Mass Spectrometry Method to Measure Membrane Proteins in Dried

Aug 22, 2017 - The dried blood spot (DBS) matrix has significant utility for applications in the field where venous blood collection and timely shipme...
6 downloads 9 Views 905KB Size
Subscriber access provided by UNIVERSITY OF ADELAIDE LIBRARIES

Article

A mass spectrometry method to measure membrane proteins in dried blood spots for the detection of blood doping practices in sport Holly D. Cox, and Daniel Eichner Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.7b02492 • Publication Date (Web): 22 Aug 2017 Downloaded from http://pubs.acs.org on August 22, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

A Mass Spectrometry Method to Measure Membrane Proteins in Dried Blood Spots for the Detection of Blood Doping Practices in Sport Holly D. Cox* and Daniel Eichner Sports Medicine Research and Testing Laboratory, 560 Arapeen Dr., Suite 150A, Salt Lake City, UT, 84108, United States ABSTRACT: The dried blood spot (DBS) matrix has significant utility for applications in the field where venous blood collection and timely shipment of labile blood samples is difficult. Unfortunately, protein measurement in DBS is hindered by high abundance proteins and matrix interference that increases with hematocrit. We have developed a DBS method to enrich for membrane proteins and remove soluble proteins and matrix interference. Following a wash in a series of buffers, the membrane proteins are digested with trypsin and quantitated by parallel reaction monitoring mass spectrometry methods. The DBS method was applied to the quantification of four cell-specific cluster of differentiation (CD) proteins used to count cells by flow cytometry, band 3 (CD233), CD71, CD45, and CD41. We demonstrate that the DBS method counts low abundance cell types, such as immature reticulocytes, as well as high abundance cell types, red blood cells, white blood cells and platelets. When tested in 82 individuals, counts obtained by the DBS method demonstrated good agreement with flow cytometry and automated hematology analyzers. Importantly, the method allows longitudinal monitoring of CD protein concentration and calculation of inter-individual variation which is difficult by other methods. Inter-individual variation of band 3 and CD45 was low, 6% and 8%, respectively, while variation of CD41 and CD71 was higher, 18% and 78%, respectively. Longitudinal measurement of CD71 concentration in DBS over an 8-week period demonstrated intra-individual variation between 17.1% – 38.7%. Thus, the method may allow stable longitudinal measurement of blood parameters currently monitored to detect blood doping practices.

Collection of athlete blood samples for anti-doping testing poses unique pre-analytical challenges.1 The blood is not collected in hospitals or clinics, but rather in the field often before or after sporting events. The sites of collection are often located a significant distance from the nearest anti-doping lab and the labile blood samples must be shipped in an expedited and temperature-controlled manner. Often, blood collections occur in remote regions where it is not possible to have the blood shipped in time to remain viable. Scheduling a phlebotomist for venous blood collection also hinders unannounced, out-of-competition testing and routine venous blood collection may be considered invasive for healthy individuals. Due to these limitations, dried blood spot (DBS) methods have been reported for most anti-doping tests that require blood and for several prohibited substances.2-4 DBS methods have been recently reported for detection of human growth hormone abuse, including fibronectin-1,5 growth hormone isoforms,6 and insulin-like growth factor-1 (IGF-1).7 To date, there is no DBS method for the detection of blood doping practices. Multiple methods of blood doping are abused by athletes to increase hemoglobin mass and oxygen transport to muscles to obtain an unfair competitive advantage. These practices include the use of recombinant erythropoietin (EPO), erythropoietin stimulating agents (ESAs), and autologous/allogeneic blood transfusions.8 The methods are used alone or in a variety of combinations designed to evade detection by current screening methods. Due to the wide variety of methods to manipulate blood, an indirect detection method has been adopted capable of detecting many forms of blood doping. This method, used since 2009 by all WADA-

accredited anti-doping laboratories, is the Athlete Biological Passport (ABP) program.9,10 The ABP method uses longitudinal monitoring of hemoglobin concentration [Hb], reticulocyte % (Ret%), and the OFF-score, which is defined as [Hb] -60 x √Ret%.11 Using this method, blood is analyzed in any anti-doping laboratory and the blood parameter data is entered into a database which contains all of the previous blood data for that athlete. Measurement of [Hb] and Ret% allows detection of unnatural increases that occur after EPO abuse or blood withdrawal, prior to autologous transfusion. However, the true advantage of the method is its ability to detect long-term erythropoietic feedback mechanisms which cause a decrease in Ret% for up to 3 weeks after abuse regardless of the method used for blood doping.12-15 Thus, by measurement of 2 blood parameters, it is possible to detect blood doping for up to 3 weeks. However, due to the blood collection limitations discussed, the frequency of testing is well below recommended levels which limits the success of the test.16 Collection of finger-prick or other capillary blood, which does not require a phlebotomist, and blood shipment on dried blood spot (DBS) cards will revolutionize the WADA ABP program and significantly increase the testing frequency. While several methods have been published to measure hemoglobin in DBS, there are no methods to measure reticulocyte%, which is the most sensitive parameter for blood doping.17,18 In the current report, a DBS method was developed to count cell types by quantification of cell-specific cluster of differentiation (CD) proteins, often used to count cells by flow

