key: cord-034191-qqb2knmo
authors: Alayi, Tchilabalo D.; Tawalbeh, Shefa M.; Ogundele, Michael; Smith, Holly R.; Samsel, Alison M.; Barbieri, Marissa L.; Hathout, Yetrib
title: Tandem Mass Tag-Based Serum Proteome Profiling for Biomarker Discovery in Young Duchenne Muscular Dystrophy Boys
date: 2020-10-06
journal: ACS Omega
DOI: 10.1021/acsomega.0c03206
sha: 
doc_id: 34191
cord_uid: qqb2knmo

[Image: see text] Blood-accessible molecular biomarkers are becoming highly attractive tools to assess disease progression and response to therapies in Duchenne muscular dystrophy (DMD) especially in very young patients for whom other outcome measures remain subjective and challenging. In this study, we have standardized a highly specific and reproducible multiplexing mass spectrometry method using the tandem mass tag (TMT) strategy in combination with depletion of abundant proteins from serum and high-pH reversed-phase peptide fractionation. Differential proteome profiling of 4 year-old DMD boys (n = 9) and age-matched healthy controls (n = 9) identified 38 elevated and 50 decreased serum proteins (adjusted P < 0.05, FDR <0.05) in the DMD group relative to the healthy control group. As expected, we confirmed previously reported biomarkers but also identified novel biomarkers. These included novel muscle injury-associated biomarkers such as telethonin, smoothelin-like protein 1, cofilin-1, and plectin, additional muscle-specific enzymes such as UTP–glucose-1-phosphate uridylyltransferase, aspartate aminotransferase, pyruvate kinase PKM, lactotransferrin, tissue alpha-l-fucosidase, pantetheinase, and ficolin-1, and some pro-inflammatory and cell adhesion-associated biomarkers such as leukosialin, macrophage receptor MARCO, vitronectin, galectin-3-binding protein, and ProSAAS. The workflow including serum depletion, sample processing, and mass spectrometry analysis was found to be reproducible and stable over time with CV < 20%. Furthermore, the method was found to be superior in terms of specificity compared to other multiplexing affinity-based methods. These findings demonstrate the specificity and reliability of TMT-based mass spectrometry methods in detection and identification of serum biomarkers in presymptomatic young DMD patients.

Duchenne muscular dystrophy (DMD) remains a serious and fatal muscle disease with a worldwide incidence of 1:5000 male births. 1 DMD is due to mutations in the X-linked dystrophin gene, leading to the loss of expression of dystrophin, an essential skeletal muscle protein that maintains the integrity and function of muscle fibers. 2 There is no cure for DMD to date except the use of corticosteroids, which are known to delay the loss of ambulation by 2 to 3 years without changing the disease course. 2, 3 Despite a number of completed and ongoing clinical trials, only two drugs received conditional approval from European Medicines Agency (EMA) and US Food and Drug Administration (FDA) in the past 15 years. These include ataluren, a stop codon read-through, and eteplirsen, an exon skipping antisense oligonucleotide that restores the reading frame during dystrophin mRNA translation. 4, 5 The slow development of new therapies in DMD has been hindered by the fact that DMD is a rare disease and that most clinical trials enroll ambulatory patients within a specific age range (4−10 years old), resulting in statistically unpowered studies to accurately assess the outcomes. Furthermore, current clinical tests to assess disease progression and response to therapies such as 6 min walk test, 10 m run/walk velocity, and other physical tests are subjective and might require longer clinical trials to determine the efficacy or failure of an investigational drug. 6−8 Major advances in biomarker discovery have been achieved in the past 7 years in DMD. 9−12 Owing to the multiplexing capabilities of affinity-based methods such as antibody multiplexing bead technology and SomaScan aptamer technology, a large number of biomarkers were discovered in DMD patients and confirmed across different laboratories and different cohorts. 13−16 Although these multiplexing affinitybased assays are superior to mass spectrometry in terms of sensitivity, dynamic range, and throughput, they are inherently not quantitative and lack accuracy in determining true fold change in biomarker levels. 17 Furthermore, these multiplexing affinity methods only detect and measure targeted sets of biomarkers and are not suitable for de novo discovery of novel biomarkers or discrimination between biomarkers that exist as multiple isoforms. Thus, alternative methods that are more versatile, specific, reproducible, and accurate and that can easily be implemented by other laboratories are needed.

In this study, we sought to optimize and standardize a serum processing workflow in combination with tandem mass tag (TMT) multiplexing strategy to systematically survey the serum proteome of young untreated DMD boys and agematched healthy controls and identify biomarkers associated in the early stages of the disease. In this study, we focused on very young DMD boys for two reasons. First is to avoid confounding variables due to glucocorticoid use 14 and second is to define blood-circulating biomarkers that might be associated with early stages of the disease. These biomarkers might eventually be used as tools to assess disease progression and response to therapies in this younger DMD population and facilitate their enrollment into clinical trials from which they are often excluded because of lack of outcome measures for younger patients. Our method successfully confirmed some of the previously reported biomarkers but also identified novel biomarkers in young untreated DMD patients.

Sample Preparation and Quality Control Check. To optimize our MS-based tandem mass tag (TMT) multiplex quantitation, we evaluated each step in the sample preparation and MS analysis through the workflow depicted in (Figure 1 ). This workflow was evaluated in triplicate using commercial serum. We first checked for the depletion kit efficiency using three different aliquots of serum samples in triplicates, 4, 8, and 12 μL containing 300, 600, and 900 μg of total proteins, respectively.

The results show a linear response between total protein used for depletion and total protein recovered after the depletion. Averages of 17.9 ± 0.2 μg (1.1% CV), 49.9 ± 1.4 μg (2.9% CV), and 81.4 ± 5.9 μg (7.3% CV) of protein were collected from the 300, 600, and 900 μg aliquots, respectively ( Figure 2A ). Together, these findings suggest that the depletion columns have enough capacity to efficiently deplete serum-abundant proteins from serum aliquots in the 300 to 900 μg total protein range in a reproducible manner with an overall CV < 10%. Moving forward, we decided to use serum aliquots containing 900 μg of total proteins as a starting material.

Depletion efficiency and reproducibility were further checked using 1D SDS-PAGE. Nondepleted serum proteins exhibited very few intense protein bands, among which human serum albumin (HSA) was the most intense band on the 1D SDS-PAGE ( Figure 2B ). The technical triplicate analysis of 3 μg of depleted serum revealed a large number of new protein bands (e.g., bands a, b, c, d, and e) on 1D SDS-PAGE and a considerable reduction in major protein bands such as HSA. Optical density analysis of randomly selected new protein bands detected in the depleted serum sample by 1D SDS-PAGE (bands a, b, c, d, and e) showed reproducibility in depletion with CVs < 10% ( Figure 2C ). Furthermore, we checked the efficiency and reproducibility of the depletion kit using western blot analysis of two proteins ( Figure 2D ), one of which is among the depleted proteins (haptoglobin, HP) and another one of which is not among the depleted proteins (gelsolin, GSN). The depletion process resulted in an enrichment of GSN protein by approximately 10-fold (CVs < 16%) and in a depletion of HP protein by approximately 2fold (CVs <14%), in comparison to nondepleted serum as expected ( Figure 2E ).

