MSACL 2016 US Abstract

Gingerbread Men: A Cookie-cutter Bottom-up Proteomics Workflow for the Hungry?

Andy Hoofnagle (Presenter)
University of Washington

Bio: Dr. Hoofnagle is Associate Professor and Head of the Division of Clinical Chemistry in the Department of Laboratory Medicine at the University of Washington. After receiving his MD PhD from the University of Colorado, he completed his residency in Clinical Pathology at UW, concurrently completing his post-doctoral training in Metabolism, Endocrinology, and Nutrition. As a junior faculty member, his laboratory focused on developing novel methods for the quantification of proteins and small molecules in clinical samples using LC-MS/MS. His grant-funded research continues to focus on using analytical chemistry to epidemiologically answer questions in vitamin D biology, cardiovascular disease, and cancer. Dr. Hoofnagle also mentors students, fellows, and residents in gaining a deeply respectful admiration for the power of mass spectrometry in the care of patients.

Authorship: Tom Laha, MT(ASCP), Clark Henderson, PhD, Andy Hoofnagle, MD PhD
Department of Laboratory Medicine, University of Washington

Short Abstract

To test the hypothesis that a widely-used monoclonal immunoassay for vitamin D binding globulin (VDBG) was significantly affected by common polymorphisms, we developed a bottom-up proteomics assay. The method was validated according to our recently recommended guidelines for publication of novel biomarkers (Clinical Chemistry) and was used to quantify VDBG and determine the haplotype in samples accrued from a clinical research study that had been previously genotyped and analyzed by immunoassay. We have now applied the same workflow to other proteins and progress toward generating a plug-and-play workflow for medium- and high-abundance proteins in human serum/plasma will be discussed.

Long Abstract

The goal of replacing immunoassays

It is widely known that clinical immunoassays, including the FDA-cleared and CE-marked devices that are broadly distributed by commercial companies to laboratories around the world, can potentially suffer from many different interferences and poor concordance between platforms. Polyclonal nephelometric assays commonly used for medium- and high-abundance proteins are generally less susceptible to interferences when compared with monoclonal sandwich assays. However, inaccurate results and poor inter-platform concordance are still possible. The direct measurement of peptides using isotope dilution-tandem mass spectrometry holds promise to become the reference measurement procedure for medium- and high-abundance proteins in serum and plasma. An established, simplified workflow for the addition of new protein analytes could facilitate more widespread adoption of routine mass spectrometric platforms to replace clinical immunoassays.

Vitamin D biology

In recent years, it has become increasingly recognized that 25-hydroxyvitamin D [25(OH)D] may not be the most useful biomarker to assess vitamin D sufficiency and the associated increased risk for cardiovascular disease, cancer, and bone fractures, particularly across all races. Analytes that have been proposed to replace 25(OH)D include free and bioavailable 25(OH)D, which can be estimated using equations that use the concentration of albumin and vitamin D binding globulin (VDBG). Many recent publications have employed a monoclonal immunoassay from R&D Systems to quantify VDBG and have shown that calculated bioavailable 25(OH)D is much more similar across races than total 25(OH)D (e.g., Powe, NEJM, 2013). We hypothesized that two commonly polymorphic amino acids in VDBG could strongly influence the monoclonal immunoassay. We set out to develop a novel bottom-up proteomics method to quantify VDBG in human samples.

Validation of novel biomarkers

In Clinical Chemistry, we have proposed the experiments that should be performed and included in all publications of novel protein biomarkers, particularly those that use LC-MS/MS for quantification (collaboration with Dr. Russell Grant, PhD, Lab Corp, Inc.). The experiments validate the LLOQ, linearity, stability, and parallelism of the method. They also examine the potential influence of interferences and peptide degradation on the quantitative results of the assay, with transparency as a central tenet. One of the most important features of the proposed list of experiments is the relatively low number of injections required (N~125). We aimed to validate our novel assay using the recommended guidelines as a proof-of-principle. All data from the validation experiments were to be included in a Skyline file on the journal website as supplemental material.


In Skyline, we used nanoflow chromatography and parallel reaction monitoring in a targeted fashion to evaluate all peptides derived from the tryptic digestion of recombinant and purified protein as well as serum/plasma. Peptide stability, the reproducibility of peptide liberation during digestion, and the observation of potential chromatographic interferences were used to initially select quantitative peptides. The peptides including the two polymorphic sites and the initial quantitative peptides were synthesized as stable isotope-labeled internal standards. The initial quantitative peptides were then evaluated further under high-flow conditions and the two most reliable peptides were chosen for quantification with a third peptide chosen for quality assurance.


Our initial experiments demonstrated that denaturation of a small amount (10 µL) of serum or plasma with 50% trifluoroethanol in 50 mM ammonium bicarbonate (150 µL total volume) and 6.6 mM dithiothreitol at 65°C, alkylation with 25 mM iodoacetamide, quenching with 6.2 mM dithiothreitol, subsequent dilution with 1.2 mL 28 mM ammonium bicarbonate, digestion with a substantial amount of trypsin (74 µg, resulting trypsin:protein ratio of ~1:10), and subsequent acidification with formic acid resulted in a mixture that could be centrifuged, transferred, diluted, and analyzed by LC-MS/MS. Using a time-course analysis, we confirmed the rapid degradation of peptides with high concentrations of trypsin, supporting the concept of spiking internal standard peptides prior to digestion. Calibration with native human sample diluted into chicken serum minimized variability between days. The results of the novel assay strongly suggest that the monoclonal immunoassay is significantly affected by the common polymorphisms.


Using the presence of polymorphic peptides to assign the haplotypes of each subject, we were 97% correct in assigning the genotype. It was unclear if polymorphisms would affect digestion/liberation of quantitative peptides, so we analyzed the crystal structure and used peptide ratios to support the hypothesis that distant digestion does not affect quantification. Importantly, these peptide ratios allow for the detection of uncommon polymorphisms in quantitative peptides, a quality assurance parameter that is unavailable to immunoassays.


The assay for VDBG possesses all of the qualities that one would want for a high-throughput assay for the quantification of proteins in clinical research studies. We have begun to apply the workflow to other proteins and this presentation will provide an update on the other proteins that are amenable to this workflow as the field works toward replacing immunoassays with LC-MS/MS assays.

References & Acknowledgements:

Financial Disclosure

GrantsyesWaters, Inc.
Board Memberno

IP Royalty: no

Planning to mention or discuss specific products or technology of the company(ies) listed above: