Discovery and verification of serum differential expression proteins for pulmonary tuberculosis

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

YTUBE1349_proof ■ 4 August 2015 ■ 1/8

Tuberculosis xxx (2015) 1e8

Contents lists available at ScienceDirect

Tuberculosis journal homepage: http://intl.elsevierhealth.com/journals/tube

MOLECULAR ASPECTS

Q3 Q2

Discovery and verification of serum differential expression proteins for pulmonary tuberculosis Cuiping Li a, 1, Xiao He b, 1, Hongtao Li c, Yi Zhou c, Ning Zang c, Shuixiu Hu d, Yanyan Zheng a, Min He a, * a

School of Public Health, Guangxi Medical University, Nanning 530021, PR China Nanning Center for Disease Control and Prevention, Nanning 530023, PR China Medical Scientific Research Center, Guangxi Medical University, Nanning 530021, PR China d Longtan Hospital of Guangxi Zhuang Autonomous Region, Liuzhou 545001, PR China b c

a r t i c l e i n f o

s u m m a r y

Article history: Received 6 February 2015 Received in revised form 3 May 2015 Accepted 7 June 2015

Pulmonary tuberculosis (PTB) is a chronic disease and has remained a severe threat to public health. Valuable biomarkers for improving the detection rate are crucial for controlling this disease. The purpose of this study was to discover potential biomarkers in sera from PTB patients compared with pneumonia patients and normal healthy controls. A total of 336 human serum specimens were enrolled in this study. Differentially expressed proteins were identified using iTRAQ method combining with MALDI-TOF-MS. Data was analyzed using relative bioinformatics methods. Potential biomarkers were further validated by IHC, ELISA and Western blot. As a result, 489 non-redundant proteins were identified in the sera, and 159 of which could be quantified by calculating their iTRAQ ratios. Compared to the controls, 26 differentially expressed proteins were recognized among PTB patients, including 16 overexpressed proteins and 10 downregulated proteins. Analysis of their functional interactions revealed that 12 proteins appeared in the center of the functional network. One of these key proteins, sex hormone binding globulin (SHBG), was found to be significantly elevated among PTB patients as compared with the controls examined by IHC, ELISA and Western blot. This result was consistent with the iTRAQ result. An independent blinded testing set to examine serum SHBG by ELISA achieved an accuracy of 78.74%, sensitivity of 75.6% and specificity of 91.5% in diagnosing PTB. In summary, iTRAQ in combination with MALDI-TOF-MS technology can efficiently screen differentially expressed proteins in sera from the PTB patients. SHBG is suggested to be a possible and novel serum biomarker for PTB. © 2015 Published by Elsevier Ltd.

Keywords: Tuberculosis Differentially expressed proteins SHBG iTRAQ MALDI-TOF-MS

1. Introduction Tuberculosis (TB) is a severe infectious disease around the world. In 2011, there were 12 million prevalent cases and 8.7 million incident cases of TB in the world was estimated by the World Health Organization (WHO), and 1.4 million people died from TB. Also in 2011, China estimated a total of 1.2e1.6 million prevalent TB cases [1]. Active TB infection has became a major disease reservoir and potential threat to the health of population. A

* Corresponding author. School of Public Health, Guangxi Medical University, No. 22 Shuangyong Road, Nanning, PR China. Tel.: þ86 771 5358146; fax: þ86 771 5350084. E-mail address: [email protected] (M. He). 1 These authors contributed equally to this work.

sensitive and reliable diagnosis method is therefore needed to efficiently detect TB from the population. Three conventional methods are used for detecting PTB in clinics, i.e., chest radiography, the sputum smear acid-fast staining test, and mycobacterial culture [2e4]. Generally, chest radiography is not sensitive enough to distinguish PTB [2,5]. In addition, sputum smear microscopy examination is infeasible for some patients with the number of sputum bacteria less than 104, and for those patients including young children who do not expectorate sputum, because patients with smear-negative pulmonary tuberculosis (SNP-TB) testing results are still considered important sources of PTB transmission. Furthermore, although mycobacterial culture is decisive, it requires quite amount of bacteria's sample size and a longer detecting period. The low efficiency and poor accuracy of these conventional screening methods hinder the early and efficient detection of PTB to a great degree.

http://dx.doi.org/10.1016/j.tube.2015.06.001 1472-9792/© 2015 Published by Elsevier Ltd.

