Abstract
Telehealth, accessing healthcare and wellness remotely, should be a cost-effective and efficient way for individuals to receive care. The convenience of having a reliable remote collection device for blood tests will facilitate access to precision medicine and healthcare. Herein, we tested a 60-biomarker health surveillance panel (HSP), containing 35 FDA/LDT assays and covering at least 14 pathological states, on 8 healthy individuals’ ability to collect their own capillary blood from a lancet finger prick and directly compared it to the traditional phlebotomist venous blood and plasma collection methods. All samples were spiked with 114 stable-isotope-labeled (SIL) HSP peptides and quantitatively analyzed by liquid chromatography-multiple reaction monitoring-mass spectrometry (LC/MRM-MS) scheduled method targeting 466 transitions from 114 HSP peptides and by a discovery data-independent acquisition mass spectrometry (DIA-MS) method. The average peak area ratio (PAR) of the HSP quantifier peptide transitions from all 8 volunteers’ capillary blood (n = 48), venous blood (n = 48), and matched plasma (n = 24) was <20% coefficients of variation (CV). Heat map analysis of all 8 volunteers demonstrated that each individual had a unique biosignature. Biological replicates from capillary blood and venous blood clustered within each volunteer in k-means clustering analysis. Pearson statistical analysis of the three biofluids indicated that there was >90% similarity. Discovery DIA-MS analysis of the same samples using a plasma spectral library and a pan-human spectral library identified 1121 and 4661 total proteins, respectively. In addition, at least 122 FDA-approved biomarkers were identified. DIA-MS analysis reproducibly quantitated (<30% CV) ∼600–700 proteins in capillary blood, ∼800 proteins in venous blood, and ∼300–400 proteins in plasma, demonstrating that an expansive biomarker panel is possible with current mass spectrometry technology. Both targeted LC/MRM-MS and discovery DIA-MS analysis of whole blood collected on remote sampling devices are viable options for personal proteome biosignature stratification in precision medicine and precision health.
Precision medicine consists of identifying the disease state of an individual so that an effective intervention or treatment can be prescribed to correct or lessen the deleterious effects of their phenotypic ailments.1 Precision health is focused on prevention, maintaining, and even optimizing health, which should be an effective way to intervene common sedentary and western diet-induced lifestyle metabolic diseases, including the onset of age-related diseases.2−4 More people are prioritizing and monitoring their personal health regime through an array of health strategies involving nutrition, physical fitness, healthy microbiome, wearable technology, and mental health well-being.5−7 Physicians, scientists, and individuals can become proactive in their self-health management, which requires appropriate assays beyond the conventional medically prescribed tests to quantitatively measure disease progression to provide actionable and personalized health insights. Both precision medicine and precision health have benefited from significant advancement of multiomics technologies that allow millions of quantitative and qualitative tests in genomics, epigenetics, transcriptomics, proteomics, and metabolomics, along with the development of powerful computational bioinformatic tools and machine learning algorithms to help understand an individual’s unique phenotype so that more effective interventions may be prescribed, customized, and monitored for desired outcomes.8−10
The blood proteome is an informative source of biomarkers representing personalized signatures of physiological phenotype, by interacting and responding to the entire body’s cellular tissue disease or health state.11 The proteome also represents an individual’s genetic predisposition and continually responds to environmental factors and infectious agents as well as drug treatments, physical exercise,6 and nutritional intervention.7,12,13 In addition, the proteome’s unique isoforms, post-translational modifications (e.g., A1C),14 longer half-lives compared to metabolites, and shorter half-lives responsive to an acute stimulus (e.g., CRP)15 allow for specific and stable targeted diagnostic tests to be developed.22 There has been significant development in biomarker assays targeting the proteome for health or disease by companies such as Olink,16 Biognosys,17 SomaLogic,18 and Seer.19 Similarly, at Cedars-Sinai Medical Center, the Advanced Clinical Biosystems Research Institute (ACBRI) and Precision Biomarker Laboratories (PBL) have developed a 60-protein health surveillance panel (HSP) that includes 35 FDA and LDT protein clinical assays20 and covers 14 physiological diseases. The HSP multiplex is composed of 1–2 peptides per protein with a quantifier (used to calculate protein concentration) and when possible, a qualifier peptide (confirms protein result).21 The current data collected via liquid chromatography-multiple reaction monitoring-mass spectrometry (LC-MRM/MS)22 and discovery data-independent acquisition mass spectrometry (DIA-MS)23 analysis with stable isotopically labeled (SIL) peptides are encouraging for the future expansion to larger multiplex biomarker panels into the hundreds of proteins to provide a more comprehensive health and disease assessment than is available with current standard-of-care medicine.24
Most health assessment screens require multiple blood draws collected at a healthcare facility by a trained phlebotomist. Samples are then packed for transport and delivered to a laboratory service provider for sample processing and analysis. The benefits of having access to a safe, viable, and reliable remote collection device are convenient and cost-effective for patients, physicians, and clinics, allowing for an array of blood tests.25 Biomarker analysis has primarily been assessed with serum or plasma, where all cellular mass, including red blood cells, white blood cells, and platelets are removed to reduce the complexity of the biofluid.11 We found that the Mitra device (Neoteryx), which absorbs 10 μL26 when collected correctly through its volumetric absorptive microsampling (VAMS) technology,27 facilitates self-administered blood collection from a finger lancet needle prick with highly reproducible replicate collection,25,28 as demonstrated with very low coefficients of variation (CV). The Mitra devices have been shown to be successfully utilized with many conventional diagnostic tests currently ordered by a physician and can be easily adapted for automation.29,30 The stability of whole blood on Mitra devices on a subset of proteins has also been assessed by multiple groups, including Van Eyk et al.30,31 The Mitra device allows testing of infectious diseases,32 small molecule drugs,33 therapeutics,34 remote patient monitoring,35 pediatric pharmacokinetic studies,36 hormones, and newborns, to name a few key areas of health. In addition, an inexpensive, easily self-assisted remote collection device will allow people in a healthcare system as well as people in underserved populations and remote regions of the world to have access to quality medical care, clinical grade assays, treatment monitoring, and ultimately help the democratization of healthcare.31,37,38
In this study, we show that the HSP multiplex assay quantitative results between capillary blood, venous blood, and plasma indicate the use of remote collection devices is a viable option for health analysis and that each individual person has a unique proteome biosignature. The convenience to patients, array of circulating tests, and sample processing efficiency on a Mitra device in addition to the technological advancements in mass spectrometry and bioinformatic tools to quantitively identify thousands of physiologically relevant proteins in whole blood may shift the paradigm away from analyzing plasma alone.
Experimental Section
Blood and Plasma Collection
Ethics Committee Cedars-Sinai Medical Center (CSMC) IRB No: STUDY00000621/MOD00004460 CSMC Date Effective: 9/23/2021 gave ethical approval for all human samples used in this study. A total of 8 volunteers signed a consent form to donate venous blood drawn by a phlebotomist at Cedars-Sinai Medical Center and collected with a 10 mL K2 EDTA vial. Plasma was isolated from the venous blood microtainer tubes (lavender, K2 EDTA). In addition, 10 μL of venous blood was pipetted onto six Mitra tips for each volunteer. Each volunteer also collected six Mitra device tips (10 μL) of their own capillary whole blood from a lancet finger prick. Mitra Device tips containing capillary whole blood, venous whole blood, and plasma (5.5 μL) were added to a 2.2 mL 96-well plate according to the experiment plate layout. As quality controls, commercially purchased pooled plasma (pool4), capillary blood, venous blood, and pooled plasma from all 8 volunteers were used in triplicate on each 96-well plate.