We then checked the reproducibility and stability of the entire workflow using 900 μg of total serum proteins as a starting material. Aliquots of 10−14 μL of each sample were processed for top 12 abundant protein depletion kit and 50 μg of depleted sample was processed for TMT 6-plex tagging, as described in the method. The obtained peptides were labeled with TMT reagent, which covalently modifies the primary amines of the N termini and lysine side chains of peptides. 19 Equal volumes of tagged samples were taken and mixed together and 18 fractions were collected from the high-pH reversed-phase spin column and analyzed by LC−MS/MS, as described above (Figure 1 ). Figure 1 . Chart depicting the overall workflow from sample preparation to mass spectrometry analysis to data processing. Serum samples (900 μg per aliquot) from 4 year-old DMD patients (n = 9) and age-matched healthy controls (n = 9) were processed for 12 most abundant proteins depletion. Resulting samples were in-solution-digested and randomized for TMT tagging, as shown in the figure. One of the control samples (dark blue) was used as a reference to normalize the data. Each sample mixture was further fractionated using a high-pH reversed-phase column and each fraction was analyzed by LC−MS/MS. Data were processed as described in the method.

The stability and reproducibility of the overall workflow ( Figure 1 ) including serum depletion step, tandem mass tag (TMT) multiplexing step, and high-pH reversed-phase peptide fractionation were evaluated by LC−MS/MS in triplicate analysis using commercial serum. After removing contaminant proteins and missing data, this standardized TMT MS method enabled the identification and quantification of 618 serum proteins across the triplicate experiment with CV < 20% for 601 proteins (97% of protein quantified), CV < 15% for 588 proteins (95% of protein quantified), and CV < 5% for 396 proteins ( Figure 2F ). The complete list of 618 quantified proteins by LC−MS/MS and associated normalized protein abundance ratios, means of normalized abundance ratios, standard deviation, and CVs of three technical replicate analysis of commercial serum are shown in Table S2 .

Serum Protein Biomarker Screening in Young DMD Patients. Serum samples from 4 year-old glucocorticoid naive DMD patients (n = 9) and age-matched healthy controls (n = 9) were processed for proteome profiling using our standardized multiplexing TMT-based mass spectrometry method described above. A total of four experiment sets of TMT MSbased assays were run in parallel including two 6-plex Figure 2 . Quality control of the immunoaffinity depletion of the 12 highly abundant proteins from serum samples. Aliquots of serum samples with different amounts of total proteins (300, 600, and 900 μg) were used in triplicate to examine the variability and quality of the immunoaffinity depletion process. (A) Linear response between the different starting amounts of total serum proteins and recovered total proteins in depleted samples. Approximately 17.9 μg (CV 1.1%), 49.9 μg (2.9%), and 81.4 μg (CV 7.3%) of proteins were recovered from serum samples containing 300, 600, and 900 μg of total proteins, respectively. (B) SDS-PAGE showing side-by-side nondepleted serum (3 μg) and depleted serum samples in triplicate (3 μg each). The nondepleted serum proteins exhibited very few intense protein bands, among which human serum albumin (HSA) was the most intense band on the 1D SDS-PAGE, while the depleted serum samples revealed large and reproducible number on new protein bands (ex. bands a, b, c, d, and e). (C) Detailed optical density analysis of randomly selected new protein bands observed on the depleted serum 1D SDS-PAGE (bands a, b, c, d, and e) allowed the comparison of the technical triplicate depletion of serum and estimated the CV < 9.1%. (D) Western blot analysis of depleted proteins (haptoglobin, HP) and nondepleted proteins (gelsolin, GSN). (E) Quantitative analysis of western blot shows an increase in GSN by approximately 10-fold (CVs < 16%) in comparison to nondepleted serum and a decrease in HP by approximately 2-fold (CVs < 14%), confirming the effectiveness and the reproducibility of the depletion kits. (F) Quality control of in-solution digestion and TMT tagging efficiency. A total of 618 serum proteins were identified and quantified. The coefficient of variation between these three technical replicates is depicted in the pie chart with the majority of quantified proteins (97%) showing CV < 20%, with 64% of proteins showing CV < 5% and 95% of proteins showing CV < 15%. experiments comparing three controls and three DMD samples each, one 5-plex experiment comparing two DMD samples and three controls, and one 4-plex experiment comparing one DMD sample and three controls. To control for intra-and inter-experiment variation, the same amount of digest of TMT 126-labeled control serum sample was used as an internal standard in all four experiments ( Figure 1 ). Using our optimized and standardized workflow, 686 proteins were quantified across all 18 samples with only 13% of missing values in all samples (Table S3 ). Further statistical analysis identified 38 significantly elevated proteins and 50 significantly decreased proteins (Student's t-test, adjusted for multiple testing, P < 0.05) in serum samples of 4 year-old DMD boys compared to 4 year-old healthy controls ( Figure 3A ,B and Table 1 ). Twenty-nine proteins out of the 88 potential biomarker candidates had missing values in some samples and these are shown in Figure S1 . fold change in protein levels in serum of young DMD patients compared to healthy controls. Serum proteome profiling was performed on 4 yearold DMD patients (n = 9) and age-matched healthy controls (n = 9) using the TMT method. Significance was defined between the two groups by independent two-sample t-test (two-sided) corrected with adjusted P value and by permutation-based FDR of 0.05. Significantly elevated and significantly decreased proteins in DMD patients relative to controls are presented in green and red, respectively. (B) Hierarchical clustering of serum proteins whose levels were significantly altered between the 4 year-old DMD (n = 9) group and age-matched healthy control (n = 9) group. Intensities of proteins are normalized abundance ratios, which were log2-transformed. As expected, a large number of identified biomarkers (50%) that were found to be elevated in sera of these young DMD boys relative to the healthy controls were of muscle origin based on Gene Ontology molecular function annotations ( Figure S2 ) and information collected using Mass Spectrometry Data Analysis tools. 20 Most of these muscle-associated proteins were previously reported by other research groups and us 9,11,14−16 except telethonin (TCAP), plectin (PLEC), smoothelin-like protein 1 (SMTNL1), myosin-7 (MYH7), PDZ and LIM domain protein 5 (PDLIM5), tropomyosin alpha-1 and beta chain (TPM1, TPM2), troponin C skeletal muscle (TNNC1, TNNC2), troponin I slow skeletal muscle (TNNI1), myomesin-2 (MYOM2), and fructose-bisphosphate aldolase C (ALDOC), which are novel to this study. The fold change difference between DMD patients and healthy controls for this class of biomarkers ranged from 2.45 to 8.66 ( Figure  3A ,B) and likely underlie the well-known sarcolemma instability in DMD and release of muscle-specific proteins. 12, 13, 21 The second class of elevated serum proteins found in young untreated DMD patients relative to controls included the well-known heme carrier protein myoglobin (MG) and muscle-specific enzymes such as creatine kinase Band M-type (CKB, CKM), fructose-bisphosphate aldolase A (ALDOA), and carbonic anhydrase 3 (CA3). Additionally, novel enzymes in this class such as fructose-bisphosphate aldolase C (ALDOC), 72 kDa type IV collagenase (MMP2), UTP−glucose-1-phosphate uridylyltransferase (UGP2), aspartate aminotransferase, cytoplasmic GOT1, pyruvate kinase PKM (PKM), ATP-dependent 6-phosphofructokinase muscle type (PFKM), proteasome subunit beta type-1 (PSMB1), and glutathione S-transferase Mu 2 (GSTM2) (Table 1, Figure 3B , and Figure S1 ) were also found. 9, 11, 22 An example of a panel of these newly identified DMD biomarkers using mass spectrometry is shown in Figure 3C ,D.