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

cytometry. While no specific CD protein exits for total reticulocytes, immature reticulocyte cells (IRC) are counted by measurement of CD71.19,20 IRCs constitute only a small fraction of the total reticulocyte population, 3%-14% in male athletes,21 which is approximately 0.03% of the total erythroid cell population. This small cell population is the first to respond to erythropoietic signals and may provide a more sensitive marker for blood doping practices.22,23 However, accurate measurement of these cells in DBS requires a method with good sensitivity and dynamic range. Additionally, band 3 (CD233) is a specific marker of RBCs or total erythrocytes,

Page 2 of 9

which can be used to calculate the IRC% of the total erythroid population. Measurement of IRC% provides a marker that is independent of plasma volume fluctuations and hyperhydration. Analysis of a volume independent marker, such as Ret% or IRC% is important because athletes often try to decrease their measured [Hb] and hematocrit values to hide blood doping practices through methods such as hyperhydration and plasma volume expansion. These methods have been demonstrated to affect OFF-score values used by the ABP method.24

Figure 1. DBS method diagram (A), %recovery of membrane proteins in DBS or liquid blood after washing (B), %recovery of sTFR when added to blood before spotting or after DBS extraction (C), stability expressed as cell number at day 9 and day 15 relative to day 1 (D). Error bars represent standard deviation.

CD proteins encompass a class of protein markers used to characterize cell types and disease states in blood.25 They are integral membrane proteins and thus, are difficult to measure in DBS by immunoassay methods due to interference from hemoglobin and antibody incompatibility with the detergents capable of membrane protein extraction. Therefore, a mass spectrometry method was needed. Several recent reports have described mass spectrometry methods to measure therapeutic proteins26 and endogenous soluble proteins in the moderate-tohigh concentration range in the DBS matrix.27,28 Protein measurement in DBS is often difficult due to the presence of high abundance blood proteins and matrix interference that often increases with hematocrit. These limitations are further compounded for membrane proteins which often exist at lower concentrations in blood. Currently, only a single DBS method has been reported for a membrane protein, ATP7b, and this method required immunopurification.29 In the current report, a novel method was developed to remove the high abundance

soluble proteins in buffer prior to trypsin digestion of the membrane proteins that remain on the spot. The DBS method is evaluated and tested in 82 individuals for comparison to current blood cell counting methods. Finally, the DBS method is tested to assess the longitudinal stability of CD71 concentrations in blood cells. While the method was developed for the detection of blood doping practices, it provides a new way to measure the concentration of CD proteins and membrane proteins on cells, which may have many applications. EXPERIMENTAL SECTION Details of the chemicals and materials are listed in the Supporting Information. Human Blood Samples. K2EDTA venous blood samples were obtained from healthy individuals from Bioreclamation (34) and from de-identified athlete samples (48) received in the laboratory. 48 males and 34 females were used for the

ACS Paragon Plus Environment

2

Page 3 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

comparison studies for CD71, CD45 and CD41. The samples were processed in 7 batches with a new calibration curve for each batch. Due to current shipping procedures, the athlete samples were not spotted until up to 48 hours after collection with storage temperatures up to 10 °C. Since band 3 is not stable in liquid blood under these conditions, it was not measured in the athlete samples. Band 3 was measured in blood obtained from 78 healthy individuals from Bioreclamation, 49 males and 29 females. Method comparison was performed by Deming regression and Bland-Altman analysis using the Analyze-it statistical software package for Excel. For the longitudinal study, venous blood samples from 5 males and 1 female were obtained in-house from volunteers with informed written consent. Six blood samples were obtained from each volunteer over an 8-week period. Compete blood count data was obtained for each sample using the Sysmex XT 2000i automated hematology analyzer. CD71+ cells were measured by flow cytometry, described below. DBS Spotting Method. Following measurement on the Sysmex XT 2000i, 20 µl of each blood sample was spotted in duplicate onto pre-treated, Whatman FTA-C cards. Prior to spotting, the cards were pre-treated with 20 µl of 2x Roche protease inhibitors and 10 mM PMSF (phenylmethylsulfonyl fluoride) and dried at room temperature. After blood spotting, the cards were dried for 3 hours and stored in sealed plastic bags with desiccant at room temperature. DBS Extraction. The full spot, 20 µl, was excised using an 8-mm punch and quartered with a razor blade. The spots were transferred to a 12 x 75 mm tube and washed in a series of 3 buffers previously demonstrated to remove soluble and peripheral membrane proteins.30,31 Each washing step, 3 ml buffer/spot, was performed for 30 min at 10 °C, with shaking, 1400 rpm. The first wash buffer was phosphate buffered saline (PBS), pH 7.4 containing 5.3 mM sodium phosphate, 1.5 mM potassium phosphate, 137 mM sodium chloride, and 1 mM PMSF. After incubation as described, the buffer was removed by vacuum suction and the spots were washed in 100 mM Na2CO3 and 1 mM PMSF, followed by the final wash in 95% ethanol containing 50 mM sodium acetate, pH 5.0. The spots were then equilibrated in 3 ml 50 mM Tris buffer, pH 8.0 for 15 minutes, 10 °C, 1400 rpm shaking. The buffer was removed by vacuum suction and the spots were transferred to a deep well 96-well plate. Proteins remaining in the spots were digested with trypsin, 2 µg/ spot, in 50 mM Tris buffer, pH 8.0 at 60 °C for 2 hours. Prior to digestion, stable isotopelabeled, cleavable, peptide internal standards were added to each sample. (The peptide internal standards are listed in the Supporting Information, Table S-1.) The digestion was stopped by addition of formic acid, 0.2% final concentration and the digestion solution was transferred to a fresh 96-well plate for mass spectrometry analysis. One peptide with internal standard was measured for each protein and cell type by mass spectrometry. For calculation of blood cell numbers from the DBS spots, a 3-point calibration curve was made from DBS spots containing blood at hematocrit (HCT) levels of 35%, 45%, and 55% with known cell numbers, measured on the Sysmex XT 2000i or by flow cytometry, for each cell type. The IRC # was calculated from the IRF% x reticulocyte #. The calibration curve spots were extracted in duplicate with each sample set. Using the calibration curve, the peak area ratio obtained by mass spectrometry analysis was converted to blood cell number for each sample. For calculation of CD71 concentration, a 3-point calibration curve was made from rat blood spots that were extracted with each sample set. After