Sample Processing
A volume of 100 μL (94.5 μL to plasma) of denaturing buffer consisting of 67% 100 mM ammonium bicarbonate (AmBic) and 33% 2,2,2-trifluoroethanol (TFE) was added to each Mitra tip in its respective well. The 96-well plate was sonicated for 15 min at RT, placed on a shaker at 1000 rpm for 30 min, and centrifuged at 1000 × g for 1 min. Samples were cysteine-reduced by adding 10 μL of 55 mM TCEP to the sample(s) while shaken at 1000 RPM for 30 min. Samples were alkylated with 10 μL of 168 mM iodoacetamide (IAM) and quenched with 10 μL of 65 mM dithiothreitol (DTT). Samples were diluted with 870 μL of 100 mM AmBic (final 5% TFE concentration) before adding 10 μL of 1 μg/μL trypsin and incubated at 37 °C for 16 h at 300 rpm. After digestion, 144.3 μL (50 μg of plasma) was pipetted into the 96-well plate, where 855.7 μL of H2O was added to bring the volume up to 1 mL. A total of 114 stable isotopically labeled (SIL) peptides representing all 60 HSP proteins were spiked into each sample solution at 20 μL of 100 pmol/mL (2000 fmol/50 μg of plasma protein). A duplicate calibration curve was also added to each MS plate from a dilution series of HSP SILs spiked into pooled trypsin-digested capillary whole blood from all 8 volunteers. The digested peptides were acidified, desalted on a Phenomenex Strata X-PRO desalting plate, dried in a TurboVap (flow rate = 40), solubilized with 100 μL of 96/4 (v/v) water/acetonitrile and 0.2% formic acid, sealed with foil, sonicated for 5 min, shaken at 1000 RPM for 10 min, and then centrifuged 1 min at 1000 × g. Resolubilized samples were added at 45 μL to two identical mass spectrometry (MS) plates (PCR Plates 96-wells and Slit Seal) according to the MS plate layout. The remaining 10 μL of each sample was pooled, mixed, and pipetted into a LC sample vial for LC/MRM-MS retention time correction (RTC, where scheduled retention times are determined and corrected to current values).
HPLC and Mass Spectrometry
LC/MRM-MS was performed on two harmonized39 systems consisting of identically configured Thermo Fisher Ultimate 3000 high-flow LC connected to a Sciex 6500+ triple quadrupole mass spectrometer. A sample list was created in Chromeleon software matched to the sample list in Sciex 6500+ Analyst software and the targeted HSP peptides and transitions method with the RTC was loaded. Each sample was injected at 15 μL (equivalent to 7.5 μg plasma and 300 fmol SIL) onto a Waters XBridge Peptide BEH C18 Column, 130 Å, 3.5 μm, 2.1 mm × 100 mm. The HPLC gradient duration was 30 min starting at 0 min 5% B 0.500 mL/min; 0.800 min 5% B to 25 min 30% B at 0.500 mL/min; 25.4 min 90% B at 0.650 mL/min; 25.7 min 30%B at 0.650 mL/min; 26.1 min 90%B at 0.650 mL/min; 26.4 min 0% B at 0.650 mL/min; 26.8 min 90% B at 0.650 mL/min; 27 min 5% B at 0.500 mL/min; 30 min stop run (mobile phase A: 98.8% H2O + 0.2% formic acid and mobile phase B: 100% acetonitrile + 0.2% formic acid). LC/MRM-MS data processing and statistical analysis were executed using SCIEX OS (version 2.1.0.55343) and PBL’s in-house ProEpic processing platform. All LC/MRM-MS raw files have been deposited to the PASSEL PeptideAtlas repository with the data set identifier PASS03796.
Data-independent acquisition mass spectrometry (DIA-MS) analysis was performed on a Thermo UltiMate 3000 LC system coupled to a Thermo Orbitrap Exploris 480. Samples were full-loop injection-loaded onto the C18 trap column (10 mm × 0.3 mm) at 15 μL/min buffer A (H2O and 0.1% formic acid) for 3 min, the valve configuration was switched to the analytical column (Phenomenex 15 cm × 300 μm, 100 Å C18 beads), and peptides were separated over a 60 min gradient at a flow rate of 9.5 μL/min, with increased buffer B (acetonitrile with 0.1% formic) from 2 to 5% in 2.5 min, then to 9% at 6 min, 27% at 45 min, and 44% at 60 min. A nanoelectrospray source with a Newomics 8-nozzle emitter (10 μm internal diameter) was used with a spray voltage of 3800 kV and capillary temperature of 320 °C. Full MS scans were acquired at 60k orbitrap resolution from 400–1000 Da with a normalized AGC target of 300% and a maximum injection time of 25 ms. The MS2 was performed over a mass range of 400–1000 Da using 12 Da windows with an overlap of 1 Da. Peptides were fragmented with a normalized collision energy of 28% in positive polarity; spectra were collected at 15k resolution in the orbitrap from 200–2000 Da with a normalized AGC target of 1000% and maximum injection time of 22 ms. All DIA-MS raw files and search results have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the data set identifier PXD038678.
Data Analysis and Statistics
PBL in-house ProEpic platform was used to analyze raw data, where raw data files information on precursor m/z, product m/z, RT, and component name from Analyst are converted (MS Convert) to mzML files that run through SmartPeak software. A custom feature for QC component rules is uploaded specifying 15 s width at 50%, and a minimum of 6 points across half height was used for all 466 transitions. The MRM report is ranked for quantifier and qualifier peptides reporting peak areas and calculating peak area ratios from endogenous peak area divided by SIL peak area, and calculating % CV. To assess the linear relationship and magnitude of association for protein intensities across biofluids, Pearson’s correlation coefficient (PCC) was calculated using the R programming language (cor.test in stats package). Average intensities for each volunteer were compared across three biofluids to produce p-value for each comparison in addition to Pearson’s r. The interpretation of the p-value indicates a statistically significant correlation between the biofluids for a given protein. A calculated p-value <0.05 equates to a statistically significant positive linear relationship (strong association) between two biofluids. Proteins with significant p-values across all biofluid comparisons are indicative of potentially robust biomarkers as they produce more consistent correlation of intensities.
Reagents and Materials
Please see the Supporting Information.
Results
Sample Collection and LC/MRM-MS Targeted Analysis
We tested an internally developed 60-protein HSP multiplex assay on 8 volunteers consisting of 5 males and 3 females of ages 27–60. Volunteers were required to collect their own capillary whole blood from a lancet finger prick using Mitra devices, which was compared to the traditional phlebotomy method of venous whole blood collection and plasma isolation; all samples were collected on the same day. All samples from the three biofluids were analyzed by both targeted LC/MRM-MS and DIA-MS analysis (Figure 1A,B). Each volunteer collected six Mitra devices of a capillary blood drop (10 μL) from a lancet finger prick that was matched to 10 μL of venous whole blood pipetted onto six Mitra devices on the same day. According to the American Red Cross, blood is estimated to be 55% plasma; we used 5.5 μL of matched isolated plasma for in-solution digestion for a theoretical 1:1:1 (capillary/venous/plasma) comparison (Figure 1A). A buffer of 33% TFE in 100 mM ammonium bicarbonate was determined internally to yield the best results. Briefly, Mitra tips were removed and placed into 96-well plates (Figures 1A and S1), trypsin-digested, and a stock containing 40 fmol of each stable isotopically labeled (SIL, Table S1), representing each targeted peptide was spiked into the solution for every 1 μg of plasma protein. The equivalent of 7.5 μg of plasma protein and 300 fmol of SIL was loaded onto the column and analyzed by targeted LC/MRM-MS using a 3000 Ultimate HPLC interfaced with a Sciex 6500+ triple quadrupole mass spectrometer (Figure 1B).