Serum proteins that were found to be decreased in young untreated DMD patients relative to healthy controls can be classified into three major groups. One group consisted of proteases such as beta-Ala-His dipeptidase (CNDP1), previously reported by others and us, 14−16 and others newly identified proteases such as lactotransferrin (LTF), tissue alpha-L-fucosidase (FUCA1), pantetheinase (VNN1), plasminogen (PLG), carboxypeptidase N catalytic chain (CPN1), glutamyl aminopeptidase (ENPEP), exostosin-1 (EXT1), ficolin-1 (FCN1), and transketolase (TKT) ( Figure 3A ,B and Figure S2 ). The second group of decreased biomarkers consisted of actin-binding proteins such gelsolin (GSN), thymosin beta-4 (TMSB4X), coactosin-like protein (COTL1), and cofilin-1 (CFL1). Circulating GSN was previously reported by others and our group to be decreased in DMD patients relative to controls, 14−16 while the decrease in the blood levels of the remaining actin-binding proteins is novel to this study. Finally, the last group of serum protein biomarkers that were found significantly decreased in young DMD patients relative to age-matched controls are also novel to this study, which include proteins involved in cell signaling such as leukosialin (SPN), macrophage receptor MARCO (MARCO), vitronectin (VTN), galectin-3-binding protein (LGALS3BP), and ProSAAS (PCSK1N). The list of prominently decreased proteins in young DMD boys relative to healthy controls with a minimum 1.5-fold change is shown in Table 1 , and a box plot example of these decreased biomarkers is shown in Figure 3E ,F. The complete list of proteins identified and quantified in 4 year-old DMD (n = 9) versus age-matched controls (n = 9) and associated statistical analysis are shown in Table S3 .

Biomarker Correlation Map Reveals Biomarker Groups Associated with Different Pathobiochemical Pathways of Muscle Pathogenesis in DMD. Serum or plasma protein levels can provide an indication on irregularities in the muscle integrity and health in DMD. For example, creatine kinase, lactate dehydrogenase, aldolase, troponin, and carbonic anhydrase CAIII are commonly used serum markers associated with muscle injury. 12, 13, 21 To examine if newly identified protein biomarkers correlate with well-known muscle injury proteins or other group of proteins, we performed pairwise correlation analysis of biomarkers, which resulted in a data matrix of 59 proteins for 18 participants (9 DMD patients compared to 9 age-matched healthy control candidates). Correlation of biomarkers to each other together with hierarchical clustering based on normalized intensity ( Figure  3B ) of each protein led to a map of 3481 Pearson correlation coefficients in total for both DMD patients and healthy controls. Interestingly, one of these groups (orthogonal rectangles indicated with black arrows, Figure 4A ) containing muscle-centric proteins such TTN, MYOM3, TNNC1, MYH7, ALDOA, and CKB is positively correlated in DMD patients (red rectangle indicated with black arrows, Figure 4A , Pearson correlation coefficient range between 0.47 and 0.82) and negatively correlated in healthy controls (blue rectangle indicated with black arrows, Figure 4A , Pearson correlation coefficient range between −0.86 and −0.53).

Furthermore, we observed an opposite correlation of the remaining muscle-derived proteins and skeletal muscle enzymes such as TNNC2, ENO3, CKM, and TPM2 with the other group of proteins. Meanwhile, this group of proteins was positively correlated with the other groups of proteins in the normal control (red rectangle indicated with green arrow, Figure 4A , Pearson correlation coefficient range between 0.32 and 0.89). These proteins turned out to be negatively correlated or have complete loss of correlation in DMD with other proteins such as plasma membrane antigen (KCTD3) known for its implication in protein−protein homo-oligomer formation, E3 ubiquitin-protein ligase (RNF34) involved in apoptotic processes, and lactotransferrin (LFT) known for its Each protein is listed with its fold change (DMD/healthy controls) and two-sided P value corrected with adjusted P value and by permutationbased FDR of 0.05. b Isoform. c Adjusted.

http://pubs.acs.org/journal/acsodf Article antimicrobial activity (blue rectangle indicated with green arrow, Figure 4A , Pearson correlation coefficient range between −0.691 and 0.1). In Drosophila, knockdown of RNF3 (dRNF34) has been reported to promote mitochondrial biogenesis, improve exercise capacity in muscle, and extends climbing time to exhaustion in moderately aged flies. 20 The reduction of RNF34 in serum of DMD could be due to a compensatory mechanism in humans where the reduction of RNF34 levels in serum in DMD correlated to the muscle protein increase in the serum due to muscle damage. More importantly, we found a group of mostly significantly decreased extracellular matrix proteins in DMD (e.g., IGFALS, PLG, C4BPA, C4BPB, CPN1, GC, VTN, and C2) to positively cluster with each other in DMD ( Figure 4B , Pearson 