washing and just prior to digestion, the rat blood spots were fortified with a recombinant standard for soluble transferrin receptor (sTFR) at 1.25 nM, 2.5 nM, and 5 nM concentration. Rat blood was selected because the amino acid sequence for rat CD71 differs from human. Liquid Chromatography-Tandem Mass Spectrometry. Liquid chromatography was performed on a Dionex UltiMate 3000 with a flow rate of 400 µl/min at 50 °C using a Waters BEH C18 column, 2.1 x 100 mm, 1.7 µm, preceded by a Waters BEH C18, 2.1 x 5 mm pre-column. Solvent A was 0.1% formic acid and Solvent B was acetonitrile. Peptides were separated on a gradient from 5% - 40% Solvent B over 18 min, then 40%-90% Solvent B, 3 min, a hold at 90% Solvent B, 3 min, then 90% - 5% Solvent B, 2 min and re-equilibration at 5% Solvent B for 4 min. Peptide quantitation by high resolution accurate mass spectrometry was performed on the Thermo QExactive Plus with 3.5 kV spray voltage, 350 °C capillary temperature, using nitrogen gas as the source and collision gas. The instrument was operated in parallel reaction monitoring mode (PRM) with 4X multiplexing. Resolution was 17500 full width at half maximum, with 200 ms maximum IT and 2.0 m/z isolation width. Peak areas were obtained from the extracted ion chromatograms of each product ion at 20 ppm mass accuracy. Details for each peptide are summarized in Supporting Information, Table S-1. Method Validation. Linearity and intra-assay precision were determined using a 7-point HCT curve, 30%, 35%, 40%, 45%, 50%, 55%, 60%, n=6. The cell number at each HCT concentration was measured on the Sysmex XT 2000i and plotted versus peak area ratio. Matrix interference was measured by addition of the pre-digested peptide IS to digested DBS extracts and to buffer. The interference was calculated as (1 – peak area IS in DBS/peak area IS buffer) x 100%. Matrix interference was measured at 30%, 35%, 40%, 45%, 50%, 55%, and 60% HCT. Additional method validation experiments including inter-day precision, selectivity, and digestion time course are described in the Supporting Information. Recovery of cell membrane proteins. To reduce matrix interference for this experiment, the blood cells were washed by 1:10 dilution in water containing 1x Roche protease inhibitors, 2 mM PMSF and vortexing for 20 sec. The cell membranes were collected by centrifugation at 20,000 x g for 5 min and the supernatant was removed. The washing step was performed 3X prior to spotting the membranes onto DBS cards. DBS spots containing washed cell membranes were either extracted in 3 buffers as described in the standard protocol or not extracted prior to spot digestion. Recovery was calculated as (peak area ratio after 3 buffer extraction/ peak area ratio with no buffer extraction) x 100%. Recovery of membrane proteins in liquid blood after washing in-solution 3x as described was determined by spotting blood before and after the washing steps followed by DBS extraction by the standard protocol. Removal of soluble transferrin receptor (sTFR). Removal of sTFR was determined by addition of recombinant sTFR to blood at 0 nM, 100 nM, 250 nM and 500 nM. sTFR was added either before spotting onto DBS cards or after spot extraction just prior to digestion. The recovery was calculated as (peak area ratio before extraction/peak area ratio after extraction) x 100%. Shipment stability. DBS spots from 14 individuals were measured after storage at room temperature for 1, 9 and 15 days. Stability was determined by comparison of the calculat-