Figure 1.
Health surveillance panel (HSP) workflow overview. (A) Outline of capillary blood (n = 6) and venous blood (n = 6) Mitra device collection matched to plasma (n = 3) from each of the 8 volunteers. (B) Outline of sample processing to analysis on the LC/MRM-MS targeted platform and on the DIA-MS discovery platform.
MS raw data were normalized and searched by using our in-house ProEpic software platform, with SmartPeak40 used to identify the quality of the peaks. ProEpic software-calculated peak area ratio (PAR = endogenous peptide/SIL) for all 114 peptides representing 60 HSP biomarker proteins across all 8 volunteers’ capillary blood (each volunteer n = 6, total 48 samples), venous blood (each volunteer n = 6, total 48 samples), and plasma samples (each volunteer n = 3, total 24 samples). The % CV ((standard deviation/mean) × 100) was calculated from across each volunteer’s 6 replicates for the quantifier peptides (Table S2). For simplicity, we compared the biological replicate averages of the quantifier peptide transitions of all HSP proteins between the capillary blood, venous blood, and matched plasma and all were below 20% CV (Figure 2A).
Figure 2.
Comparison of each HSP biomarker’s coefficient of variation (CV) and peak area ratio (PAR) in capillary blood, venous blood, and plasma. (A) Average % CV calculated for each HSP biomarker measured from all 8 volunteers’ replicate samples of capillary blood (n = 48), venous blood (n = 48), and matching plasma (n = 24) was <20% CV in all three biofluids. (B) Comparison of each HSP biomarker’s calculated PAR averages from all 8 volunteers’ replicate samples of capillary blood (n = 48), venous blood (n = 48), and matched plasma (n = 24).
The average PAR of all 8 volunteers’ biological replicates was compared to determine if there were any major differences between the three biofluid collection methods (Figure 2B, raw data in Table S2). The majority of HSP proteins demonstrated a relatively similar PAR intensity between the biofluids. The lower abundant HSP proteins were highlighted in a zoomed graph so that the relative quantities could be seen between biofluids (Figure 2B).
Interestingly, each volunteer had a unique PAR intensity biosignature pattern for many of the HSP proteins (e.g., CRP, VWF, SHBG, APOE, IGHA1, ITIH1, HP, CLUS, and GSN, to highlight a few), which were consistent across each biofluid of capillary blood, venous blood, and plasma (Figure 3). One of the major benefits of collecting whole blood on Mitra devices is that differential hemolysis due to phlebotomist collection and plasma isolation is avoided for the red blood cell-specific proteins HBA and PRDX2 (Figure 3). This was demonstrated in each volunteer having a similar PAR intensity biosignature for HBA and PRDX2 in both capillary and venous blood. When we looked at the levels of HBA and PRDX2 in plasma for each volunteer, the PAR did not match either the capillary or venous pattern; however, the HBA pattern matched the PRDX2 PAR pattern for each volunteer in the plasma samples, indicating inadvertent hemolysis typically seen in plasma collection. This observation demonstrates that even a small amount of red blood cell perturbation results in significant differences in plasma levels related to mechanical fractionation of plasma rather than biological differences.
Figure 3.
Each volunteer has a unique HSP protein biosignature that is similar across three biofluids. Comparison of PAR averages (capillary n = 6, venous n = 6, plasma n = 3) for each volunteer for C-reactive protein (CRP), von Willebrand factor (VWF), sex hormone-binding globulin (SHBG), apolipoprotein E (APOE), immunoglobin heavy constant α 1 (IGHA1), inter-α-trypsin inhibitor heavy chain H (ITIH1), haptoglobin (HP), clusterin (CLUS), and gelsolin (GSN).
Statistical Analysis of Similarity of HSP Proteins in Capillary Blood, Venous Blood, and Plasma
The PAR data of all 8 volunteers, including all of their biological replicates, were analyzed by several statistical techniques to determine if there was any inconsistent variance from the other volunteers when compared to the different biofluid collection methods. In Figure S4, the Pearson correlation coefficient is plotted against p-values for all of the HSP proteins to demonstrate the similarity of most HSP proteins in each biofluid comparison and highlights the proteins that are significantly different.
Protein Biosignatures of 8 Healthy/Normal Individuals
The multiplex HSP protein biomarker data was used to analyze each volunteer’s set of biological replicates within each biofluid of capillary blood (n = 6), venous blood (n = 6), and matched plasma (n = 3) in comparison to all of the volunteers. The heat map analysis demonstrated that each individual volunteer had a unique biosignature of upregulated and downregulated proteins in each biofluid (Figure 4). Three of the volunteers (5, 6, and 7) had a number of inflammatory proteins including CRP upregulated. Interestingly, volunteer 5′s results correlated with a same-day blood analysis of Covid-19 IgG antibodies and N capsid antibodies that indicated a past infection (Table S5).
Figure 4.
Each volunteer has a unique heat map biomarker biosignature across capillary blood, venous blood, and plasma. Hierarchical clustering heat map distribution of HSP proteins from 8 volunteers’ capillary blood, venous blood, and plasma. The averages of replicates (capillary blood n = 6, venous blood n = 6, plasma n = 3) were grouped for each volunteer using the Euclidean distance measure and Ward clustering from T-test/NOVA calculations.
Since we did not have an established disease cohort to compare within this study, we further tested if replicates from capillary blood collected on Mitra devices (n = 6) would cluster with replicates from venous blood (n = 6) collected on Mitra devices (Figure 5). For volunteer 5, capillary and venous blood replicates all clustered together in Cluster 1 distinctly away from the other volunteers (Figure 5). In addition, for volunteer 6 and volunteer 7, who have respectively reported a history of hypertension and inflammation, biological replicates from their capillary blood and venous blood both clustered and overlapped in Cluster 2 and Cluster 6. While an active healthy male volunteer 2′s biological replicates clustered in Cluster 8, and two other healthy males, volunteer 1 in Cluster 4 and volunteer 8 in Cluster 7 overlapped, as did two healthy females, volunteer 3 and volunteer 4 clustered near each other encompassed by the red dotted line. Dendrogram hierarchical clustering of all volunteers’ biological replicates demonstrated the replicates clustered for each volunteer in both their capillary blood and venous blood (Figure S4).
Figure 5.
Each volunteer’s capillary blood and venous blood similarly clustered into distinct biosignatures. K-means clustering of the 8 volunteers’ biological replicate samples from both capillary blood (n = 6) and venous blood (n = 6). Gender and general health status listed with each volunteer.
HSP Physiological Concentration in Capillary Blood, Venous Blood’ and Plasma
To determine if the volunteers’ physiological endogenous HSP proteins were within the linear range of calibration, SIL peptides were spiked in duplicate into pooled trypsin-digested capillary whole blood from all 8 volunteers to create calibration curves on both sample plate 1 and plate 2 and then analyzed on two harmonized LC/MRM-MS systems. Linear regression was used to calculate the slopes for each transition (Table S6). Since the 300 fmol of SIL spiked into each sample is in the linear range of the calibration curve for all HSP proteins, the 300 fmol was used to calculate the endogenous HSP protein concentration, by multiplying each HSP protein’s PAR (peptide/SIL) by 300 fmol (Table S7). Interestingly, the majority of the volunteers’ HSP proteins were within published physiological ranges found in the Plasma Proteome Database (http://www.plasmaproteomedatabase.org).43,41,42 The calculated concentration of each HSP protein for all 8 volunteers was plotted to visualize the dynamic range comparison between capillary blood, venous blood, and plasma (Figure 6).
Figure 6.