Please cite this article in press as: Li C, et al., Discovery and verification of serum differential expression proteins for pulmonary tuberculosis, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.06.001

55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

YTUBE1349_proof ■ 4 August 2015 ■ 2/8

2

C. Li et al. / Tuberculosis xxx (2015) 1e8

Detection of serum biomarkers is recognized as an effective approach for diagnosis, monitoring treatment efficacy and prognostic evaluation of PTB [5]. Some reports suggested that several markers in blood might be as important as TB markers, such as Creactive protein (CRP) [6], IP-10 [7], human epididymis protein 4 [8], activator receptor, procalcitonin [9], krebs von den Lungen-6 [10], serum amyloid A and transthyretin [11]. These TB-associated proteins, however, have not been well established in clinical application because of their inadequate sensitivity and specificity [7]. In particular, no biomarker so far was reported to be able to distinguish the differences among TB patients, normal controls and pneumonia. ITRAQ Labeling combining with MALDI-TOF-MS is able to quantify hundreds of proteins simultaneously, either relatively or absolutely. Thus, it has become a powerful analytical technique for discovering disease biomarkers and drug targets. This strategy has recently been employed successfully to identify new markers for certain diseases, such as adult obstructive sleep apnea [12], autoimmune diseases [13,14], head-and-neck cancer [15], lung Cancer [16], and hepatocellular carcinoma [17]. Our aim was to detect multiple differentially expressed proteins of TB in sera collected from smear-positive pulmonary tuberculosis (SPP-TB) patients, SNP-TB patients, and drug resistant tuberculosis (DR-TB) patients respectively in a single experimental setting by using iTRAQ Labeling and MALDI-TOF-MS technology. Differentially expressed proteins were further analyzed using bioinformatics methods, and SHBG protein was selectively validated by ELISA, Western blot and immunohistochemical staining (IHC). The predictive value of serum SHBG for distinguishing among PTB cases, pneumonia patients, as well as the normal controls was evaluated. 2. Materials and methods 2.1. Serum samples This study was approved by the Ethics Committee of Guangxi Medical University. Written informed consent from all participants was collected. In the PTB group, sera were collected from PTB inpatients in the Fourth People's Hospital of Nanning City in September 2008. All PTB patients were diagnosed according to the combined clinical criteria from the WHO [18] including clinical, radiological, and bacteriological examinations. Sera from pneumonia patients were collected from the First Affiliated Hospital of Guangxi Medical University. Sera collection was underwent a full diagnostic assessment. The diagnosis of pneumonia was based on the presence of acute signs and symptoms suggesting lower respiratory tract infection on admission and radiographic evidence of a pulmonary infiltrate that had no other known cause. Sera from a control healthy group were also collected from the Guangxi Medical University First Affiliated Hospital and also underwent a full diagnostic assessment to exclude tuberculosis and pneumonia. Screening study and Western blot validation study: A total of 87 participants (61 male and 26 female patients) were verified, including 30 healthy controls (mean age, 45.3 ± 3.6 years), 7 DR-TB (mean age, 42.7 ± 4.5 years), 10 SPP-TB (mean age, 43.2 ± 4.3 years), 30 SNP-TB (mean age, 42.8 ± 5.7 years), and 10 pneumonia cases (mean age, 42.8 ± 5.7 years). ELISA validation study: 207 participants (133 male and 74 female patients) were verified, including 32 healthy controls (mean age, 42.3 ± 3.6 years), 35 DR-TB (mean age, 42.7 ± 4.5 years), 55 SPP-TB (mean age, 42.4 ± 3.5 years), 70 SNP-TB (mean age, 40.1 ± 5.7 years), and 15 pneumonia cases (mean age, 43.5 ± 5.7 years). These five groups were matched for age and sex. There were 50 patients who did not accept PTB drug treatment including 33 male and 17 female patients (mean age, 44.7 ± 4.5 years) and 110 patients who accepted rifampicin treatment

