel multibiomarker test as an objective measure of RA disease activity to aid in the management of RA in patients with this condition.

Application of high hydrostatic pressure to dissociate aggregates and refold proteins.

Non-denaturing pressures of around 2000 bar are effective for eliminating and refolding protein aggregates and may be applicable in various phases of protein manufacturing to decrease aggregate levels in products and improve process yields. Lower aggregate levels can result in reduced immunogenicity of proteins and enable the correct refolding of proteins that might not be recovered with traditional techniques. High pressure treatment can also be used to conduct selective PEGylation and protease cleavage reactions while minimizing protein aggregation. High pressure processes have been used in the food industry for over 50 years and large scale (300 L) systems are commercially available, enabling production of proteins on the kilogram scale. This review summarizes the utility of high pressure refolding to remove and refold protein aggregates, enhance therapeutic proteins, and facilitate manufacturing improvements at industrial scales.

Confronting high-throughput protein refolding using high pressure and solution screens.

Over-expression of heterologous proteins in Escherichia coli is commonly hindered by the formation of inclusion bodies. Nevertheless, refolding of proteins in vitro has become an essential requirement in the development of structural genomics (proteomics) and as a means of recovering functional proteins from inclusion bodies. Many distinct methods for protein refolding are now in use. However, regardless of method used, developing a reliable protein refolding protocol still requires significant optimization through trial and error. Many proteins fall into the category of "Challenging" or "Difficult to Express" and are problematic to refold using traditional chaotrope-based refolding techniques. This review discusses new methods for improving protein refolding, such as implementing high hydrostatic pressure, using small molecule additives to enhance traditional protein refolding strategies, as well as developing practical methods for performing refolding studies to maximize their reliability and utility. The strategies examined here focus on high-throughput, automated refolding screens, which can be applied to structural genomic projects.