Comparison of HSP biomarker’s calculated physiological concentrations in capillary blood, venous blood, and plasma. The average concentration (mg/L) of each HSP protein was calculated for all 8 volunteers’ replicates in capillary blood (n = 48, blue), venous blood (n = 48, orange), and plasma (n = 24, gray). Elevation of CRP (mg/L), indicating inflammatory response, in volunteers 5, 6, and 7 capillary blood, venous blood, and plasma is highlighted in a magnified inset graph. A whisker and box plot of the albumin concentration (g/mL) for capillary blood, venous blood, and plasma is inset into the graph.
In addition, the CRP (mg/L) concentrations for each volunteer’s biological replicates in capillary, venous, and plasma were calculated and plotted in a box and whisker graph to compare the detected levels within each biofluid (Figure 6). Interestingly, the three volunteers, 5, 6, and 7, all had elevated physiological levels approaching 10 mg/L, indicating mild inflammation. This CRP elevated response correlated with perturbances in inflammatory proteins. All other healthy volunteers had normal CRP levels (about 0.5–1 mg/mL). The detection of elevated CRP in volunteers assumed normal/healthy demonstrates its utility as a biomarker within the health surveillance panel for detecting inflammatory responses. Since ALB is the most abundant protein in blood (∼50% of plasma proteins), we compared the calculated concentration for all 8 volunteers’ replicates between the three biofluids in a box and whisker graph (Figure 6).
Data-Independent Analysis (DIA) of Capillary Blood, Venous Blood, and Plasma
Duplicate sample plates were analyzed via the discovery DIA-MS platform using the Exploris 480 mass spectrometer (Figure 1B) to determine if the HSP proteins could also be identified in an unbiased manner in whole blood samples. Using the twin plasma spectral library,44 1121 total unambiguous proteins were identified. While using the pan-human spectral library,45 which incorporated many of the proteins found in whole blood cells, a total of 3811 unambiguous proteins at 1% FDR (Table S10) were identified (4661 proteins were identified when FDR was set to 5%, Table S11). There are significantly more proteins identified in all three biofluids using the pan-human spectral library. In addition, many other biologically relevant proteins were identified in the whole blood, presumably from red blood cells, white blood cells, platelets, etc. that are typically too low abundant in plasma alone.
The overlap of the total identified proteins in the three biofluids was compared in Venn diagrams for the twin library and pan-human library (1% FDR and 5% FDR, Figure 7A). The DIA-MS analysis using the pan-human library at 1% FDR resulted in the shared identification of 2944 proteins between capillary blood (93%) and venous blood (94%), while plasma alone had 451 (25%) unique proteins identified, indicating that although plasma shared 1172 identified proteins with both capillary blood and venous blood, as expected, plasma and whole blood (∼55% plasma) had distinct proteomes due to the cellular differences of each biofluid. Using the data from the pan-human library search with the stringent 1% FDR filter, the identified proteins with CV < 10%, CV < 20%, and CV < 30% were compiled into stacked columns for each individual volunteer into three separate charts for capillary blood, venous blood, and plasma (Figure 7B) in order to visualize the upper limits of reproducible relative quantitation for each biofluid. The DIA-MS analysis is able to reproducibly quantitate ∼600–700 proteins in capillary blood, ∼800 proteins in venous blood, and ∼300–400 proteins in plasma. Heat map analysis of each volunteer’s capillary blood proteome (<20% CV) demonstrated each individual has a unique proteomic biosignature. The DIA analysis also demonstrated reproducible quantitation of APOE, IGHA1, and ITIH1 (n = 3) in each biofluid, and each volunteer had a unique biosignature pattern (Figure 7C) similar to that seen in the LC/MRM-MS targeted analysis (Figure 4).
Figure 7.
Data-independent acquisition (DIA) of all 8 volunteers’ capillary blood, venous blood, and plasma. (A) Venn diagram of total proteins identified in capillary blood, venous blood, and plasma from twin plasma spectral library search and pan-human spectral library search at 1% FDR and 5% FDR. (B) Column graph of coefficient of variance of proteins identified by DIA-MS in 3 biological replicates for each volunteer in capillary blood, venous blood, and plasma (volunteer 5 in capillary and plasma volunteer 4 is n = 2). Heat map analysis of all 8 volunteers’ capillary blood proteins indicates each individual has a unique proteomic biosignature. (C) Each volunteer has a unique HSP biosignature pattern across biofluids, as seen with APOE, IGHA1, and ITIH1 (n = 3).
The DIA-MS data also contained at least 122 FDA-approved biomarkers between all three biofluids (Figure 8). The average % CV across all 8 volunteers (3 biological replicates, n = 24) was calculated for each FDA-approved biomarker and plotted for each biofluid to observe which biomarkers in the DIA-MS analysis had the least variation (<10% CV, <20% CV, and <40% CV). Surprisingly, without SILs, the capillary blood, venous blood, and plasma, respectively, had 30, 38, and 36 FDA biomarkers <20% CV. In the future, choosing the most reliable peptides from each biomarker and creating matching SILs for targeted analysis will decrease the variance. A total of 83 FDA biomarkers were shared between the three biofluids, while 27 and 5 FDA biomarkers were seen in whole blood and plasma, respectively (Figure 8B Venn diagram). The quantitative reproducible identification of >600 proteins by discovery DIA-MS analysis of remote collection capillary whole blood indicates that even more proteins will be quantitatively detected with the addition of representative SIL peptides allowing for the development of a comprehensive proteomic biosignature assay.
Figure 8.
At least 122 FDA-approved biomarkers are identified in the DIA-MS analysis of the three biofluids. (A) Capillary blood, venous blood, and plasma contained 113, 108, and 105 FDA biomarkers, respectively. Each FDA biomarker is the calculated average % CV of all 8 volunteers (n = 24), and only the biomarkers with <50% CV were plotted. (B) Venn diagram of the FDA-approved biomarkers in capillary blood, venous blood, and plasma. All FDA biomarkers are listed for each biofluid.
Discussion
The combination of having a targeted 60-protein biomarker HSP multiplex assay and the convenience of a robust remote collection device will facilitate access to precision medicine and precision health. The HSP assay consisting of approximately 35 FDA and LDT protein biomarker clinical assays20 (Table S7) and representing 14 physiological diseases is an excellent starting point for covering a broad range of biologically relevant proteins to assess the overall health of an individual. We determined that there were quantitative similarities between the three biofluids analyzed with the HSP assay. The ability of an individual to lance their own finger and collect a drop of blood cannot be underestimated. Our preliminary studies indicated not only that people are squeamish about pricking their own finger but also are not able to reliably collect 60–100 μL of their own capillary blood required for some remote collection devices, while the traditional microtainers that require 250 μL of blood, pose even more of a challenge for self-collection. One major benefit of the Mitra device is that its VAMS technology allows for the accurate measurement of 10 μL of a single drop of blood, which resulted in an overall coefficient of variance for each HSP protein well below 20% CV, demonstrating the overall robustness of the HSP targeted assay with all and that large multiplex biomarker panels are possible with remote collection devices. Conducting a targeted analysis using SIL peptides for each protein corrects for any sample processing from the time of desalting, drying peptides down, resolubilizing, and any variability during the instrument analysis or from different LC/MRM-MS platforms. The <10% CV of four sets of QCs consisting of purchased plasma, pooled volunteer plasma, pooled peptides from capillary blood, and pooled peptides from venous blood analyzed at the beginning, middle, and end of each sample plate demonstrated the significant reproducibility of two harmonized LC/MRM-MS systems (Figure S2).