including 74 male and 36 female patients (mean age, 43.6 ± 5.2 years). All of the subjects were diagnosed as HIV-negative. All patients and controls were from the same geographic region (Guangxi, China) and ethnic origin (Han ethnicity). The general characteristics of the SNP-TB, SPP-TB, DR-TB and pneumonia cases for the control group are shown in Table 1. IHC validation study: 50 participants (16 male and 34 female patients) were verified, including 10 controls (mean age, 43.6 ± 15.6 years), and 40 PTB patients (mean age, 53.4 ± 10.4 years). Controls included 4 cancer adjacent normal pneumonic tissues and 6 normal pneumonic tissues. 2.2. ITRAQ and LC-MS/MS Fourteen highest abundance proteins were extracted from the sera at room temperature with the Agilent 1200 HPLC system (Agilent Technologies, Waldron, Germany) and installed on the MARS Human 14 column (4.6 mm id  100 mm, Agilent Technologies, Inc.). Briefly, prior to immunodepletion, the pooled serum samples from each group were diluted with Buffer A at a ratio of 1:3 (Agilent Technologies, Inc.), transferred to a 0.22 mm spin filter, and then centrifuged at 16,000 g for 1 min to remove particles. After collection of the less abundant protein fractions, the MARS columns were washed and the bound proteins were eluted with 100% buffer B. The procedures were conducted according to the protocol provided by the manufacturer. The collected fractions were further concentrated and desalted using 3000 MWCO Hydrosart Vivaspin 2 spin concentrators (Sartorius Stedim Biotech, Gottingen, Germany) at 8000  g for 99 min for three times. On each occasion, the sample solution was buffer exchanged with 50 mM triethylammonium bicarbonate (TEAB, pH 8.5 buffer, SigmaeAldrich Corporation, Saint Louis, MO, USA). The concentrated samples were determined using the BCA Protein Assay kit (Pierce, Ill, USA), and each 100 mg of protein was packed into 1.5 mL Eppendorf (EP) tubes, and dried and stored at 20  C. Trypsin digestion and iTRAQ labeling were performed according to the manufacturer's protocol (Applied Biosystems) using 8-plex iTRAQ reagent and buffer kits (ABI, Framingham, MA). The samples were labeled with iTRAQ reagents as follows: controls, 113; DR-TB, 117; SPP-TB, 118; SNP-TB, 119; and pneumonia, 121. The peptide mixture was fractionated by SCX chromatography at a basic pH (pH 3) to reduce the complexity of the mixture using a polysulfoethyl column (2.1  200 mm, PolyLC, Columbia, MD). Data acquisition was accomplished in a positive ion mode using a 5800 analyzer equipped with time-of-flight (TOF)/TOF ion optics (Applied Biosystems). ProteinPilot software v.3.0 (AB Sciex) was used for protein identification and quantitation. The search parameters were set by MMTS as follows: homo sapiens, trypsin cleavage, and cysteine alkylation. Each mass spectrometry (MS/MS) spectrum was searched against the UniProt database. Proteins with at least one peptide and above the 95% confidence level (unused ProtScore > 1.3) were recorded. The normalization tools and statistical package from the ProteinPilot software were utilized. A p < 0.05 and an average iTRAQ ratio >9 or <0.1 were considered to be high significance. The protein was considered stably expressed if the differences in expression values in the triplicate experiments were less than 1. MS/MS data has been deposited in the PRIDE database with accession number PXD002298. 2.3. Informatics analysis The cellular component, molecular function, and biological process were annotated by the GO database (http://www. geneontology.org/). The proteineprotein interaction network was

Please cite this article in press as: Li C, et al., Discovery and verification of serum differential expression proteins for pulmonary tuberculosis, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.06.001

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

YTUBE1349_proof ■ 4 August 2015 ■ 3/8

C. Li et al. / Tuberculosis xxx (2015) 1e8

3

Table 1 Characteristics of TB, pneumonia patients and controls in this Study.

Screening study iTRAQ-labeled samples Number of patients Age* Sex (male/female) HIV-negative ELISA validation study Number of patients Age* Sex (male/female) HIV-negative

Healthy control

DR-TB

SPP-TB

SNP-TB

Pneumonia

113 30 45.3 ± 3.6 27/3 30

117 7 42.7 ± 4.5 4/3 7

118 10 43.2 ± 4.3 6/4 10

119 30 39.1 ± 5.7 17/13 30

121 10 42.8 ± 5.7 7/3 10

32 42.3 ± 3.6 14/10 32

35 42.7 ± 4.5 23/12 35

55 42.4 ± 3.5 35/20 55

70 40.1 ± 5.7 50/20 70

15 43.5 ± 3.7 11/4 15

DR-TB: drug resistant tuberculosis; SPP-TB: smear-positive pulmonary tuberculosis; SNP-TB: smear-negative pulmonary tuberculosis. * Age shown as years with mean ± SD.

analyzed by search tool for the retrieval of interacting genes/proteins (STRING) software (http://string.embl.de/). 2.4. IHC examined tissue concentrations of differentially expressed proteins Differentially expressed protein was selected for immunohistochemical verification using an independent sample set. Tissue chips for PTB (US Biomax) and antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA) were obtained commercially. Antibody was first optimized when diluting (the dilution factor was 1:150). For antigen retrieval, the buffer (1 mM EDTA, pH 8.0 or 0.01 M sodium citrate buffer, pH 6.0) was heated to about 95  C and then the slide arrays were placed in the buffer for 10e15 min. The blocking antibody (normal goat serum) was applied and incubated for 20 min at room temperature. Tissue sections were first incubated with the primary antibody for 1 h at room temperature, washed and then incubated with a biotin-conjugated secondary antibody at 20e37  C for 20 min. Slides were washed with three times PBS (0.1 M, pH: 7.4) between every step. Finally the sections were counterstained with the DAB Kit. The immunopositive staining was evaluated in five areas. Sections were scored as positive if cells showed immunopositivity in the cytoplasm, plasma membrane, and/or nucleus when judged independently by two scorers who were blinded to the clinical outcome. First, a quantitative score was determined by estimating the percentage of immunohistochemically positive stained cells: 0, <5% cells; 1, 5e25% cells; 2, 25e50% cells; 3, 50e75% cells; and 4, >75% cells. Second, the intensity of the staining was scored by evaluating the average staining intensity of the positive cells (0, none; 1, weak; 2, intermediate; and 3, strong). Finally, a total score (ranging from 0 to 12) was obtained by multiplying the quantitative score and the intensity score. A combined staining score of 0 was considered to be negative staining (no expression); a score between 1 and 4 was considered to be weak staining (expression); a score between 5 and 8 was considered to be moderate staining (expression); and a score between 9 and 12 was considered to be strong staining (high expression). The immunohistochemical data were subjected to statistical analysis using Wilcoxon's rank test. 2.5. Western blotting analysis Western blotting was performed as previously described [19]. The primary antibody, mouse anti-human monoclonal antibody, was used at a concentration of 1:2000 (Santa Cruz, USA). The secondary antibody, IRDye-labeled goat anti-mouse (LI-COR, USA), was used at a concentration of 1:10,000; a quantitative analysis was performed using Odyssey (LI-COR, USA).

