Anna Merrill (Presenter)
University of Washington
Bio: Dr. Anna Merrill is a Clinical Chemistry Fellow at the University of Washington in Seattle.
Authorship: Anna Merrill (1), Hannah Pflaum (1), Thomas Laha (1), Pratistha Ranjitkar (1), Clark Henderson (1), Rhona Jack (2), Jane Dickerson (2), Andrew Hoofnagle (1)
(1) Department of Laboratory Medicine, University of Washington, Seattle, WA, (2) Seattle Children’s Hospital, Seattle, WA.
| Short Abstract Retinol-binding protein (RBP), a 21-kDa polypeptide that transports vitamin A from hepatic storage to peripheral target locations, is commonly quantified in the clinical laboratory using nephelometry, a methodology with notable shortcomings. We validated a “plug-and-play” clinical LC-MS/MS assay for RBP that circumvents many of these challenges, improving testing quality and utility for multiple patient groups, including preterm neonates. Assay calibration was accomplished using an external calibration scheme. Both pre-validation and validation studies were carried out according to CLSI document C62-A. During assay validation, we witnessed sporadic failures that were eventually linked to faulty trypsin digestion. Symptoms of this problem will be discussed, along with troubleshooting strategies and suggestions for quality monitoring in high-throughput clinical mass spectrometry assays. |
Long Abstract
Retinol-binding protein (RBP) is a 21-kDa polypeptide that, together with transthyretin, transports fat-soluble vitamin A from hepatic storage to peripheral target locations, where it is essential for vision, immune function, reproduction, and cellular communication. Inappropriate levels of vitamin A can cause severe epithelial degeneration, such as night blindness in cases of deficiency and hair loss in hypervitaminosis. Serum RBP and vitamin A concentrations are frequently paired to compute the vitamin A-RBP molar ratio, an important marker of vitamin A nutritional status in various pediatric patient populations. Moreover, multiple studies suggest that elevated serum RBP may be associated with insulin resistance and pathogenesis of type 2 diabetes [1, 2].
In the clinical laboratory, RBP is commonly quantified using immunonephelometry, a methodology with a few notable shortcomings. First, these FDA-cleared RBP assays are only available on a limited number of commercial platforms, which can render in-house testing impractical depending on the laboratory’s other testing needs. Second, nephelometric approaches lack the sensitivity to quantify low RBP concentrations (< 1 mg/dL) often found in preterm neonates and other patients with impaired liver synthetic capacity [3]. Finally, assays based on light scattering are susceptible to interferences from matrix constituents that increase sample turbidity, a phenomenon we have observed in children treated with ketogenic diets for intractable seizures. We reasoned that liquid chromatography-tandem mass spectrometry (LC-MS/MS) could circumvent many of the aforementioned challenges plaguing the current clinical RBP assay and improve the quality and utility of measured RBP concentrations for multiple patient populations.
The protein sequence of RBP was obtained from UniProtKB and imported into Skyline for in silico generation of tryptic peptide precursor ion isolation lists and collision energies. Four candidate peptides (FSGTWYAMAK, YWGVASFLQK, LIVHNGYCDGR, QEELCLAR) were selected based on amino acid composition, precision, stability, and overall performance. Individual peptide transitions used for RBP quantification were selected empirically.
The approach used for RBP quantification is based on our published methodology for quantification of vitamin D-binding globulin (VDBG) [4]. In this regard, RBP serves as an exemplary model for “plug-and-play” proteomics, where clinical mass spectrometry workflows are intentionally designed for applicability to multiple protein markers with minimal changes in reagents and equipment. All sample preparation steps are compatible with both individual tube and multi-well plate formats. In brief, serum proteins are denatured with trifluoroethanol (TFE), reduced with dithiothreitol, and alkylated with iodoacetamide. Trypsin is added following sample dilution, with proteolytic digestion allowed to proceed for 30 minutes at 37° C with agitation. The resulting peptides are further diluted prior to analysis by reversed-phase LC-MS/MS (Xevo TQMS, Waters).
Assay calibration was accomplished using an external calibration scheme. A set of five reference calibrators was constructed by spiking recombinant RBP into salmon serum. We used quantitative amino acid analysis (Molecular Structure Facility, University of California, Davis, CA) to establish the concentration of RBP in the stock reference calibrator solution. Interestingly, this approach revealed a ~30% positive bias in RBP concentrations measured by a commonly used nephelometric method. Working calibrators consisted of pooled human serum spiked with recombinant RBP to a final concentration of approximately 10 mg/dL, which was used undiluted or diluted with pooled human serum and salmon serum to approximately 7.5, 5.0, 2.5, and 1.0 mg/dL.
Assay pre-validation was carried out according to CLSI document C62-A. Specificity and selectivity were assessed by quantifying background peaks in IS-containing salmon serum, which, for three of the four candidate peptides, were found to be, on average, 2.6% of the IS at the expected retention time. Surprisingly, we discovered that a peptide in the salmon serum interfered with transitions for the QEELCLAR peptide, which had not been predicted by BLAST search. At this point, the QEELCLAR peptide was eliminated from consideration in all future experiments. To evaluate potential contributions of matrix effects, multiple sample matrices, including salmon serum, the highest working calibrator, and patient specimens, were combined at 80/20, 50/50, and 20/80 proportions. Recoveries varied by peptide and transition but, in general, were deemed acceptable because at least one transition per peptide recovered between 85% and 115% for all sample mixes tested. Repeatability was preliminarily determined to be < 15% CV at three points over the calibration range (110% LLMI, 90% ULMI, and MI mid-range).
Areas of special focus throughout the validation process included improving functional sensitivity and benchmarking performance using an extensive method comparison study. The design and monitoring of robust quality assurance metrics is another critical component. During assay validation, we witnessed sporadic failures that were eventually linked to faulty trypsin digestion. Symptoms of this problem will be discussed in detail, along with troubleshooting strategies and suggestions for quality monitoring in high-throughput clinical mass spectrometry assays.
References & Acknowledgements:
1. Yang Q, Graham TE, Mody N, Preitner F, Peroni OD, Zabolotny JM, et al. Serum retinol binding protein 4 contributes to insulin resistance in obesity and type 2 diabetes. Nature 2005; 436: 356-62.
2. Graham TE, Yang Q, Bluher M, Hammarstedt A, Ciaraldi TP, Henry RR, et al. Retinol-binding protein 4 and insulin resistance in lean, obese, and diabetic subjects. N Engl J Med 2006; 354: 2552-63.
3. Shenai JP, Chytil F, Jhaveri A, Stahlman MT. Plasma vitamin A and retinol-binding protein in premature and term neonates. J Pediatr 1981; 99: 302-5.
4. Henderson CM, Lutsey PL, Misialek JF, Laha TJ, Selvin E, Eckfeldt JH, et al. Measurement by a novel LC-MS/MS methodology reveals similar serum concentrations of vitamin D-binding protein in blacks and whites. Clin Chem 2016; 62: 179-87.
| Description | Y/N | Source |
| Grants | no | |
| Salary | no | |
| Board Member | no | |
| Stock | no | |
| Expenses | no |
IP Royalty: no
| Planning to mention or discuss specific products or technology of the company(ies) listed above: | no |