The majority of the HSP proteins had similar levels of each HSP protein across the three biofluids (Figure 2B). Of note, haptoglobin was higher in plasma than whole blood; however, the ratios of haptoglobin in each volunteer relative to the other volunteers followed the same pattern throughout the three biofluids (Figure 3). Subject-specific patterns for other HSP proteins are also seen throughout the biofluids, especially in CRP, VWF, SHBG, and IGHA1 as well as other HSP proteins, indicating that there were specific proteomic biosignatures in each “healthy/normal” individual that were not affected by the different biofluids. Pearson statistical analysis conducted on all volunteers and all biological replicates also supported the concept that the majority of HSP proteins have similar and reproducible biosignatures across biofluids. When capillary blood was directly compared to venous blood, there were only three proteins, C1R, SERPINF2, and SERPINC1, which were not statistically similar in the different collection methods, demonstrating that a remote collection device may be used to achieve overall similar results as a traditional phlebotomist blood draw. The variability in the SERPINs may be explained by the activation of blood clotting events in capillary blood during the finger lancet since these proteins are important in controlling proteolytic cascades in coagulation pathways. As expected, red blood cell proteins HBA and PRDX2 were found, respectively, in upward to 1000-fold and >200-fold higher concentrations in whole blood compared to plasma, which explains the significant difference in the Pearson statistical analysis. However, it remains to be seen if any of these few differences between biofluids would have an effect on differentiating an individual’s health within a defined disease group.
Hierarchical clustering heat maps indicated that each individual volunteer had their own unique biosignature compared to the other volunteers’ subject-specific profiles. Three of the volunteers, 5, 6, and, 7 had elevated HSP proteins that correlated with inflammation. Interestingly, on the same day the volunteers’ samples were collected, volunteer 5 had also had blood collected for the Cedars-Sinai EMBARC study, where blood is qualitatively analyzed for IgG against nucleocapsid (N) protein (2.52, where >1.4 is positive) and quantitively analyzed for IgG for the spike (S) protein (34117.98 AU/mL, where ≥50 AU/mL is positive) of Covid-19. The antibody (IgG) results indicate that volunteer 5 may have had a recent infection with Covid-19. The HSP proteins associated with an elevation of inflammation, CRP, VWF, HP, and IGHA1 immunoglobin were also elevated in volunteer 5. Volunteers 6 and 7 also had elevated levels of CRP compared to the other volunteers (Figures 3, 4, and 6), which also correlated with a history of hypertension and inflammation, respectively. According to the Cleveland Clinic, normal CRP levels are ≤9mg/L and moderate elevation is between 10.0–100.0 mg/L, where >2.0 mg/L is at higher risk of heart disease, while a hs-CRP test <2.0 mg/L is designated lower risk of heart disease. So, although all volunteers’ CRP levels were below ≤9mg/L, it was interesting to see stratification of the 8-volunteer “healthy/normal” cohort with three individuals correlating with signs of mild inflammation, detected with the HSP assay.
The SIL calibration curves allowed us to calculate the physiological concentrations of the endogenous HSP proteins in the volunteers in each biofluid (mg/L and nmol/L; Table S7). In the future, intact isotopically labeled proteins could be added to the HSP assay to assess the absolute physiological concentrations of proteins, in order to meet the FDA-regulated Tier 1 assay46 of physician-prescribed blood tests. The dynamic range of HSP proteins in capillary blood (Figure 6C) is all within the lower limits of detection (LLOD).
The unbiased protein discovery DIA-MS analysis identified 3811 unambiguous proteins at 1% FDR (4661 proteins at 5% FDR) in the three biofluids when using the pan-human spectral library45 compared to the 1121 proteins identified using a twin plasma spectral library,44 and 2.21-fold more protein in capillary whole blood and 2.57-fold more protein in venous whole blood, while only a 0.06-fold difference in plasma, demonstrating the significant gain of protein identifications when analyzing whole blood over plasma alone and using a comprehensive spectral library. Interestingly, most HSP proteins were identified in the whole blood samples, demonstrating that the HSP assay could be incorporated into a DIA-hybrid discovery and quantitative assay. More importantly, a larger protein/biomarker multiplex panel could be created with the addition of SILs that represent the 122 FDA-approved biomarkers, LDT proteins, and many of the DIA-MS-discovered proteins that have known biological relevance in health and disease, including oxidative stress proteins (e.g., superoxide dismutase, catalase, etc.), longevity (e.g., telomere-related proteins), energy metabolism (e.g., creatine kinases, adenylate kinases, etc.), heart and muscle proteins (e.g., Troponin C, myosins, tropomyosins, actin, desmin, FHL2, etc.), and oncogene-related proteins (e.g., vav, myc, src, etc.), among many other proteins24 for comprehensive health surveillance panel. In addition, the biomarker panel could be expanded to include a broad range of low abundant organ-specific protein signaling moieties circulating in the bloodstream such as myokines (skeletal muscle),3 osteokines (bone),47 cardiokines (heart),48 adipokines (white adipose tissue),49 and neurokines (neurons)50 that may be monitored in health and disease, where intervention through exercise43 can stimulate their release, promoting health in other tissues.51 In addition, the protein activity/function in health and disease can be more specifically monitored through post-translational modifications (PTMs) such as phosphorylation, glycosylation, and acetylation as well as monitoring single nucleotide polymorphisms (SNPs) by creating the representative SILs.52
Conclusions
A growing population of people are taking control of managing their personal health, and there has been a rise in companies that are promoting systems to improve consumer access to medical resources and quantifiable diagnostic tests.20,53 A comprehensive health surveillance panel of proteins would add significant insight into the effectiveness of the treatment, better assess prognosis, and monitor unforeseen side effects. In addition, a comprehensive biomarker panel would help individuals correlate quantitative metrics to the effectiveness of their personalized fitness and nutrition intervention regimes to optimize their health and slow the onset of age-related metabolic diseases.51,54 The combination of having an easy to use and reliable remote sampling device to fully utilize the power of mass spectrometry technology advancements to reach an expansive biomarker panel will be critical in the stratification of individual biosignatures for personalized medicine and health.17,24,55
Acknowledgments
The authors would like to acknowledge key contributions for the success of this study. The authors would like to thank Cedars-Sinai for the support of Precision Biomarkers Laboratories (PBL). The authors would like to thank the phlebotomists at the Human Physiology Laboratory at Cedars-Sinai for collecting venous blood samples from our 8 volunteers. Shruti Rao coordinated compliance with the IRB protocol, consent forms, and blood draws as well as provided critical review and comments on the manuscript. The original patent was “Highly multiplexed and mass spectrometry based methods to measuring 72 human proteins” patent application No. 62/448.319 (international publication number WO, 2018/136825 A1). The authors would like to thank ACBRI bioinformatics group for ProEpic data analysis platform and PBL’s Jonathan Bui’s contributions. The authors would also like to thank Angel Keoseyan for her work with the DIA platform. The authors would like to thank Ana Gabriela Loayza and Dragana Noe for art contributions. The authors acknowledge BioRender for use of images in Figure 1 and TOC abstract (e.g., HPLC, mass spectrometers, hand blood drop, pipette, and test tube with blood), in which Cedars-Sinai has a license. In loving memory of Boone Miller.
Supporting Information Available
The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.analchem.3c01189.