2.6. ELISA examined serum concentrations of differentially expressed protein A human SHBG ELISA kit was purchased from R & D Systems. The dilution factor was 1:100 and performed in duplicate according to the ELISA manufacturer's instructions. It was used to measure concentrations of protein in each serum sample in the validation set (n ¼ 207, 32 healthy controls, 15 pneumonia and 160 PTB patients). To consider the effect of rifampicin, we made a comparison of serum SHBG concentration. PTB patients included 50 patients without rifampicin and 110 patients with rifampicin. The PTB patients with rifampicin meant that PTB patients received rifampicin treatment one year ago or longer and received rifampicin treatment again because of the recurrence of TB in our study. The serum samples of these retreated PTB patients were taken before receiving the second rifampicin treatment. The PTB patients without rifampicin meant that new PTB patients never received rifampicin treatment. SPSS version 20.0 was used for statistical analyses. The experimental data were presented as mean ± SD. All p-values were twotailed and a p-value of less than 0.05 was considered statistically significant. Parametric data were compared between different groups using one-way ANOVA. The serum levels of SHBG were used to construct a diagnostic model. The diagnostic scores of healthy control and pneumonia patients were set as “0” and those of pulmonary TB patients were set as “1”. ROC curve analysis was conducted in this study.

3. Results 3.1. Differentially expressed proteins According to the criteria for protein quantification, 398, 376 and 384 proteins were quantified using triplicate iTRAQ labeling and LC-MS/MS analyses. A total of 434 proteins could be repeatedly identified in run 1, run 2 and run 3. The false discovery rates (FDRs) for protein identification based on searching against a reversed database in the triplicate experiments were 0.0056, 0.0127 and 0.0143, respectively. In the sera of PTB patients compared to healthy controls and pneumonia patients, a total of 26 differentially expressed proteins were identified in both runs with an unused ProtScore >1.3, an average iTRAQ ratio >9 (upregulated) or <0.1 (downregulated), p < 0.05, and error factor <2. Of the 26 differentially expressed proteins, 16 proteins were upregulated and 10 were downregulated. CO1A1 was upregulated the most (TB/control: 87.1001) and IGJ (TB/control: 0.0107) was downregulated the most among PTB patients. Table 2 summarizes the detail information on the differentially expressed proteins, including accession

Please cite this article in press as: Li C, et al., Discovery and verification of serum differential expression proteins for pulmonary tuberculosis, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.06.001

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

YTUBE1349_proof ■ 4 August 2015 ■ 4/8

4

C. Li et al. / Tuberculosis xxx (2015) 1e8

Table 2 The differentially expressed proteins in PTB when compared with normal and pneumonia. Accession

Abbreviation

Protein name

Description

Mean

P-value

spjP02452 spjP07900 spjP08571

CO1A1 HSP90A CD14

Collagen alpha-1(I) chain Heat shock protein HSP 90-alpha Monocyte differentiation antigen CD14

87.1001 54.113 48.3463

0.0149 0.0226 0.0271

12 3 2

spjP02741 spjO75882

CRP ATRN

C-reactive protein Attractin

17.7606 17.378

0.01 0.0291

6 6

spjP02735 spjP20742

SAA PZP

Serum amyloid A protein Pregnancy zone protein

17.0133 16.1856

0.019 0.0176

6 43

spjP05155

IC1

Plasma protease C1 inhibitor

15.8995

0.0062

20

spjP02745

C1QA

12.3193

0.023

9

spjP01011 spjP02750 spjP02747

AACT A2GL C1QC

11.6974 10.429 9.8401

0.004 0.007 0.0076

82 22 16

spjQ96IY4 spjP48740

CBPB2 MASP1

9.7863 9.6573

0.0322 0.0337

5 3

spjP00450 spjP04278 spjP35542 spjP02671

CERU SHBG SAA4 FIBA

Complement C1q subcomponent subunit A Alpha-1-antichymotrypsin Leucine-rich alpha-2-glycoprotein Complement C1q subcomponent subunit C Carboxypeptidase B2 Mannan-binding lectin serine protease 1 Ceruloplasmin Sex hormone-binding globulin Serum amyloid A-4 protein Fibrinogen alpha chain