http://pubs.acs.org/journal/acsodf Article correlation coefficient range between 0.32 and 0.93) while exhibiting variable correlation in healthy controls. These proteins are involved in cell adhesion and tissue remodeling (e.g., IGFALS, VTN, and PLG) as well as complement cascade activation and inflammation (C4BPA, C4BPB, C2, and CPN1). This finding suggests a regression or deficit in immunity in DMD, as reported previously. 23 We also found a group of proteins for which levels decrease in DMD and whose positive correlation in normal controls is disturbed in DMD patients ( Figure 4C ). These proteins are mostly located in the extracellular matrix and plasma membrane. Proteins such as AGT, CPN2, BASP1, SPP1 and CNDP1, FCN1, LGALS3BP, and DEFA1 are involved in vasoconstriction, extracellular matrix assembly, complement activation, cell defense response, organ tissue development (e.g., diaphragm), biomineral cell differentiation (e.g., osteoblast), and positive regulation of bone resorption. These proteins are positively correlated to each other in the normal control (Pearson correlation coefficient range between 0.38 and 0.89), while in DMD, their levels were significantly decreased, and the correlation was disturbed ( Figure 4C ). This latest finding suggests that organ growth and bone development in natural defense systems are significantly impaired in DMD. This result corroborates with clinical observations from other groups. 23−25 Confirmation of a Subset of Newly Identified DMD Biomarkers Using ELISA. To verify the reliability of our new multiplex workflow method approach, we performed ELISA assays to confirm a set of biomarker candidates using the same cohort of DMD patients and healthy controls as in MS analysis. The confirmation analysis was performed using ELISA assay for CFL1, FCN1, and PLEC. ELISA analysis was performed in duplicate for both the standard curve and the samples with CV < 15% in all analysis. The ELISA data agree with mass spectrometry data and show that CFL1 and FCN1 were both significantly decreased in DMD ( Figure 4D ,E) by approximately 1.9-fold (P = 0.009) and 3-fold (P = 0.017), respectively. Contrary to mass spectrometry data, PLEC was found to be slightly decreased by 1.2-fold (P = 0.008) using ELISA in DMD patients compared to healthy subjects ( Figure  4F ). Plectin is a large actin-binding protein (highly expressed in muscle and heart) that links intermediate filaments (IF) to dystrophin−glycoprotein complexes (DGC) and integrin complexes and anchors intermediate filaments to desmosomes. 26, 27 The ELISA result of plectin analysis could be explained with the lack of specificity or interference observed with the immunoassay. 17, 28 Indeed, the antibody could recognize a common sequence of a protein present in different proteoforms without being able to distinguish between them. 29 Human plectin protein (approximately 500 kDa) has several isoforms (nine isoforms in universal protein resource (UniProt)) with 96% sequence homology produced by alternative splicing. 26 Antibody specificity for a specific isoform is needed for data interpretation but is not often possible. An example of such interferences has been reported for antibodies against the glycated hemoglobin form HBA1c that also crossreacted with the glycosylated form of hemoglobin. 30 Mass spectrometry was able to quantify two to six unique peptides of isoform-1 of plectin-1 (PLEC) and AQQQAEAER and QVEEAER peptide sequences across all samples used for statistical analysis.

Development of blood-accessible biomarkers in DMD had become attractive in the past few years owing to the growing number of clinical trials and the need of reliable and sensitive outcome measures to assess disease progression and response to therapies. 31 Current outcome measures such as timed functional tests, although important, are often subjective, can be used for ambulatory patients only, and might require longer clinical trials to observe meaningful changes. 4−6,8 Serum biomarkers are rather objective and can be used as monitoring tools in clinical trials if associated with the disease outcomes and drug effect. In the past few years, affinity-based serum proteome profiling methods such as SomaScan aptamer-based technology and antibody bead array have contributed to the discovery of a large number of serum protein biomarkers in DMD. 8, 14, 16 Although they are highly multiplexing and powerful, these techniques are inherently less quantitative and suffers from the epitope effect and specificity for certain target proteins that exist as multiple isoforms. 17, 32 Furthermore, potential biomarker candidates for which there are no aptamers or antibodies available can be overlooked. In this study, we optimized and standardized a TMT multiplex assay workflow featuring high specificity and reproducibility. The workflow consisted of depletion of the 12 most abundant serum proteins, followed by in-solution digestion, TMT tagging, high-pH fractionation, and LC−MS/MS analysis.

The overall workflow tested with three different starting amounts of total serum proteins and in triplicate for each experiment was found to be reproducible and stable over time. More than 600 proteins on average were identified and quantified across all samples with CVs < 16% for the serum protein depletion process and CV < 15% for 95% of protein quantification using the TMT-based mass spectrometry approach.

Our standardized workflow was implemented to compare the serum proteome profiles of untreated 4 year-old DMD patients (n = 9) and age-matched healthy controls (n = 9). We reliably identified 38 elevated and 50 decreased proteins in serum samples of the DMD group relative to the control group (P < 0.05, adjusted for multiple testing). As expected, the majority of elevated proteins were of muscle origin and reflect sarcolemma instability even in young preasymptomatic patients. Although 60% of these muscle injury-associated biomarkers were previously reported by others and our group, 8,9,11,14−16 additional muscle-associated proteins were identified and are novel to this study. These included telethonin (TCAP), plectin (PLEC), smoothelin-like protein 1 (SMTNL1), PDZ and LIM domain protein 5 (PDLIM5), and tropomyosin alpha-1 and beta chain (TPM1, TPM2), which to the best of our knowledge, were not described before. Importantly, the fold change of these muscle injury proteins in DMD patients relative to controls was somewhat lower than the fold changes reported in earlier studies using affinity-based assays such as antibody bead arrays or SomaScan aptamer assay. For instance, circulating CK-M and CA3 were found to be 65-fold and 75-fold higher in young DMD patients relative to age-matched controls using the SomaScan aptamer 14 method, while these same proteins were found to be only 6.3-fold and 4.4-fold higher in DMD patients relative to controls using the TMT mass spectrometry method. The 4.4fold value in CA3 levels in DMD patients relative to controls determined by the TMT method was closer to the 10-fold ACS Omega http://pubs.acs.org/journal/acsodf Article value obtained for this same protein using an absolute quantitative assay. 33 The discrepancy in fold change observed for CK-M between TMT mass spectrometry data and previously reported SomaScan data could be due to the fact that different techniques measure different things. In DMD, circulating creatine kinase exists as CK-M and CK-B and as homodimers and heterodimers. SomaScan might lack specificity and might measure both CK-M and CK-B and the different dimers. Indeed, we have superimposed SomaScan data of CK-M and CK-B using the same set of samples and the R 2 was almost equal to 1 (data not shown), indicating that SomaScan was not able to distinguish between circulating CK-M and circulating CK-B. Mass spectrometry, on the other hand, enabled measurement of CK-M and CK-B separately. Further experiments using highly reliable absolute quantification methods are needed to resolve this discrepancy seen in the CK fold changes between DMD patients and age-matched healthy controls using these different techniques. Other elevated biomarker candidates identified in this study included some glycolytic enzymes such as ATP-dependent PFKM, UGP2, GSTM2, and PSMB1. These enzymes are also abundant in skeletal muscle and their release into the blood circulation could reflect sarcolemma instability.