ACS Paragon Plus Environment

3

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ed cell number on day 1 to the calculated cell number on day 9 and day 15. Inter-individual variation. Inter-individual variation in the CD71, CD45, and CD41 protein concentration per cell was estimated by calculation of the LC-MS/MS response divided by the cell number (measured on the Sysmex XT 2000i) for each individual, n = 82, and calculation of the average and %relative standard deviation of these values. Inter-individual variation in band 3 was calculated using the same method from 78 individuals. Flow Cytometry. Flow cytometry analysis was performed in the Flow Cytometry Core Facility at the University of Utah School of Medicine. Twenty microliters of blood was diluted in 500 µl of Cell Buffer containing PBS, 0.5% bovine serum albumin and 2 mM EDTA. Anti-human FITC-CD71 antibody and anti-human PE-CD235a antibody were added, 1 µl of each per sample, and incubated in the dark for 30 min. The cells were then washed with 2 ml Cell Buffer and collected by centrifugation, 200 x g, 5 min. Cells were resuspended in Cell Buffer and 20 µl of AccuCount fluorescent counting beads were added. Cells positive for both CD71 and CD235a were counted using the BD FACSCanto analyzer and normalized to counting bead number for calculation of the CD71+ cells/µl. To calculate CD71+ cell numbers in DBS samples, a 3-point calibration curve was made at HCT 35%, 45%, and 55% and the CD71+ cells were counted by flow cytometry prior to DBS spotting and extraction. Table 1. Method Validation Results

Band 3 CD71

HCT Range (%)

Linearity (r2)

30 - 60

0.982

30 - 60

0.985

Intra-assay Precision (CV%) Ave (*Range) 7.1 (11.1 – 10.9) 10.7 (9.2 – 15.9)

Matrix Interference % Ave (*Range) 53 (47-58) 22 (17-24)

CD45

Page 4 of 9 30 - 60

0.906

9.3 (11.5 – 14.0) CD41 30 - 60 0.984 7.4 (12.1 – 10.9) *Range represents the values for 30% and 60% HCT.

31 (31-35) 33 (31-37)

RESULTS AND DISCUSSION Figure 1A is a schematic outline of the method used to count cell numbers in DBS. Cell number was determined by measurement of band 3 (CD233) for RBC, CD71 for IRC, CD45 for WBC, and CD41 for platelets. Since these proteins are integral membrane proteins, the DBS spots were extracted in a series of 3 buffers that remove soluble proteins30,31 while leaving cell membranes on the DBS spot. After extraction, the membrane proteins in the spot were digested with trypsin and the released peptides were quantified by high resolution accurate mass spectrometry. One peptide and the corresponding peptide internal standard were measured for each protein. Since CD proteins often contain a large soluble domain, it was possible to obtain efficient digestion using short digestion times and in the absence of denaturants and reducing/alkylation agents. Digestion efficiency is shown in Figure S-2, Supporting Information. Cell number was calculated using a calibration curve made from blood spots at HCT 35%, 45% and 55% whose cell numbers were determined on the Sysmex. The chromatogram in Figure 1A shows the extracted product ion chromatogram for each endogenous peptide measured in the DBS spot. Recovery on DBS cards. Since the spots are extracted in buffer three times prior to digestion, it was important to determine the recovery of the membrane proteins on the spot. Additionally, it was possible that the recovery may vary depending upon the cell type and lipid/protein contents of the plasma membrane. Recovery of the membrane proteins on the DBS spots after the 3 extraction steps was 88%, 74%, 62%, and 81% for band 3, CD71, CD45, and CD41,

Table 2. Deming regression and Bland-Altman analysis Deming Regression

Cell Type/ Protein

RBC/ band3 WBC/ CD45 Platelet/ CD41 IRC (Sysmex)/ CD71 CD71+(Flow)/ CD71

Sample number

r2

78 82 82 82 31

0.71 0.93 0.76 0.77 0.91

Slope

1.006 1.193 0.852 1.153 0.954

Intercept

0.5 -116.3 30.4 -88.8 28.8

Bland-Altman %Relative Difference Mean Difference (%) 0.5 -5.2 -3.2 -4.6 11.4

ACS Paragon Plus Environment

InterIndividual Variation

95% Limit of Agreement (%)

(%RSD)

-11.7 – 12.6 -35.6 – 25.2 -53.2 – 46.9 -81.4 – 72.3 -38.1 – 60.9

6 8 18 78 33

4

Page 5 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Figure 2. Deming regression graphs comparing DBS cell number to Sysmex cell number for RBCs (A), IRCs (B), WBCs (C), and Platelets (D).