It belongs to the fibrillar collagen family It is molecular chaperone It belongs to the lipopolysaccharide (LPS) receptor It belongs to the pentaxin family It plays a role in melanocortin signaling pathways It is a major acute phase protein It belongs to the protease inhibitorfamily It belongs to the PDGF/VEGF growth factor family It is the first component of complement system. It belongs to the serpin family Its tissue specificity was plasma C1q associates with the proenzymes C1r and C1s to yield C1 It belongs to the peptidase M14 family It belongs to the peptidase S1 family

9.4011 9.2461 0.0824 0.0704

0.00001 0.0246 0.0009 0.0258

spjP02656

APOC3

Apolipoprotein C-III

0.0625

0.0057

6

spjO14791 spjP02654 spjP02655

APOL1 APOC1 APOC2

Isoform 2 of Apolipoprotein L1 Apolipoprotein C-I Apolipoprotein C-II

0.061 0.0484 0.0198

0.0039 0.0085 0.0203

3 6 7

spjP01857 spjP01593 spjP00739 spjP01591

IGHG1 KV101 HPTR IGJ

Ig gamma-1 chain C region Ig kappa chain V-I region AG Haptoglobin-related protein Immunoglobulin J chain

It is a blue, copper-binding glycoprotein It is an androgen transport protein It belongs to the SAA family It acts as a cofactor in platelet aggregation. It belongs to the apolipoprotein C3 family It belongs to the apolipoprotein L family It belongs to apolipoprotein C1 family It belongs to the apolipoprotein C2 family Its subcellular location is secreted It is a BenceeJones protein It belongs to the peptidase S1 family Its subcellular location is secreted

0.0122 0.0111 0.0111 0.0107

0.0441 0.0406 0.0479 0.0013

8 8 42 5

number, abbreviation, protein name, description of location and function, average iTRAQ ratio, P-value, and number of peptides detected. Figure 1 shows the identification of the SHBG peptide by MS/MS. 3.2. Functional classification and interaction of differentially expressed proteins The functional classifications of the 26 differentially expressed proteins were analyzed. Their main locations were in the extracellular region and extracellular region part, which are shown in Figure 2A. Major molecular functions were protein binding and enzyme inhibitory activity (Figure 2B). Two main biological processes were biological regulation and response to stress (Figure 2C). An interaction diagram was simulated for 26 differentially expressed proteins using the database described at www.string-db. org (Figure 2D). Proteins HSP90, CD14, IC1, C1QA, C1QC, CRP, CERU, SHBG, HPTR, APOC3, CBPB2 and FIBA appeared in the center of the functional network intersection, suggesting their important roles in the protein interactions. 3.3. SHBG as a biomarker for PTB validated by IHC Protein SHBG appeared in the center of the functional network intersection and the high expression levels of SHBG found among PTB patients were stable in the triplicate experiments of MS analysis. SHBG was chosen for further verification using immunohistochemistry to detect the expression levels of SHBG in an independent set of tissue specimens which includes PTB patients

No.of peptides

175 3 5 16

and controls. The results of IHC showed that 33 tissue specimens of PTB patients were considered to be strong staining and 7 tissue specimens of PTB patients were considered to be moderate staining; however, all controls was considered to be negative staining. Compared to the controls, there was a significant increase in the expression level of SHBG among PTB patients; this result also supports the above mass spectrometry (MS) findings (shown in Figure 3 and Table 2). 3.4. Western blotting In Western blotting, the expression level of SHBG in DR-TB, SPP-TB, SNP-TB groups was significantly increased when compared with the healthy controls and pneumonia group (Figure 4). This result was consistent with iTRAQ-LC- MS/MS consequence. 3.5. Serum concentrations for SHBG further examined by ELISA ELISA was performed to detect the expression level of SHBG in serum in one independent set of PTB patients, normal healthy controls and pneumonia patient serum samples. Serum SHBG concentrations in the DR-TB, SPP-TB, SNP-TB, pneumonia and control groups were 181.76 ± 200.09 nmol/L, 160.43 ± 75.45 nmol/ L, 159.75 ± 75.47 nmol/L, 60.63 ± 59.38 nmol/L and 33.39 ± 34.24 nmol/L, respectively. As shown in Figure 5A, compared with pneumonia and healthy control, the SHBG expression level progressively increased in the DR-TB, SPP-TB and SNP-TB groups (all p values < 0.0001), which was consistent with the

Please cite this article in press as: Li C, et al., Discovery and verification of serum differential expression proteins for pulmonary tuberculosis, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.06.001

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

YTUBE1349_proof ■ 4 August 2015 ■ 5/8

C. Li et al. / Tuberculosis xxx (2015) 1e8

5

4. Discussion

Figure 1. The identification of the peptide of SHBG by MS/MS. A. Quantitative information for peptide. B. The peptide fragments, including b-ion and y-ion series. The peptide digests in healthy control group, DR-TB group, SPP-TB, SNP-TB and pneumonia group were labeled with iTRAQ-113, 117, 118, 119 and 121 tags.

findings in the iTRAQ-LC-MS/MS analysis. However, there was no significant difference in SHBG expression levels among the DR-TB, SPP-TB and SNP-TB groups. Moreover, the serum SHBG concentration before rifampicin treatment was 164.17 ± 76.13, and the serum SHBG concentration after using rifampicin was 165.08 ± 128.17 nmol/L. SHBG expression levels showed no significant difference regardless of rifampicin drug therapy, as shown in Figure 5B. A diagnostic model was established for the analysis of the serum concentration data of SHBG (47 healthy controls and 160 PTB patients) using SPSS 20.0. The results showed that the AUC and critical value of the ROC curve for SHBG were 0.904 and 96.555 nmol/L, respectively. The accuracy, sensitivity and specificity of SHBG were 78.74%, 75.6% and 91.53%, respectively, in diagnosing PTB (Figure 5C).