In this study, we have also identified 50 proteins whose circulating levels were significantly decreased in young DMD boys compared to controls. Several of these decreased biomarker candidates in DMD patients relative to controls, namely, CNPD1, AGT, SPP1, GSN, OMD, and PLG, were previously reported to be decreased in young DMD patients relative to controls using an independent SomaScan technique. 14 Additionally, several proteins that were found to be decreased in DMD patients relative to controls using TMT technology were reported to be unchanged in their levels between DMD patients and controls using the SomaScan technique. 14 These included carbonic anhydrase 2 (CA2), TKT, FCN1, and galectin-3-binding protein (LGALS3BP). This discrepancy between the TMT mass spectrometry assay and SomaScan assay could be due to the epitope effect that might hinder the binding of the aptamer in the SomaScan assay in case the target is in complex with another protein and thus provide different quantitative results than mass spectrometry where all proteins are denatured and dissociated before analysis. The newly identified biomarkers that decreased in untreated young DMD patients relative to controls included TMSB4X, COTL1, CFL1, TKT, BLVRB, among others. The serum biomarker proteins that were found to be decreased in young DMD patients relative to controls have mostly enzymatic activity, for instance, already described DMD biomarkers such as CNDP1 9 but also new biomarkers such as LTF, FUCA1, VNN1, PLG, CPN1, ENPEP, EXT1, and TKT. CPN1 is an enzyme that protects the body from potent vasoactive and inflammatory peptides. CPN1 cleaves Cterminal arginine and lysine residues from complement anaphylatoxins (inactivation of C3a and C4a with significant reduction of C5a, kinins, and CKM) found in the bloodstream. 34, 35 We hypothesize that the decrease in circulating CPN1 is to reduce the complement activation cascade, a retro control mechanism used to reduce inflammation and attack of organ following fibrosis, muscle damage, and protein leakage in the bloodstream. 34−36 A small proportion of decreased serum biomarkers are of muscle origin, for example, known DMD biomarkers such as GSN but also new biomarkers such as TMSB4X, COTL1, and CFL1. These groups of proteins all interact with and bind actin. As reported for circulating GSN, these proteins might be involved in scavenging toxic circulating actin. 37 Their decline in the circulation could be attributed to the fact that the formed complex with actin is cleared by the liver. 38 The last group of decreased biomarkers includes those involved in G protein-coupled receptor binding or transmembrane signaling receptor activity and scavenger receptor activity. This latest protein group includes new biomarkers such as SPN, MARCO, LGALS3BP, PCSK1N, and VTN. VTN is a glycoprotein largely found in serum or plasma, which is produced mainly in the liver. It has different roles in complement system regulation. 39 Reduced level of VTN in serum has been shown to be related to liver disease. 40 Indeed, in DMD and Becker muscular dystrophy (BMD), several severe cases of defects in hepatocytes and nonalcoholic fatty liver disease were reported in pediatric patients. 41, 42 Recently, cases of a reduction of the plasma VTN were reported in patients with myasthenia gravis (MG), a neuromuscular disease with similar symptoms as DMD, mainly muscle weakness and rapid fatigue. 43 Another protein found in this group, ProSAAS (PCSK1N), is a neuroendocrine with multiple functions including fetal neuropeptide pro. 44 ProSAAS is known to inhibit proprotein convertase 1 (PC1/ 3), a major protein produced in neuroendocrine cells, which regulates the proteolytic cleavage of neuroendocrine peptide precursors. 45, 46 PCSK1N is enzymatically cleaved into different neurosignaling peptides including PEN, LEN, and SAAS with roles as neurotransmitters. The alteration in PCK1N levels can possibly affect PC1/3 and lead to endocrine system dysregulation reported in DMD 47, 48 or in other neuromuscular diseases such as fibromyalgia (FM). 49 As ProSAAS-derived peptides, two peptides were identified and quantified, one of which was specific to ProSAAS (AEAQEAEDQQAR) and not to PEN, LEN, or SAAS fragment of neurotransmitters. This suggests that the change in the level of ProSAAS is driven by this peptide and not by the neurotransmitter peptides, which can bias ProSAAS protein quantitation. Taken together, these findings add deeper insight into the understanding of the DMD disease mechanism and pathology in young patients.

We correlated all serum biomarkers filtered based on 100% quantitative data completeness. This led to a data matrix of 59 proteins for 18 participants. The biomarkers' correlation to each other together with hierarchical clustering based on normalized intensity of each protein generated a map of 3481 Pearson correlation coefficients in total for both DMD patients and healthy controls. This unbiased analysis led remarkably to the clustering of proteins into two main groups. Our findings revealed impairments in the natural defense system, translated with decreases in the levels of a group of proteins in sera of DMD patients relative to healthy controls. These proteins are involved in cell adhesion and tissue remodeling (e.g., IGFALS, VTN, and PLG) as well as proteins involved in complement cascade activation, regulation, and inflammation (C4BPA, C4BPB, C2, and CPN1). This finding suggests a regression or deficit in immunity in DMD, as previously reported. 23 Also, we identified a group of proteins mostly located in the extracellular matrix and plasma membrane, such as AGT, CPN2, BASP1, SPP1 and CNDP1, FCN1, LGALS3BP, and DEFA1. These proteins are involved in vasoconstriction, extracellular matrix assembly, complement activation, cell defense response, organ tissue development (e.g., diaphragm), biomineral cell differentiation (e.g., osteoblast), and positive regulation of bone ACS Omega http://pubs.acs.org/journal/acsodf Article resorption. This finding indicates that organ growth and bone development are significantly disturbed in DMD. This result corroborates with clinical observation from other groups. 23−26 We recognize that sample size is one of the limitations of our current study. This is a recurrent question, especially when dealing with a rare disease such as DMD. We had some challenges enrolling and collecting more samples from 4 yearold boys during the period of the study. However, in this pilot study, although it is preliminary and uses a small sample size, we have confirmed previously reported biomarkers. We have also identified novel biomarkers that were significantly different between young DMD boys and age-matched healthy controls with a good P value adjusted for multiple testing. We have not noticed any large variations in the level of candidate biomarkers between patients. Nevertheless, follow-up studies using a larger sample size are needed to validate these novel biomarker candidates and define their physiological significance and relation with early-stage DMD pathogenesis.

The second limitation of our study is the identification and quantification of a lower number of serum proteins after the depletion step compared to previous studies. 50, 51 This is mainly due to differences in sample preparation between this study and previous studies. Indeed, previous studies used intensive sample fractionation (>80 fractions per sample) and online chromatographic separation during the fractionation step, requiring more resources and machine time than our current workflow where only 18 fractions were collected from the spin column and analyzed by LC−MS/MS. Our simplified method will be easy to implement by others for future validation studies. The third limitation of our study is the lack of a specific and validated ELISA assay to confirm newly identified candidate biomarkers. This is also a recurrent question when it comes to validating proteomics data. Biomarker discovery and validation are an ongoing effort and more specific and sensitive ELISA assays or alternative orthogonal methods are needed to achieve this goal.