Additional experimental details; materials; methods; results; map of Plate 1 and Plate 2; overall average %CV of HSP proteins in quality control (QC) samples from sample plate1 and 2; graphical representation of Pearson statistical analysis results plotting correlationcoefficient vs p-values, and tables (PDF)
HSP precursor ions and 466 transition list for LC/MRM-MS targetted analysis (Table S1) (XLSX)
PAR and %CV of HSP quantifier peptides from capillary whole blood, venous whole blood and plasma. (WB = whole blood, C = capillary, V = venous, Pl = plasma) (Table S2) (XLSX)
Quality control (QC) targeted LC/MRM-MS of HSP peptides on plate 1 and plate 2 (Table S3) (XLSX)
Pearson statistical analysis comparing similarity of HSP proteins PAR signal between capillary blood and venous blood, capillary blood and plasma and venous (Table S4) (XLSX)
Average slopes and %CV calculated from calibration curves from plate 1 and plate 2 (Table S6) (XLSX)
PAR averages of 8 volunteers and the biological replicates for capillary blood (n = 48), venous blood (n = 48) and matched plasma (n= 24) was used to calculate (Table S7) (XLSX)
DIA 5%FDR using twin plasma spectral library (Table S9) (XLSX)
DIA 1%FDR using pan-human spectral library. Total proteins 3811 (Table S10) (XLSX)
DIA 5%FDR using pan-human spectral library. (Total Protein 4661) (Table S11) (XLSX)
Author Contributions
S.A.W. and S.M.M. conceptualized, managed, and planned the execution of the study and are corresponding authors. S.A.W. performed the sample collection, sample processing, LC/MRM-MS experimentation, data analysis, and statistical analysis and constructed and finalized all figures and wrote, edited, and submitted the final manuscript. SMM provided critical review and comments to the manuscript. N.H. and A.M. developed the DIA platform, analyzed samples, compiled data, and provided critical review and comments to the manuscript. Z.L.D. performed the Pearson correlation coefficient statistical analysis and provided critical review and comments on the manuscript. Q.F. is a coinventor on the HSP patent used for this study and reviewed the manuscript. J.V.E. is a coinventor on the HSP patent, provided key elements for the HSP assay, and provided critical review and comments on the manuscript.
The authors declare no competing financial interest.
Supplementary Material
References
- van Eyk J. E.; Snyder M. P. Precision Medicine: Role of Proteomics in Changing Clinical Management and Care. J. Proteome Res. 2019, 18, 1–6. 10.1021/acs.jproteome.8b00504. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gorzelitz J.; Trabert B.; Katki H. A.; Moore S. C.; Watts E. L.; Matthews C. E. Independent and Joint Associations of Weightlifting and Aerobic Activity with All-Cause, Cardiovascular Disease and Cancer Mortality in the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial. Br. J. Sports Med. 2022, 56, 1277–1283. 10.1136/bjsports-2021-105315. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Little H. C.; Tan S. Y.; Cali F. M.; Rodriguez S.; Lei X.; Wolfe A.; Hug C.; Wong G. W. Multiplex Quantification Identifies Novel Exercise-Regulated Myokines/Cytokines in Plasma and in Glycolytic and Oxidative Skeletal Muscle. Mol. Cell. Proteomics 2018, 17, 1546–1563. 10.1074/mcp.RA118.000794. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tucker W. J.; Fegers-Wustrow I.; Halle M.; Haykowsky M. J.; Chung E. H.; Kovacic J. C. Exercise for Primary and Secondary Prevention of Cardiovascular Disease: JACC Focus Seminar 1/4. J. Am. Coll. Cardiol. 2022, 80, 1091–1106. 10.1016/j.jacc.2022.07.004. [DOI] [PubMed] [Google Scholar]
- Sempionatto J. R.; Lasalde-Ramírez J. A.; Mahato K.; Wang J.; Gao W. Wearable Chemical Sensors for Biomarker Discovery in the Omics Era. Nat. Rev. Chem. 2022, 6, 899–915. 10.1038/s41570-022-00439-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wang L.; Wang J.; Yu P.; Feng J.; Xu G.; Zhao X.; Wang T.; Lehmann H. I.; Li G.; Sluijter J. P. G.; Xiao J. METTL14 Is Required for Exercise-Induced Cardiac Hypertrophy and Protects against Myocardial Ischemia-Reperfusion Injury. Nat. Commun. 2022, 13, 6762 10.1038/s41467-022-34434-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Galié S.; García-Gavilán J.; Papandreou C.; Camacho-Barcía L.; Arcelin P.; Palau-Galindo A.; Rabassa A.; Bulló M. Effects of Mediterranean Diet on Plasma Metabolites and Their Relationship with Insulin Resistance and Gut Microbiota Composition in a Crossover Randomized Clinical Trial. Clin. Nutr. 2021, 40, 3798–3806. 10.1016/J.CLNU.2021.04.028. [DOI] [PubMed] [Google Scholar]
- Hasin Y.; Seldin M.; Lusis A.. Multi-Omics Approaches to Disease. In Genome Biology; BioMed Central Ltd.2017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cai Z.; Poulos R. C.; Liu J.; Zhong Q.. IScience Machine Learning for Multi-Omics Data Integration in Cancer. https://doi.org/10.1016/j.isci. [DOI] [PMC free article] [PubMed]
- Miao Z.; Humphreys B. D.; McMahon A. P.; Kim J. Multi-Omics Integration in the Age of Million Single-Cell Data. Nat. Rev. Nephrol. 2021, 17, 710–724. 10.1038/s41581-021-00463-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ignjatovic V.; Geyer P. E.; Palaniappan K. K.; Chaaban J. E.; Omenn G. S.; Baker M. S.; Deutsch E. W.; Schwenk J. M. Mass Spectrometry-Based Plasma Proteomics: Considerations from Sample Collection to Achieving Translational Data. J. Proteome Res. 2019, 18, 4085–4097. 10.1021/acs.jproteome.9b00503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Longo V. D.; Anderson R. M. Nutrition, Longevity and Disease: From Molecular Mechanisms to Interventions. Cell 2022, 185, 1455–1470. 10.1016/j.cell.2022.04.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wang D. D.; Hu F. B. Precision Nutrition for Prevention and Management of Type 2 Diabetes. Lancet Diabetes Endocrinol. 2018, 6, 416–426. 10.1016/s2213-8587(18)30037-8. [DOI] [PubMed] [Google Scholar]
- Noar A.; Welsh J.; Price D. A.. 870-P: A1C Reductions and Improved Patient-Reported Outcomes Following CGM Initiation in Insulin-Managed T2D Diabetes 2020; Vol. 69 10.2337/db20-870-P. [DOI] [Google Scholar]
- Kao P. C.; Shiesh S.-C.; Wu T.-J.. Review: Serum C-Reactive Protein as a Marker for Wellness Assessment; 2006. www.annclinlabsci.org. [PubMed]