respectively, Figure 1B (black bars). It should be noted that band 3 is a 12-transmembrane domain protein, while CD71, CD45 and CD41 are single transmembrane domain proteins, but this difference did not appear to affect the recovery. The data suggests that the Whatman filter paper provides a simple matrix for quantitative recovery and measurement of membrane proteins in blood. Even when a standard liquid blood sample is available, washing of blood cells in-solution to remove matrix results in loss of membranes, which are not fully recovered without ultracentrifugation.30,31 Figure 1B (striped bars) shows the %recovery of the same membrane proteins in liquid blood when washed three times in-solution. The recovery was 49%, 43%, 14% and 40% for band 3, CD71, CD45, and CD41, respectively. Recovery of WBC membranes containing CD45, was lower, at 14%, than other cell types, suggesting that the recovery in liquid blood may vary based upon the content of the membrane for that cell type. Since significantly better recovery was observed on DBS membranes, it is possible that the DBS extraction method may have utility even when liquid blood is available. Removal of soluble proteins. CD71 is an established marker for IRCs measured by flow cytometry and is likewise used in this method to measure the same cell type. However, in vivo, the CD71 protein is cleaved by a protease on the extracellular side to release soluble transferrin receptor (sTFR) into the circulation. Thus, it is possible that sTFR in the plasma fraction of the DBS spot could interfere with measurement of CD71 protein if it were not completely removed prior to sample digestion. To measure recovery of sTFR remaining after spot extraction, recombinant sTFR was added to blood either before spotting onto DBS cards or after spot extraction and just prior to digestion. The fortified sTFR concentrations, 100 nM, 250 nM and 500 nM, represent a 4 – 20-fold excess

of endogenous concentrations. As shown in Figure 1C, recovery of sTFR remains constant at ≤ 2.5%, with increasing concentration, thus indicating efficient removal of the soluble protein. Stability. In the context of anti-doping, one significant advantage of blood collection on DBS cards is the low cost of sample shipment and enhanced sample stability which will allow increased testing, even in remote regions where athletes often reside and train. Blood collected for the current WADA ABP program must comply with specific shipping temperatures and a collection-to-analysis time of less than 72 hours.32 Thus, it was important to determine the shipment stability of the blood cell proteins on the DBS cards. Stability of the proteins in 14 individual DBS samples was determined after storage at room temperature for 1, 9 and 15 days. As shown in Figure 1D, the calculated cell numbers at day 9 were 99%, 97%, 118% and 98% of that calculated on day 1 for RBC, IRC, WBC, and platelets, respectively. On day 15, the calculated cells numbers were 92%, 76%, 119%, and 67% of that calculated on day 1 for RBC, IRC, WBC and platelets, respectively. Using a guideline of bias ≤ 15%, the proteins are stable for between 9-15 days, which is more than sufficient to receive cards shipped from any location. Since blood samples in the ABP program are measured immediately upon receipt for longitudinal data analysis, there is no need for retrospective analysis and long term stability was not determined. Method Validation. Method validation results are summarized in Table 1 and in the Supplemental Information. Linearity of the method was measured using a 7-point curve of DBS with increasing HCT% and cell number. The peak area response for each peptide was linear with increasing cell number, r2 ≥ 0.901. Intra-assay precision was tested at each hematocrit level and was ≤15.9%. The range of the intra-assay

ACS Paragon Plus Environment

5

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

precision demonstrates that hematocrit level did not affect precision. Matrix interference was largest for the band 3 peptide. The range of matrix interference for each peptide demonstrates that it was not substantially affected by hematocrit. Selectivity was determined in rat and chicken blood DBS spots because the peptide sequences differ from the human sequences for each protein. There were no significant interfering peaks observed at the retention time for each peptide. Selectivity is shown in the extracted product ion chromatograms in Figure S-1, Supporting Information. Data for interday precision and the digestion time course are also shown in Supporting Information. The method was designed for excision of the full 20 µl spot, which requires volumetric application of the blood sample. This may be performed by several methods including a recently described, microfluidic-based DBS device.33 Excision of a partial spot for use with uncontrolled blood drop application was not tested, but would likely have error resulting from the volumetric differences that occur at different hematocrit levels, i.e. the hematocrit effect. Although it should not affect percentage values, such as IRC%, total cell numbers would be affected. For partial spot excision, use of hematocrit correction methods may correct the error, as previously reported.34,35 Comparison of DBS Cell Number to Sysmex Values. The calculated cell numbers using the DBS method were compared to the cell numbers using the Sysmex XT 2000i, as mandated by WADA. Figure 2 shows the Deming regression analysis and the statistics are listed in Table 2. For each cell type, the DBS method demonstrated good correlation with the Sysmex cell numbers: RBC slope = 1.006, r2 = 0.71; IRC slope = 1.153, r2 = 0.77; WBC slope = 1.193, r2 = 0.93; and platelets slope = 0.852, r2 = 0.76. Using this method, a calibration curve made from one individual’s blood was used to calculate cell number in blood from many different individuals. Thus, it was important to estimate the inter-individual variation in the average concentration of CD protein per cell. As shown in Table 2, band 3 and CD45 have very low interindividual variation, 6% and 8%, respectively. CD41 has higher variation, at 18%, which is consistent with a previous report that calculated the inter-individual variation of CD41 at 24%.36 Finally, CD71 had a very high variation of 78%, which is expected given its regulation during reticulocyte maturation. The trend in inter-individual variation agreed well with the Bland-Altman results, listed in Table 2 and shown in Supporting Information Figure S-3. The % mean difference was low for each cell type, RBC = 0.5%, IRC = -4.6%, WBC = -5.2%, and platelets = -3.2%. However, the 95% limits of agreement were better for the CD proteins with low interindividual variation, band 3 and CD45. Comparison of DBS Cell Number to Flow Cytometry Values. In the case of IRC measurement, it is important to note that automated hematology analyzers, such as the Sysmex XT 2000i, count immature reticulocytes as the population of erythrocytes with the highest intensity of RNA staining.20 While reticulocyte RNA content correlates well with CD71 expression, the two cell populations are not identical. To obtain a more direct comparison of CD71 cell number, the population of CD71+ and CD235a+ cells was counted by flow cytometry in blood collected from 31 individuals and this was compared to the DBS CD71+ cell results. As shown in Figure 3 and Table 2, correlation of the DBS CD71+ cell number with the flow cytometry cell number was good with Deming slope = 0.954 and r2 = 0.91. The Bland-Atlman % mean dif-