We implemented and standardized a serum proteome profiling workflow, which implements serum-abundant protein depletion and the sensitivity gained from high-pH reversed-phase peptide fractionation and MS-based tandem mass tag (TMT) multiplex quantitation approach. We evaluated our workflow on 4 year-old untreated DMD cases and age-matched healthy controls to define candidate biomarkers associated with the early stage of the disease. A large number of these biomarkers confirmed the previous studies performed by other groups and us. However, several new biomarkers were identified in this study. These findings support the efficiency, reliability, sensitivity, and capability of our serum biomarker study workflow in biomarker discovery. Also, we brought new insight into the assessment of the disease pathogenesis in young DMD patients with a new panel of biomarker signatures such as proteases and actin-binding proteins. Further examination of serum protein biomarkers using our workflow in presymptomatic very young DMD patients less than 2 years old might provide further insight into muscle pathogenesis and provide tools to assess disease severity and response to intervention in DMD infants.

Collection. In this study, we used serum samples from a subset of young DMD patients (average age, 4.38 ± 0.24 years old; n = 9) and from age-matched healthy controls (average age, 4.47 ± 0.32; n = 9). DMD patients were enrolled through the Cooperative International Neuromuscular Research Group−Duchenne Natural History Study (CINRG-DNHS). 18 The study protocol was approved by Institutional Review Boards at all participating institutions and informed written consent was obtained from the parents of the participants or their legal guardians. Serum samples from healthy control pediatric donors were obtained from a third party company (BIOIVT, Hicksville, NY, USA). Detailed patient demographics and characteristics are listed in Table S1 . Serum samples were prepared following a rigorous standardized operating protocol and stored in small workable aliquots at −80°C in polypropylene cryogenic vials (Thermo Scientific Nalgene) to avoid repetitive freeze-thaw cycles. The commercial serum (catalogue #H4522) was purchased from Sigma-Aldrich.

Top 12 Serum-Abundant Protein Depletion and Quality Control. Protein concentration was determined using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, lL, USA) according to the manufacturer's instructions and measurements were performed using the SpectraMax i3x microplate reader (Molecular Devices, San Jose, CA). Serum aliquots containing 300, 600, or 900 μg of total proteins (corresponding to 5−14 μL of serum) were processed for depletion of the top 12 serum-abundant proteins using Pierce Top 12 abundant protein depletion spin columns (Thermo Fisher Scientific, lL, USA). Briefly, the depletion spin columns were equilibrated at room temperature (30 min), the column's screw caps were removed, and an adequate amount of total protein serum was directly added to the resin slurry into each column. Columns were recapped and inverted several times until the resin was completely suspended in the solution and then incubated with gentle end-over-end mixing for 90 min at room temperature using a tube revolver (Thermo Fisher Scientific, lL, USA) set at 19 rpm. After the incubation, the bottom closure of the columns was twisted off and the cap was loosened. The column was then placed in 2 mL Protein LoBind tubes (Eppendorf AG, Hamburg, Germany) and centrifuged at 1000g for 5 min to collect the remaining serum depleted sample (430−470 μL).

Sample processing quality control was checked by gel electrophoresis using precast NuPAGE 4−12% Bis-Tris protein gels (Thermo Fisher Scientific, lL, USA). Aliquots containing 3 μg of total proteins of depleted serum samples were mixed with 5 μL of NuPAGE LDS sample buffer and 2 μL of 500 mM Pierce dithiothreitol (DTT) (Thermo Fisher Scientific, lL, USA) and heated at 70°C for 5 min. Samples were uploaded into the wells and gel electrophoresis was run for 1 h using MOPS SDS running buffer (Thermo Fisher Scientific, lL, USA) at a constant voltage of 200 V in an Xcell SureLock Electrophoresis Cell (Thermo Fisher Scientific, lL, USA). The gel was stained with 15 mL of Bio-Safe Coomassie (Bio-Rad Laboratory, CA, USA) and then destained overnight with distilled water. Gel images were taken using an Azure c300 imager (Azure Biosystems, CA, USA) and analyzed using open-access software (GelQuant.NET).

In-Solution Digestion of Depleted Serum Samples and TMT Derivatization. All the reagents used in this experiment were from TMTsixplex Isobaric Mass Tagging Kit (Thermo Fisher Scientific, lL, USA) unless indicated. Aliquots of 140−250 μL containing 50 μg of total proteins from each sample were vacuum-dried to 30 μL and denatured with 20 μL of 1% sodium dodecyl sulfate (SDS) heated at 50°C for 20 min. Disulfide bonds were reduced by adding 5 μL of 200 mM Tris(2-carboxyethyl) phosphine (TCEP) followed by incubation at 55°C for 1 h. Free thiols were then alkylated by adding 5 μL of 375 mM iodoacetamide (IAA) followed by incubation at 40°C for 1 h in the dark. Proteins were then precipitated by adding 370 μL of pure chilled acetone (−20°C), followed by centrifugation at 15000 rpm for 30 min using an Eppendorf centrifuge 5424 (Eppendorf AG, Hamburg, Germany). The precipitate from each sample was kept at −20°C overnight (16 h). The mixture was then centrifuged again the next day and the supernatant was gently removed from the protein pellet. Then, 100 μL of 50 mM triethyl ammonium bicarbonate (TEAB) was added to the protein pellet and vortexed for 1 min. For protein digestion, a total of 20 μL of 0.05 μg/μL Pierce Trypsin Protease, MS Grade (Thermo Fisher Scientific, lL, USA), was added to each sample in two steps for a total enzyme/protein ratio of 1:50 (w/w). For the first 4 h of proteolysis, only 10 μL of the trypsin was added to each sample and the remaining 10 μL was added to each sample for overnight protein digestion. During the entire digestion process, samples were incubated at 37°C using a Thermomixer C (Eppendorf AG, Hamburg, Germany) set 450 rpm.

The peptide derivatization with TMT label reagents (Thermo Fisher Scientific, lL, USA) was performed according to the manufacturer's procedure with slight modification. Vials of TMT reagent (containing 0.8 mg of each tag reagent) were reconstituted in 100 μL of anhydrous acetonitrile and 50 μL of resulting solutions was added to the corresponding digested samples (50 μg of protein digested) and incubated at 27°C for 2 h with shaking at 450 rpm. To quench each reaction, 8 μL of 5% hydroxylamine was added to each sample and incubated for 1 h at 27°C with shaking. To multiplex the sample before the LC−MS/MS analysis, 42 μL of each TMT-labeled sample was combined in either 4-plex, 5-plex, or 6-plex in new Eppendorf tubes, dried, and stored at −80°C until high-pH reversedphase fractionation. For the 4 year-old serum study, combined samples shared in common the same digested control sample labeled with TMT 126 tag, which was selected for the normalization of data.