- Tognetti M.; Sklodowski K.; Müller S.; Kamber D.; Muntel J.; Bruderer R.; Reiter L. Biomarker Candidates for Tumors Identified from Deep-Profiled Plasma Stem Predominantly from the Low Abundant Area. J. Proteome Res. 2022, 21, 1718–1735. 10.1021/acs.jproteome.2c00122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zhou Y.; Tan Z.; Xue P.; Wang Y.; Li X.; Guan F. High-Throughput, in-Depth and Estimated Absolute Quantification of Plasma Proteome Using Data-Independent Acquisition/Mass Spectrometry (“HIAP-DIA”). Proteomics 2021, 21, 2000264 10.1002/pmic.202000264. [DOI] [PubMed] [Google Scholar]
- Javaheri A.; Diab A.; Zhao L.; Qian C.; Cohen J. B.; Zamani P.; Kumar A.; Wang Z.; Ebert C.; Maranville J.; Kvikstad E.; Basso M.; van Empel V.; Richards A. M.; Doughty R. N.; Rietzschel E.; Kammerhoff K.; Gogain J.; Schafer P.; Seiffert D. A.; Gordon D. A.; Ramirez-Valle F.; Mann D. L.; Cappola T. P.; Chirinos J. A. Proteomic Analysis of Effects of Spironolactone in Heart Failure With Preserved Ejection Fraction. Circ.: Heart Failure 2022, 15, e009693 10.1161/CIRCHEARTFAILURE.121.009693. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ferdosi S.; Tangeysh B.; Brown T. R.; Everley P. A.; Figa M.; Mclean M.; Elgierari M.; Zhao X.; Garcia J.; Wang T.; Chang M. E. K.; Riedesel K.; Chu J.; Mahoney M.; Xia H.; O’brien E. S.; Stolarczyk C.; Harris D.; Platt T. L.; Ma P.; Goldberg M.; Langer R.; Flory M. R.; Benz R.; Tao W.; Cuevas J. C.; Batzoglou S.; Blume J. E.; Siddiqui A.; Hornburg D.; Farokhzad O. C.; Mclean M.; Mahoney M.; Harris D.; Hornburg D. Engineered Nanoparticles Enable Deep Proteomics Studies at Scale by Leveraging Tunable Nano-Bio Interactions. Proc. Natl. Acad. Sci. U.S.A. 2022, 119, e2106053119 10.1073/pnas.2106053119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bhowmick P.; Roome S.; Borchers C. H.; Goodlett D. R.; Mohammed Y. An Update on MRMAssayDB: A Comprehensive Resource for Targeted Proteomics Assays in the Community. J. Proteome Res. 2021, 20, 2105–2115. 10.1021/acs.jproteome.0c00961. [DOI] [PMC free article] [PubMed] [Google Scholar]
- van den Broek I.; Mastali M.; Mouapi K.; Bystrom C.; Merz C. N. B.; van Eyk J. E. Quality Control and Outlier Detection of Targeted Mass Spectrometry Data from Multiplex Protein Panels. J. Proteome Res. 2020, 19, 2278–2293. 10.1021/acs.jproteome.9b00854. [DOI] [PubMed] [Google Scholar]
- Freue G. V. C.; Borchers C. H. Multiple Reaction Monitoring (MRM): Principles and Application to Coronary Artery Disease. Circ.: Cardiovasc. Genet. 2012, 5, 378 10.1161/circgenetics.111.959528. [DOI] [PubMed] [Google Scholar]
- Venable J. D.; Dong M. Q.; Wohlschlegel J.; Dillin A.; Yates J. R. Automated Approach for Quantitative Analysis of Complex Peptide Mixtures from Tandem Mass Spectra. Nat. Methods 2004, 1, 39–45. 10.1038/nmeth705. [DOI] [PubMed] [Google Scholar]
- Kennedy J. J.; Whiteaker J. R.; Ivey R. G.; Burian A.; Chowdhury S.; Tsai C. F.; Liu T.; Lin C.; Murillo O. D.; Lundeen R. A.; Jones L. A.; Gafken P. R.; Longton G.; Rodland K. D.; Skates S. J.; Landua J.; Wang P.; Lewis M. T.; Paulovich A. G. Internal Standard Triggered-Parallel Reaction Monitoring Mass Spectrometry Enables Multiplexed Quantification of Candidate Biomarkers in Plasma. Anal. Chem. 2022, 94, 9540–9547. 10.1021/acs.analchem.1c04382. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Protti M.; Mandrioli R.; Mercolini L. Tutorial: Volumetric Absorptive Microsampling (VAMS). Anal. Chim. Acta 2019, 1046, 32–47. 10.1016/j.aca.2018.09.004. [DOI] [PubMed] [Google Scholar]
- van den Broek I.; Fu Q.; Kushon S.; Kowalski M. P.; Millis K.; Percy A.; Holewinski R. J.; Venkatraman V.; van Eyk J. E. Application of Volumetric Absorptive Microsampling for Robust, High-Throughput Mass Spectrometric Quantification of Circulating Protein Biomarkers. Clin. Mass Spectrom. 2017, 4–5, 25–33. 10.1016/j.clinms.2017.08.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rudge J.; Kushon S. Volumetric Absorptive Microsampling: Its Use in COVID-19 Research and Testing. Bioanalysis 2021, 13, 1851–1863. 10.4155/bio-2021-0102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fu Q.; Kowalski M. P.; Mastali M.; Parker S. J.; Sobhani K.; van den Broek I.; Hunter C. L.; van Eyk J. E. Highly Reproducible Automated Proteomics Sample Preparation Workflow for Quantitative Mass Spectrometry. J Proteome Res. 2018, 17, 420–428. 10.1021/acs.jproteome.7b00623. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fu Q.; Johnson C. W.; Wijayawardena B. K.; Kowalski M. P.; Kheradmand M.; van Eyk J. E. A Plasma Sample Preparation for Mass Spectrometry Using an Automated Workstation. J. Visualized Exp. 2020, 2020, e59842 10.3791/59842. [DOI] [PubMed] [Google Scholar]
- Mc Ardle A.; Binek A.; Moradian A.; Chazarin Orgel B.; Rivas A.; Washington K. E.; Phebus C.; Manalo D. M.; Go J.; Venkatraman V.; Coutelin Johnson C. W.; Fu Q.; Cheng S.; Raedschelders K.; Fert-Bober J.; Pennington S. R.; Murray C. I.; van Eyk J. E. Standardized Workflow for Precise Mid- and High-Throughput Proteomics of Blood Biofluids. Clin. Chem. 2022, 68, 450–460. 10.1093/clinchem/hvab202. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fuller G.; Mouapi K. N.; Joung S.; Shufelt C.; van den Broek I.; Lopez M.; dhawan S.; Mastali M.; Spiegel C.; Merz N. B.; van Eyk J. E.; Robinson A. Feasibility of Patient-Centric Remote Dried Blood Sampling: The Prediction, Risk, and Evaluation of Major Adverse Cardiac Events (PRE-MACE) Study. Biodemography Soc. Biol. 2020, 65, 313–322. 10.1080/19485565.2020.1765735. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Klumpp-Thomas C.; Kalish H.; Drew M.; Hunsberger S.; Snead K.; Fay M. P.; Mehalko J.; Shunmugavel A.; Wall V.; Frank P.; Denson J. P.; Hong M.; Gulten G.; Messing S.; Hicks J.; Michael S.; Gillette W.; Hall M. D.; Memoli M. J.; Esposito D.; Sadtler K. Standardization of ELISA Protocols for Serosurveys of the SARS-CoV-2 Pandemic Using Clinical and at-Home Blood Sampling. Nat. Commun. 2021, 12, 113 10.1038/s41467-020-20383-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Moon S. J.; Lee S. E.; Kwak Y. G.; Kim M. G. Validation of Lc-Ms/Ms Method for Determination of Rosuvastatin Concentration in Human Blood Collected by Volumetric Absorptive Microsampling (Vams). Transl. Clin. Pharmacol. 2021, 29, 125–134. 10.12793/tcp.2021.29.e13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Li H.; Bigwarfe T.; Myzithras M.; Waltz E.; Ahlberg J. Application of Mitra Microsampling for Pharmacokinetic Bioanalysis of Monoclonal Antibodies in Rats. Bioanalysis 2019, 11, 13–20. 10.4155/bio-2018-0228. [DOI] [PubMed] [Google Scholar]