Page 6 of 9

ference was 11.4% and the 95% limit of agreement was 38.1% - 60.9%. While the agreement of the DBS CD71+ cell number with flow cytometry values is good, it does not fully account for the cell-to-cell variability in CD71 expression. Since the CD71 concentration per cell declines during reticulocyte maturation, the average CD71 concentration measured in DBS is dependent not only on cell number but also on the average maturity of the reticulocyte population. Thus, an alternative calibration curve was tested for CD71. This curve consisted of rat blood DBS spots that were extracted with each batch and were fortified with recombinant sTFR/CD71, at three concentrations, just prior to digestion. The curve allowed conversion of the DBS peak area ratios to CD71 concentration rather than to IRC number. Additionally, the new curve may provide a more stable calibration reference over time. Since many CD proteins contain a large soluble domain, like CD71, it may be possible to produce recombinant standards of these soluble domains for use in calibration curves. The stability of CD71 measurements using the new calibration curve was tested in the longitudinal experiment discussed below.

Figure 3. Comparison of DBS cell number to flow cytometry cell number by Deming regression (top) and Bland-Altman (bottom).

Longitudinal Measurements. Successful detection of blood doping practices relies on stable, longitudinal measurement of an athlete’s blood parameters. The data for a single athlete’s longitudinal profile may consist of measurements collected from 34 different WADA accredited labs. While each laboratory is required to use the same Sysmex instru-

ACS Paragon Plus Environment

6

Page 7 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

ment, the inter-laboratory imprecision in the reticulocyte% values must be very low and this variability has been problematic.37 Additionally, since it was not previously possible to monitor average CD71 concentrations over time, it was unclear if this value would be a stable biomarker. Thus, the DBS method was tested by longitudinal measurement of CD71 over an 8-week period in 6 individuals. As shown in Figure 4, the average CD71 concentration in DBS spots ranged from 0.9 – 4.6 nM across 6 individuals and the intra-individual variation over 8 weeks was between 17.1% – 38.7%. The data suggest that longitudinal measurement of the average CD71 concentration in DBS may provide a stable parameter to monitor over time.

Figure 4. Longitudinal measurement of CD71 concentration in DBS with table showing the values for each subject.

CONCLUSIONS In this report, we describe the first published method to quantitatively measure CD proteins and count cell types in DBS. The DBS method represents a new way of measuring membrane proteins on cells, which may have several applications. Since CD proteins often contain large soluble domains, it may be possible to produce recombinant forms of the soluble domains for use as calibrators, as was demonstrated with CD71/sTFR. Likewise, the soluble domain of band 3 was recently obtained from a commercial source and will be tested in future experiments as a calibrator. When used in combination with recombinant protein calibrators, the DBS method can measure the average molar concentration of CD proteins on cells, which is difficult by other methods. The concentration data may be compared between experiments over time or between labs, which is difficult by flow cytometry. One potential application includes quantification of band 3 for the diagnosis of hereditary spherocytosis, which is currently performed, qualitatively, by flow cytometry.38 Additionally, the DBS method may be used in place of a semi-quantitative flow cytometry method reported to measure CD71 as a marker of iron status in patients with chronic renal failure.39 When used in combination with blood counts from a hematology analyzer, the DBS method is able to measure the inter-individual variation of CD protein per cell, which is difficult by other methods. However, an estimation of the inter-individual variation for platelet proteins, such as CD41, was previously reported