High-pH Reversed-Phase Peptide Fractionation. Vacuum-dried samples were reconstituted in 300 μL of 0.1% trifluoroacetic acid (TFA) and fractionation was performed using Pierce high-pH reversed-phase peptide fractionation kit (Thermo Fisher Scientific, IL, USA) according to the manufacturer's procedure with slight modification. Briefly, columns were conditioned with successive cleaning steps using pure ACN (300 μL twice) and washing steps with 0.1% TFA (300 μL twice) with centrifugation set at 1000g for a 4 min duration. Each sample (300 μL) was loaded onto the columns and centrifuged at 500g for 8 min. The operation was repeated once again and the flow-through (FT) was collected. The columns were then washed with water (200 μL twice) by centrifugation at 500g for 6 min, and the wash (W) was then collected. Peptide elution was performed with 300 μL of different elution buffer mixtures made with ACN and 0.1% triethlyamine (TEA) at 1000g centrifugation force for 8 min each. In total, 16 consecutive fractions were collected with elution buffers gradually increasing by 5% ACN increments up to 80%. The FT, W, and 16 collected factions were dried using a SpeedVac vacuum concentrator and stored at −20°C until nanoLC−MS/MS analysis.

Liquid Chromatography−Tandem Mass Spectrometry Analysis. Collected peptide fractions were reconstituted in 20 μL of water containing 0.1% formic acid (FA) (v/v) and 10 μL of samples was injected. Peptide separation was carried out using an Ultimate 3000 RSLCnano system (Dionex/Thermo Fisher Scientific). For each analysis, the sample was loaded into an Acclaim PepMap trap column (2 cm × 75 μm inner diameter, C18, 3 μm, 100 A; Dionex, CA, USA) at 3.5 μL/min with aqueous solution containing 0.1% FA and 2% ACN (v/v). After 8 min, the trap column was set online with an EASY-Spray Acclaim PepMap RSLC analytical column (50 cm × 75 μm inner diameter, C18, 2 μm, 100 A; Dionex, CA, USA). Peptides were eluted by applying a mixture of solvents A and B. Solvent A consisted of HPLC-grade water with 0.1% FA (v/ v), and solvent B consisted of HPLC-grade acetonitrile (80% ACN) with 0.1% FA (v/v). Separations were performed using a linear gradient of 5 to 50% solvent B at 300 nL/min over 100 min followed by a 4 min linear increase of ACN percentage up to a washing step (9 min at 99% solvent B) and a 5 min linear decrease of ACN percentage up to an equilibration step (20 min at 2% solvent B). The total analysis run time was 145 min. LC−MS/MS data-dependent acquisition was performed using a Q-Exactive HF mass spectrometer (Thermo Scientific, Bremen, Germany) in positive mode. For ionization, an EASY-Spray ES233 (Thermo Scientific, Bremen, Germany) was used with a voltage set at 2 kV, and the capillary temperature set at 350°C. Full MS scans were acquired in the Orbitrap mass analyzer over an m/z 375−1800 range with a resolution set at 60,000 for m/z 200. The target automatic gain control value of 3 × 10 6 was used with a maximum allowed injection time (Maximum IT) of 80 ms. For MS/MS, an isolation window of 1.2 m/z was utilized. The 20 most intense peaks with a charge state between 2 and 5 were selected for fragmentation using high-energy collision-induced dissociation with a stepped normalized collision energy of 27−32. The tandem mass spectra were acquired with fixed first mass of m/z 110 in the Orbitrap mass analyzer with a resolution set at 30,000 for m/z 200 and an automatic gain control of 5 × 10 5 . The ion intensity selection threshold was 9.1 × 10 3 , and the maximum injection time was 110 ms. The dynamic exclusion time was 45 s for the total run time of 145 min.

Mass Spectrometry Data Processing, Protein Identification, and Quantitation Analysis. All data files were collected and processed with a specific workflow designed in Proteome Discoverer 2.2 (Thermo Fisher Scientific). Searches were performed against Homo sapiens (TaxID = 9606) protein sequence database downloaded from www.uniprot.org on 15 November 2017 (42,182 entries) using Sequest HT (Thermo Fisher Scientific) with precursor and fragment mass tolerance respectively set at ±10 ppm and ±0.05 Da and the following dynamic modifications: carbamidomethyl on cysteine, acetyl on protein N terminus, oxidation on methionine, TMT 6-plex modification on lysine, and N terminus of peptides. The targetdecoy database search allowed us to control and estimate the false positive discovery rate at 1% for the peptide and protein as well. Quantitation of 4-plex (126, 127, 128, and 129), 5-plex (126, 127, 128, 129, and 130) , and 6-plex (126, 127, 128, 129, ACS Omega http://pubs.acs.org/journal/acsodf Article 130, and 131) were set up in Proteome Discoverer by the activation or deactivation of unused channels from original TMT 6-plex quantification method. The reporter ion quantifier tolerance was set at 0.05 Da with the integration method as the most confident centroid. The peptides used for the quantitation are only unique peptides and the spectra with missing channels were rejected. The reporter quantitation was corrected with isotopic impurity of reporter values for minimum signal-to-noise ratio (S/N) set at 1.5. The protein abundance was calculated as the sum of abundance of the corresponding protein in each fraction. All data were normalized to the same control sample labeled with TMT 126, which was used across all multiplexing analyses. The mass spectrometry proteomics data are made available via ProteomeXchange under the project named DMD Biomarkers in 4 year-old boys, DMD patient vs control study, with identifier PXD013174. Enzyme-Linked Immunosorbent Assay (ELISA). A subset of identified biomarker candidates including cofilin-1 (CFL1), ficolin-1 (FCN1), and plectin (PLEC) were examined using ELISA assay with kits corresponding respectively to catalogue numbers OKEH01091, OKDD00565, and OKBB002505 (AVIVA Systems Biology, San Diego, CA, USA) on the same serum samples analyzed by mass spectrometry. Serum samples from DMD patients and agematched healthy controls were diluted at 1:300 for CFL1, 1:250 for FCN1, and 1:400 for PLEC and the analyses were performed according to the manufacturer's instructions using the SpectraMax i3x microplate reader (Molecular Devices, San Jose, CA).