As an approach to differentiate PTB patients, pneumonia patients, and healthy control subjects, the identification of proteins with altered expression is crucial in discovering new biomarkers to improve the accuracy of TB detection rate. Using a shotgun quantitative proteomics approach, this study identified proteins with expression altered in sera from PTB patients. The shotgun quantitative proteomics approach, namely, is a new proteomic technology based on differential labeling of peptides with iTRAQ reagents prior to their separation and analysis by multidimensional LC coupled to MS. Recently, a proteomics-based discovery of four new biomarkers in PTB serum was reported using a Surface Enhanced Laser Desorption/Ionization (SELDI)-based technique [11,20]. This SELDI method, although providing a convenient throughput that enables screening of a large number of samples, has the limitation of reducing the analysis to a specific proteomic subset. Only those proteins that remain bound to the selected protein chip could be detected, according to their biochemical properties. This approach leads to “MS peaks” whose corresponding protein cannot always be identified. Moreover, this technology is controversial due to several disadvantages, such as being short on standard procedures, timeconsuming and with low reproducibility. Unlike the SELDI technique, the iTRAQ-based method theoretically covers the entire proteomic profile of the samples. Biomarker discovery by iTRAQ Reagent Labeling coupled to LC-MS/MS has been documented to be reproducible [21]. Moreover, the results of validation demonstrate that candidate biomarkers which were screened and identified using iTRAQ-based method reliably reflect the developmental stages of disease, yielding a high sensitivity and specificity [16]. For patients with PTB, the iTRAQ-coupled 2D LCMS/MS technique has been applied successfully to identify a panel of potential serum biomarkers compared with healthy

Figure 2. Data mining of the set of PTB serum biomarker candidates. A. Predicted localization. B. Predicted molecular function. C. Predicted biological process. D. Interaction diagram of 26 differentially expressed proteins in PTB. Proteins HSP90, CD14, CRP, IC1, C1QA, C1QC, CBPB2, CERU, SHBG, FIBA, APOC3, and HPTR appeared in the center of the functional network intersection.

Please cite this article in press as: Li C, et al., Discovery and verification of serum differential expression proteins for pulmonary tuberculosis, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.06.001

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

YTUBE1349_proof ■ 4 August 2015 ■ 6/8

6

C. Li et al. / Tuberculosis xxx (2015) 1e8

Figure 3. Immunohistochemical verification of iTRAQ-discovered potential marker SHBG in PTB and controls tissues. Positive staining is brown and is intense in PTB tissues. A showed the distributions of negative staining, weak staining, moderate staining and strong staining in PTB and control tissues. B showed the PTB tissues. C showed the control tissues. All panels showed 200 magnifications. The results of IHC showed that 33 tissue specimens of PTB patients were considered to be strong staining and 7 tissue specimens of PTB patients were considered to be moderate staining; however, all controls was considered to be negative staining. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

controls and determine APOC2, CD5L, HABP2, and RBP4 to be potential protein biomarkers for PTB [22]. Gene-ontology was performed to obtain a better and deeper understanding of the biological significance of the 26 differentially expressed proteins. The data provided valuable information for the further study of molecular mechanisms that govern the normal to the infection process of Mycobacterium tuberculosis. Some previous studies have reported certain proteins may be potential biomarkers of PTB, which were also identified in our study. For example, as one member of the apolipoprotein family, APOC2 was found to be a potential protein biomarker of PTB [22]. Our results also showed similar trends with the finding of a previous study that indicated that the serum APOC2 expression level was downregulated. Heat shock proteins might be a potential biomarker in

Figure 4. Validation of the differential expression of SHBG protein by Western blot analysis. Western blot analysis was carried out using serum and antibodies against SHBG. The expression of SHBG has similar trends with the iTRAQ results. Each serum sample contained 50 mg proteins. The results of Western blotting revealed that the expression level of SHBG in DR-TB, SPP-TB, SNP-TB groups was significantly increased when compared with the healthy controls and pneumonia group.