using spectral counting, an alternative mass spectrometry method.36 We demonstrate that the DBS method has good precision, and linearity, r2 ≥ 0.906, between 30% - 60% HCT. Even after 3 washing steps, recovery of membrane proteins on the spot was between 62%-88% for the four proteins measured. The demonstrated removal of recombinant sTFR to ≤ 2.5% after extraction demonstrates that sTFR does not interfere with measurement of membrane-associated CD71. Finally, the proteins were stable at room temperature for 9-15 days, which is more than sufficient for shipment of DBS cards from any location. Cell numbers calculated by the DBS method using a 3-point calibration curve demonstrated good agreement with Sysmex values, which is the instrument currently required for the WADA passport. It should be noted that the DBS method measures the average CD protein concentration over the entire cell population. Thus, the method will most accurately reflect cell number when there is low cell-to-cell variation and low inter-individual variation in the CD protein concentration per cell. This is the case for band 3, CD45, and CD41. Since CD71 protein concentrations decline during reticulocyte maturation and the inter-individual variation is higher, it is more accurate to use a calibration curve fortified with recombinant sTFR which allows calculation of CD71 protein concentration rather than cell number. For detection of blood doping practices in an anti-doping setting, the DBS method can significantly improve the current WADA passport. Venous blood collections are expensive and difficult to collect and ship to WADA accredited laboratories world-wide, especially, from the African continent. Use of the DBS method can circumvent these issues and may provide improved precision for longitudinal measurement. Measurement of a marker for IRCs (CD71) rather than reticulocyte% may also increase the sensitivity of the test, since it is more responsive to changes in erythropoiesis.22,23 However, CD71 is also increased in iron deficient anemia 40, which may be observed, particularly in female marathon runners, elite runners, and triathletes.41,42 Concomitant measurement of hemoglobin concentrations in DBS will aid in detection of anemic athletes. Additional testing in controlled human administration studies is underway to determine the utility of CD71 and the new DBS method for detection of blood doping practices.

ASSOCIATED CONTENT Supporting Information Additional material as noted in the text. The Supporting Information is available free of charge on the ACS Publications website. Materials list, method validation tables and figures, and BlandAltman plot figure. (PDF)

AUTHOR INFORMATION Corresponding Author * E-mail: [email protected]. Phone: 801-994-9454

Author Contributions The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

Notes The authors declare no competing financial interests.

ACS Paragon Plus Environment

7

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACKNOWLEDGMENTS This work was supported by the Partnership for Clean Competition, #404284R212. The authors would like to thank Dr. Larry Bowers for his critical insight and support of the project.

REFERENCES (1) Bidlingmaier, M. Clin Chem 2016, 62, 1296-1298. (2) Moller, I.; Thomas, A.; Geyer, H.; Schanzer, W.; Thevis, M. Anal Bioanal Chem 2012, 403, 2715-2724. (3) Ryona, I.; Henion, J. Anal Chem 2016, 88, 11229-11237. (4) Thomas, A.; Geyer, H.; Schanzer, W.; Crone, C.; Kellmann, M.; Moehring, T.; Thevis, M. Anal Bioanal Chem 2012, 403, 1279-1289. (5) Ferro, P.; Ventura, R.; Perez-Mana, C.; Farre, M.; Segura, J. Drug Test Anal 2016. (6) Reverter-Branchat, G.; Bosch, J.; Vall, J.; Farre, M.; Papaseit, E.; Pichini, S.; Segura, J. Clin Chem 2016, 62, 1353-1360. (7) Cox, H. D.; Rampton, J.; Eichner, D. Anal Bioanal Chem 2013, 405, 1949-1958. (8) Jelkmann, W.; Lundby, C. Blood 2011, 118, 2395-2404. (9) Robinson, N.; Saugy, M.; Vernec, A.; Pierre-Edouard, S. Clin Chem 2011, 57, 830-832. (10) Pottgiesser, T.; Schumacher, Y. O. Bioanalysis 2012, 4, 1245-1253. (11) Gore, C. J.; Parisotto, R.; Ashenden, M. J.; Stray-Gundersen, J.; Sharpe, K.; Hopkins, W.; Emslie, K. R.; Howe, C.; Trout, G. J.; Kazlauskas, R.; Hahn, A. G. Haematologica 2003, 88, 333-344. (12) Damsgaard, R.; Munch, T.; Morkeberg, J.; Mortensen, S. P.; Gonzalez-Alonso, J. Haematologica 2006, 91, 1006-1008. (13) Russell, G.; Gore, C. J.; Ashenden, M. J.; Parisotto, R.; Hahn, A. G. Eur J Appl Physiol 2002, 86, 442-449. (14) Morkeberg, J.; Belhage, B.; Ashenden, M.; Borno, A.; Sharpe, K.; Dziegiel, M. H.; Damsgaard, R. Int J Sports Med 2009, 30, 285-292. (15) Leuenberger, N.; Barras, L.; Nicoli, R.; Robinson, N.; Baume, N.; Lion, N.; Barelli, S.; Tissot, J. D.; Saugy, M. Am J Hematol 2016, 91, 467-472. (16) Banfi, G. Clin Chem Lab Med 2011, 49, 1417-1421. (17) Bejder, J.; Aachmann-Andersen, N. J.; Bonne, T. C.; Olsen, N. V.; Nordsborg, N. B. Drug Test Anal 2016, 8, 1049-1055. (18) Zorzoli, M.; Rossi, F. Drug Test Anal 2012, 4, 854-858. (19) Malleret, B.; Xu, F.; Mohandas, N.; Suwanarusk, R.; Chu, C.; Leite, J. A.; Low, K.; Turner, C.; Sriprawat, K.; Zhang, R.; Bertrand, O.; Colin, Y.; Costa, F. T.; Ong, C. N.; Ng, M. L.; Lim, C. T.; Nosten, F.; Renia, L.; Russell, B. PLoS One 2013, 8, e76062. (20) Kono, M.; Kondo, T.; Takagi, Y.; Wada, A.; Fujimoto, K. Clin Chim Acta 2009, 404, 105-110. (21) Nunes, L. A.; Grotto, H. Z.; Brenzikofer, R.; Macedo, D. V. Biomed Res Int 2014, 2014, 349182. (22) Major, A.; Bauer, C.; Breymann, C.; Huch, A.; Huch, R. Br J Haematol 1994, 87, 605-608. (23) Biesma, D. H.; Kraaijenhagen, R. J.; Dalmulder, J.; Marx, J. J.; Van De Wiel, A. Br J Haematol 1994, 86, 30-35. (24) Bejder, J.; Hoffmann, M. F.; Ashenden, M.; Nordsborg, N. B.; Karstoft, K.; Morkeberg, J. Scand J Med Sci Sports 2016, 26, 338-347.