Western Blot. Aliquots containing 20 μg of total proteins from each sample including molecular weight markers, iBright Prestained Protein Ladder ranging from 11 to 250 kDa, were loaded in different wells of precast Criterion XT 12% Bis-Tris (1 mm, 18-well) gels. Gels were run at 200 V for 1 h and transferred onto the nitrocellulose membrane (Thermo Fisher) at 0.07 A for 17 h. Membranes were then washed with PBST and cut at the 55 kDa ladder marker for separate antibody incubations for haptoglobin and gelsolin. The membranes were then washed in Bio-Rad Blotting-Grade Blocker and separately incubated in anti-haptoglobin (ab131236) and anti-gelsolin (ab75832) at 1:10000 and 1:25000 dilutions, respectively. Membranes were washed again and incubated in anti-rabbit IgG-HRP (NA934V) antibodies at 1:3000 dilutions. Membranes were revealed using an Azure c300 imager (Azure Biosystems, CA, USA) and image-analyzed using open-access software (GelQuant.NET) with fluorescence settings at 10 s of absorbance.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.0c03206.

Supporting table descriptions; hierarchical clustering of other biomarkers found with missing data; Gene Ontology molecular function analysis (PDF) DMD patient and age-matched healthy control demographics; list of proteins identified and quantified in triplicate experiment using commercial serum; list of proteins identified and quantified in serum samples of young DMD patients relative to healthy controls; GO molecular function ontology (MFO) analysis of the 38 elevated and 50 decreased proteins in serum samples of DMD patients relative to healthy controls (XLSX) ■ AUTHOR INFORMATION Corresponding Author

United States; Email: yhathout@binghamton.edu Authors Tchilabalo D. Alayi − Department of Pharmaceutical Science, School of Pharmacy and Pharmaceutical Sciences

United States Holly R. Smith − Department of Pharmaceutical Science, School of Pharmacy and Pharmaceutical Sciences and Department of Biochemistry

Children's National Health System

Children's Hospital of Richmond at VCU

The CINRG-DNHS was funded by the U.S. Department of Education/NIDRR (#H133B031118 and #H133B090001)

Evidence-Based Path to Newborn Screening for Duchenne Muscular Dystrophy

The Pathogenesis and Therapy of Muscular Dystrophies

Corticosteroid Treatments in Males With Duchenne Muscular Dystrophy: Treatment Duration and Time to Loss of Ambulation

Advances in the Treatment of Duchenne Muscular Dystrophy: New and Emerging Pharmacotherapies

Restoring Dystrophin Expression in Duchenne Muscular Dystrophy: Current Status of Therapeutic Approaches

The 6-Minute Walk Test and Other Endpoints in Duchenne Muscular Dystrophy: Longitudinal Natural History Observations over 48 Weeks from a Multicenter Study

Revised North Star Ambulatory Assessment for Young Boys with Duchenne Muscular Dystrophy

Discovery of Serum Protein Biomarkers in the Mdx Mouse Model and Cross-Species Comparison to Duchenne Muscular Dystrophy Patients

Proteomics Profiling of Urine Reveals Specific Titin Fragments as Biomarkers of Duchenne Muscular Dystrophy

Muscle-Derived Proteins as Serum Biomarkers for Monitoring Disease Progression in Three Forms of Muscular Dystrophy

Affinity Proteomics within Rare Diseases: A BIO-NMD Study for Blood Biomarkers of Muscular Dystrophies

Aartsma-Rus, A. Tracking Disease Progression Non-Invasively in Duchenne and Becker Muscular Dystrophies

Longitudinal Serum Biomarker Screening Identifies Malate Dehydrogenase 2 as Candidate Prognostic Biomarker for Duchenne Muscular Dystrophy

The Fundamental Flaws of Immunoassays and Potential Solutions Using Tandem Mass Spectrometry

Cinrg Investigators. The Cooperative International Neuromuscular Research Group Duchenne Natural History Study-a Longitudinal Investigation in the Era of Glucocorticoid Therapy: Design of Protocol and the Methods Used

III Isobaric Labeling-Based Relative Quantification in Shotgun Proteomics

Proteomics Software Suite for in-Depth Mass Spectrometry Data Analysis Using Grid Computing

Biochemical Markers of Muscular Damage

Mass Spectrometry-Based Identification of Muscle-Associated and Muscle-Derived Proteomic Biomarkers of Dystrophinopathies

Role of Complements and Immunoglobulins in Duchenne Muscular Dystrophy

Short Stature in Duchenne Muscular Dystrophy: A Study of 34 Patients

Patterns of Growth in Ambulatory Males with Duchenne Muscular Dystrophy

Plectin Isoforms as Organizers of Intermediate Filament Cytoarchitecture

Plectin 1 Links Intermediate Filaments to Costameric Sarcolemma through -Synemin, -Dystrobrevin and Actin

Advantageous Uses of Mass Spectrometry for the Quantification of Proteins

A Review of Variant Hemoglobins Interfering with Hemoglobin A1c Measurement

Aptamers They Trust: The Caveats of the SOMAscan Biomarker Discovery Platform from SomaLogic. Circulation

Carbonic Anhydrase III in Serum in Muscular Dystrophy and Other Neurological Disorders: Relationship with Creatine Kinase

A Pleiotropic Regulator of Inflammation

Structure and Function of Human Plasma Carboxypeptidase N, the Anaphylatoxin Inactivator

Targeted Disruption of the Gene Encoding the Murine Small Subunit of Carboxypeptidase N (CPN1) Causes Susceptibility to C5a Anaphylatoxin-Mediated Shock

Plasma Gelsolin: Indicator of Inflammation and Its Potential as a Diagnostic Tool and Therapeutic Target

Role of Plasma Gelsolin and the Vitamin D-Binding Protein in Clearing Actin from the Circulation

The Complement FactsBook

Quantitation of Vitronectin in Serum: Evaluation of Its Usefulness in Routine Clinical Practice

The role of liver in muscular dystrophy

Duchenne and Becker Muscular Dystrophy Presenting as Nonalcoholic Fatty Liver Disease

Plasma Vitronectin Is Reduced in Patients with Myasthenia Gravis: Diagnostic and Pathophysiological Potential

Coexpression of Proprotein Convertase SPC3 and the Neuroendocrine Precursor ProSAAS 1

A Novel Function for ProSAAS as an Amyloid Anti-Aggregant in Alzheimer's Disease

Prohormone Convertase 1/3 Is Essential for Processing of the Glucose-Dependent Insulinotropic Polypeptide Precursor

Endocrine Aspects of Duchenne Muscular Dystrophy

Muscular Dystrophy Tracking and Research Network MD STAR. Bone Health and Endocrine Care of Boys with Duchenne Muscular Dystrophy: Data from the MD STARnet

Systematic Analysis of the Cerebrospinal Fluid Proteome of Fibromyalgia Patients

Quantitative Workflow for Sensitive Biomarker Discovery in Plasma Yields Novel Candidates for Early Myocardial Injury

The PRIDE Database and Related Tools and Resources in 2019: Improving Support for Quantification Data