pulmonary and extrapulmonary tuberculosis and may be useful for diagnosis and for understanding the pathogenesis of pulmonary and extrapulmonary TB [23]. HSP90 was also identified in our study. In addition to the above-mentioned proteins, we also found some other proteins that may be associated with PTB among the differentially-expressed proteins found, which are described in the following sections. MASP1 performs a key role in innate immunity by recognizing pathogens through patterns of sugar moieties and neutralizing them. IC1 may play a potentially crucial role in regulating important physiological pathways including complement activation, blood coagulation, fibrinolysis and the generation of kinins. Serum proteomic profiles from patients with active tuberculosis and normal controls were used in a SELDI-based technique to identify serum amyloid A protein and transthyretin as potential biomarkers for tuberculosis [11]. Interestingly, we also found serum amyloid A protein in our study. For other proteins, the relationships with disease were unclear. However, unlike the genome, the proteome is highly dynamic. It is very likely that proteins such as CRP are acute phase reactants stimulated by inflammation, which are not TB specific and only responding to environmental changes, pathological situations, etc. Therefore, these proteins may still hold out the promise of providing explanations for the biological mechanism underlying the PTB process. In this study, the interaction diagram showed that 12 proteins, HSP90, CD14, IC1, C1QA, C1QC, CRP, CERU, SHBG, HPTR, APOC3, CBPB2 and FIBA, appeared in the center of the functional network intersection. SHBG was not found in the key nodes in the protein interactions in this study, and there was also no previous evidence on the association between SHBG and PTB. However, the interaction diagram proteins showed that SHBG was closely linked to CRP, which CRP has been reported to be a PTB biomarker [24]. Moreover, SHBG is indirectly related to the apolipoprotein family including APOC1, APOC2 and APOC3. Recently, a fairly close involvement of PTB development with apolipoprotein metabolism has been found

Please cite this article in press as: Li C, et al., Discovery and verification of serum differential expression proteins for pulmonary tuberculosis, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.06.001

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

YTUBE1349_proof ■ 4 August 2015 ■ 7/8

C. Li et al. / Tuberculosis xxx (2015) 1e8

7

and members of this family may probably act as protective or pathogenic factors [25e27]. Therefore, SHBG may still play an important role in PTB. SHBG, functioning as an androgen transport protein, is produced by the liver and binds to testosterone with high affinity and with lower affinity to estradiol (E2). SHBG may also be involved in receptor-mediated processes and regulate the plasma metabolic clearance rate of steroid hormones by controlling their plasma concentration. Additionally, it has also been proposed as a biomarker candidate in diabetics [28]. Our results of IHC, ELISA and Western blot showed that SHBG was overexpressed in the PTB group but is present at significantly lower levels in healthy controls and in the pneumonia group. This result is consistent with the results of iTRAQ-LC-MS/MS. When SHBG was tested for its ability to distinguish the PTB patients from the controls and pneumonia group, the results yielded a high sensitivity and specificity in the blinded test set. The results also showed that 24.4% of PTB patients were not sensitive to the diagnostic model and 8.5% of the cases were false negatives. Moreover, SHBG was overexpressed in the DR-TB and SNP-TB group and yielded high sensitivity and specificity in the blinded test set, which is significant for the detection of DR-TB and SNP-TB. So far, limited research was on using serum proteomic spectra to classify DR-TB, SNP-TB patients, and non-TB pneumonia controls, and the association between SHBG and PTB was not reported. Our data is the first to suggest the potential application for SHBG in distinguishing PTB patients, pneumonia patients, and healthy controls. Two papers showed that SHBG in plasma was found to be increased by rifampicin treatment [29,30]. However, our results showed that the serum SHBG expression level was up-regulated among TB patients who did not accept drug treatment including rifampicin. Therefore, we may draw a preliminary conclusion that the increase in serum SHBG protein expression level was not associated with rifampicin treatment. In conclusion, this study was designed to screen serum biomarkers among PTB patients, establish respective diagnostic models to identify PTB patients with iTRAQ-LC-MS/MS, and explore the potential value in improving detection of DR-TB and SNP-TB. Twenty-six differentially expressed proteins were discovered among PTB patients. Twelve of these key proteins, such as HSP90, were associated with the development of PTB. Because SHBG obtained value as an indicator in differentiating PTB from pneumonia and healthy controls, it is suggested to be a possible novel serum biomarker for PTB. Acknowledgments We are thankful to all the participants of our study. The work was supported by grants from the Science and Technology Department of Guangxi and Nanning (Grant Nos.: 110400302,0991011,201109063C), and Program for Innovaive Research Team of Intellectual Highland in High School of Guangxi (Guijiaoren [2010]38 Funding:

None.

Competing interests: interest. Figure 5. Analysis of SHBG in serum. (A) Levels of SHBG were measured by ELISA in serum of healthy controls (n ¼ 24), pneumonia group (n ¼ 15), PTB patients (n ¼ 50) before rifampicin used, and PTB patients (n ¼ 110) after rifampicin used. (B) Levels of SHBG were measured by ELISA in serum of healthy controls (n ¼ 24), PTB patients (n ¼ 160), and pneumonia group (n ¼ 15). Median values were shown by a horizontal line. The p-value was calculated with the ManneWhitney U test. (C) ROC curve analysis. ROC curves analysis of SHBG to discriminate PTB patients from healthy controls and pneumonia. The AUCs were 0.904 for SHBG.