Page 8 of 9

(25) Zola, H.; Swart, B.; Nicholson, I.; Aasted, B.; Bensussan, A.; Boumsell, L.; Buckley, C.; Clark, G.; Drbal, K.; Engel, P.; Hart, D.; Horejsi, V.; Isacke, C.; Macardle, P.; Malavasi, F.; Mason, D.; Olive, D.; Saalmueller, A.; Schlossman, S. F.; Schwartz-Albiez, R., et al. Blood 2005, 106, 3123-3126. (26) Sleczka, B. G.; D'Arienzo, C. J.; Tymiak, A. A.; Olah, T. V. Bioanalysis 2012, 4, 29-40. (27) Chambers, A. G.; Percy, A. J.; Yang, J.; Borchers, C. H. Mol Cell Proteomics 2015, 14, 3094-3104. (28) Henderson, C. M.; Bollinger, J. G.; Becker, J. O.; Wallace, J. M.; Laha, T. J.; MacCoss, M. J.; Hoofnagle, A. N. Proteomics Clin Appl 2017. (29) Jung, S.; Whiteaker, J. R.; Zhao, L.; Yoo, H. W.; Paulovich, A. G.; Hahn, S. H. J Proteome Res 2017, 16, 862-871. (30) Fujiki, Y.; Hubbard, A. L.; Fowler, S.; Lazarow, P. B. J Cell Biol 1982, 93, 97-102. (31) Pasini, E. M.; Kirkegaard, M.; Mortensen, P.; Lutz, H. U.; Thomas, A. W.; Mann, M. Blood 2006, 108, 791-801. (32) Robinson, N.; Kirchbichler, A.; Banuls, O.; Mader, M.; Aikin, R.; Sottas, P. E.; D'Onofrio, G. Int J Lab Hematol 2016, 38, 685-693. (33) Verplaetse, R.; Henion, J. Anal Chem 2016, 88, 6789-6796. (34) Capiau, S.; Stove, V. V.; Lambert, W. E.; Stove, C. P. Anal Chem 2013, 85, 404-410. (35) Capiau, S.; Wilk, L. S.; Aalders, M. C.; Stove, C. P. Anal Chem 2016, 88, 6538-6546. (36) Burkhart, J. M.; Vaudel, M.; Gambaryan, S.; Radau, S.; Walter, U.; Martens, L.; Geiger, J.; Sickmann, A.; Zahedi, R. P. Blood 2012, 120, e73-82. (37) Ashenden, M.; Sharpe, K.; Habel, J.; Vicentin, A.; Schischka, S. J Clin Pathol 2013, 66, 232-237. (38) King, M. J.; Zanella, A. Int J Lab Hematol 2013, 35, 237-243. (39) Soininen, K.; Punnonen, K.; Matinlauri, I.; Karhapaa, P.; Rehu, M. Clin Chem Lab Med 2010, 48, 1239-1245. (40) Ervasti, M.; Matinlauri, I.; Punnonen, K. Clin Chim Acta 2007, 383, 153-157. (41) Bruinvels, G.; Burden, R.; Brown, N.; Richards, T.; Pedlar, C. PLoS One 2016, 11, e0149881. (42) Coates, A.; Mountjoy, M.; Burr, J. Clin J Sport Med 2016.

Insert Table of Contents artwork here

ACS Paragon Plus Environment

8

Page 9 of 9

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment

9