- Shufelt C.; Dzubur E.; Joung S.; Fuller G.; Mouapi K. N.; van den Broek I.; Lopez M.; Dhawan S.; Arnold C. W.; Speier W.; Mastali M.; Fu Q.; van Eyk J. E.; Spiegel B.; Merz C. N. B. A Protocol Integrating Remote Patient Monitoring Patient Reported Outcomes and Cardiovascular Biomarkers. NPJ Digital Med. 2019, 2, 84 10.1038/s41746-019-0145-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Moorthy G. S.; Vedar C.; Zane N. R.; Downes K. J.; Prodell J. L.; DiLiberto M. A.; Zuppa A. F. Development and Validation of a Volumetric Absorptive Microsampling- Liquid Chromatography Mass Spectrometry Method for the Analysis of Cefepime in Human Whole Blood: Application to Pediatric Pharmacokinetic Study. J. Pharm. Biomed. Anal. 2020, 179, 113002 10.1016/j.jpba.2019.113002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lei B. U. W.; Prow T. W. A Review of Microsampling Techniques and Their Social Impact. Biomed. Microdevices 2019, 21, 81 10.1007/s10544-019-0412-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Mathew B. S.; Mathew S. K.; Aruldhas B. W.; Prabha R.; Gangadharan N.; David V. G.; Varughese S.; John G. T. Analytical and Clinical Validation of Dried Blood Spot and Volumetric Absorptive Microsampling for Measurement of Tacrolimus and Creatinine after Renal Transplantation. Clin. Biochem. 2022, 105–106, 25–34. 10.1016/j.clinbiochem.2022.04.014. [DOI] [PubMed] [Google Scholar]
- Rockwood A. L.; Lowenthal M. S.; Bystrom C. Harmonization of Liquid Chromatography–Tandem Mass Spectrometry Protein Assays. Clin. Lab. Med. 2018, 38, 499–513. 10.1016/j.cll.2018.05.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kutuzova S.; Colaianni P.; Röst H.; Sachsenberg T.; Alka O.; Kohlbacher O.; Burla B.; Torta F.; Schrübbers L.; Kristensen M.; Nielsen L.; Herrgård M. J.; McCloskey D. SmartPeak Automates Targeted and Quantitative Metabolomics Data Processing. Anal. Chem. 2020, 92, 15968–15974. 10.1021/acs.analchem.0c03421. [DOI] [PubMed] [Google Scholar]
- Nanjappa V.; Thomas J. K.; Marimuthu A.; Muthusamy B.; Radhakrishnan A.; Sharma R.; Ahmad Khan A.; Balakrishnan L.; Sahasrabuddhe N. A.; Kumar S.; Jhaveri B. N.; Sheth K. V.; Kumar Khatana R.; Shaw P. G.; Srikanth S. M.; Mathur P. P.; Shankar S.; Nagaraja D.; Christopher R.; Mathivanan S.; Raju R.; Sirdeshmukh R.; Chatterjee A.; Simpson R. J.; Harsha H. C.; Pandey A.; Prasad T. S. K. Plasma Proteome Database as a Resource for Proteomics Research: 2014 Update. Nucleic Acids Res. 2014, 42, D959–D965. 10.1093/nar/gkt1251. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Geyer P. E.; Holdt L. M.; Teupser D.; Mann M. Revisiting Biomarker Discovery by Plasma Proteomics. Mol. Syst. Biol. 2017, 13, 942 10.15252/msb.20156297. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gaither C.; Popp R.; Mohammed Y.; Borchers C. H. Determination of the Concentration Range for 267 Proteins from 21 Lots of Commercial Human Plasma Using Highly Multiplexed Multiple Reaction Monitoring Mass Spectrometry. Analyst 2020, 145, 3634–3644. 10.1039/c9an01893j. [DOI] [PubMed] [Google Scholar]
- Liu Y.; Buil A.; Collins B. C.; Gillet L. C.; Blum L. C.; Cheng L.; Vitek O.; Mouritsen J.; Lachance G.; Spector T. D.; Dermitzakis E. T.; Aebersold R. Quantitative Variability of 342 Plasma Proteins in a Human Twin Population. Mol. Syst. Biol. 2015, 11, 786 10.15252/msb.20145728. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rosenberger G.; Koh C. C.; Guo T.; Röst H. L.; Kouvonen P.; Collins B. C.; Heusel M.; Liu Y.; Caron E.; Vichalkovski A.; Faini M.; Schubert O. T.; Faridi P.; Ebhardt H. A.; Matondo M.; Lam H.; Bader S. L.; Campbell D. S.; Deutsch E. W.; Moritz R. L.; Tate S.; Aebersold R. A Repository of Assays to Quantify 10,000 Human Proteins by SWATH-MS. Sci. Data 2014, 1, 140031 10.1038/sdata.2014.31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Boja E. S.; Fehniger T. E.; Baker M. S.; Marko-Varga G.; Rodriguez H. Analytical Validation Considerations of Multiplex Mass-Spectrometry-Based Proteomic Platforms for Measuring Protein Biomarkers. J. Proteome Res. 2014, 13, 5325–5332. 10.1021/pr500753r. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lu W.; Xiao W.; Xie W.; Fu X.; Pan L.; Jin H.; Yu Y.; Zhang Y.; Li Y. The Role of Osteokines in Sarcopenia: Therapeutic Directions and Application Prospects. Front. Cell Dev. Biol. 2021, 9, 735374 10.3389/fcell.2021.735374. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Linscheid N.; Santos A.; Poulsen P. C.; Mills R. W.; Calloe K.; Leurs U.; Ye J. Z.; Stolte C.; Thomsen M. B.; Bentzen B. H.; Lundegaard P. R.; Olesen M. S.; Jensen L. J.; Olsen J. V.; Lundby A. Quantitative Proteome Comparison of Human Hearts with Those of Model Organisms. PLoS Biol. 2021, 19, e3001144 10.1371/journal.pbio.3001144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zhang Y.; Yu M.; Dong J.; Wu Y.; Tian W. Identification of Novel Adipokines through Proteomic Profiling of Small Extracellular Vesicles Derived from Adipose Tissue. J. Proteome Res. 2020, 19, 3130–3142. 10.1021/acs.jproteome.0c00131. [DOI] [PubMed] [Google Scholar]
- Xu Y.; Song X.; Wang D.; Wang Y.; Li P.; Li J. Proteomic Insights into Synaptic Signaling in the Brain: The Past, Present and Future. Mol. Brain 2021, 14, 37 10.1186/s13041-021-00750-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gonzalez-Gil A. M.; Elizondo-Montemayor L. The Role of Exercise in the Interplay between Myokines, Hepatokines, Osteokines, Adipokines, and Modulation of Inflammation for Energy Substrate Redistribution and Fat Mass Loss: A Review. Nutrients 2020, 12, 1899 10.3390/nu12061899. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fert-Bober J.; Murray C. I.; Parker S. J.; van Eyk J. E. Precision Profiling of the Cardiovascular Post-Translationally Modified Proteome Where There Is a Will, There Is a Way. Circ. Res. 2018, 122, 1221–1237. 10.1161/CIRCRESAHA.118.310966. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kosack C. S.; Page A. L.; Klatser P. R. A Guide to Aid the Selection of Diagnostic Tests. Bull. W. H. O. 2017, 95, 639–645. 10.2471/BLT.16.187468. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kirk B.; Feehan J.; Lombardi G.; Duque G. Muscle, Bone, and Fat Crosstalk: The Biological Role of Myokines, Osteokines, and Adipokines. Curr. Osteoporosis Rep. 2020, 18, 388–400. 10.1007/s11914-020-00599-y. [DOI] [PubMed] [Google Scholar]
- Lou R.; Tang P.; Ding K.; Li S.; Tian C.; Li Y.; Zhao S.; Zhang Y.; Shui W. Hybrid Spectral Library Combining DIA-MS Data and a Targeted Virtual Library Substantially Deepens the Proteome Coverage. iScience 2020, 23, 100903 10.1016/j.isci.2020.100903. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.