Ethical approval:

We declare that we have no conflict of

Not required.

References [1] 2012. Gtr. Http://www.Who.Int/tb/publications/global_report/gtbr12_main. Pdf. 2012.

Please cite this article in press as: Li C, et al., Discovery and verification of serum differential expression proteins for pulmonary tuberculosis, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.06.001

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 Q1 117 118 119 120 121 122 123 124 125 126 127 128 129 130

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

YTUBE1349_proof ■ 4 August 2015 ■ 8/8

8

C. Li et al. / Tuberculosis xxx (2015) 1e8

[2] Hopewell PC, Pai M, Maher D, Uplekar M, Raviglione MC. International standards for tuberculosis care. Lancet Infect Dis 2006;6:710e25. [3] Andreu J, Caceres J, Pallisa E, Martinez-Rodriguez M. Radiological manifestations of pulmonary tuberculosis. Eur J Radiol 2004;51:139e49. [4] Blumberg HM, Burman WJ, Chaisson RE, et al. American thoracic society/ centers for disease control and prevention/infectious diseases society of america: treatment of tuberculosis. Am J Respir Crit Care Med 2003;167: 603e62. [5] Heo EY, Chun EJ, Lee CH, et al. Radiographic improvement and its predictors in patients with pulmonary tuberculosis. Int J Infect Dis 2009;13:e371e376. [6] Peresi E, Silva SM, Calvi SA, Marcondes-Machado J. Cytokines and acute phase serum proteins as markers of inflammatory regression during the treatment of pulmonary tuberculosis. J Bras Pneumol 2008;34:942e9. [7] Hong JY, Lee HJ, Kim SY, et al. Efficacy of ip-10 as a biomarker for monitoring tuberculosis treatment. J Infect 2014;68:252e8. [8] Liu W, Yang J, Chi PD, et al. Evaluating the clinical significance of serum he4 levels in lung cancer and pulmonary tuberculosis. Int J Tuberc Lung Dis 2013;17:1346e53. [9] Lee SH, Lee EJ, Min KH, et al. Procalcitonin as a diagnostic marker in differentiating parapneumonic effusion from tuberculous pleurisy or malignant effusion. Clin Biochem 2013;46:1484e8. [10] Miwa S, Suzuki Y, Shirai M, et al. Assessment of serum kl-6 as a prognostic marker in pulmonary tuberculosis patients. Int J Tuberc Lung Dis 2013;17: 240e2. [11] Agranoff D, Fernandez-Reyes D, Papadopoulos MC, et al. Identification of diagnostic markers for tuberculosis by proteomic fingerprinting of serum. Lancet 2006;368:1012e21. [12] Jurado-Gamez B, Gomez-Chaparro JL, Munoz-Calero M, et al. Serum proteomic changes in adults with obstructive sleep apnoea. J Sleep Res 2012;21:139e46. [13] Serada S, Fujimoto M, Ogata A, et al. Itraq-based proteomic identification of leucine-rich alpha-2 glycoprotein as a novel inflammatory biomarker in autoimmune diseases. Ann Rheum Dis 2010;69:770e4. [14] Fernandez-Puente P, Mateos J, Fernandez-Costa C, et al. Identification of a panel of novel serum osteoarthritis biomarkers. J Proteome Res 2011;10: 5095e101. [15] Ralhan R, Desouza LV, Matta A, et al. Discovery and verification of head-andneck cancer biomarkers by differential protein expression analysis using itraq labeling, multidimensional liquid chromatography, and tandem mass spectrometry. Mol Cell Proteomics 2008;7:1162e73. [16] Zeng GQ, Zhang PF, Deng X, et al. Identification of candidate biomarkers for early detection of human lung squamous cell cancer by quantitative proteomics. Mol Cell Proteomics 2012;11. M111 013946.

[17] He X, Wang Y, Zhang W, et al. Screening differential expression of serum proteins in afp-negative hbv-related hepatocellular carcinoma using itraq -maldi-ms/ms. Neoplasma 2014;61:17e26. [18] Organization WH. International union against tuberculosis and lung disease; royal netherlands tuberculosis association: revised international definitions in tuberculosis control. Int J Tuberc Lung Dis 2001;5. [19] Mathews ST, Plaisance EP, Kim T. Imaging systems for westerns: chemiluminescence vs. Infrared detection. Methods Mol Biol 2009;536:499e513. [20] Liu Q, Chen X, Hu C, et al. Serum protein profiling of smear-positive and smear-negative pulmonary tuberculosis using seldi-tof mass spectrometry. Lung 2010;188:15e23. [21] Song X, Bandow J, Sherman J, et al. Itraq experimental design for plasma biomarker discovery. J Proteome Res 2008;7:2952e8. [22] Xu DD, Deng DF, Li X, et al. Discovery and identification of serum potential biomarkers for pulmonary tuberculosis using itraq-coupled two-dimensional lc-ms/ms. Proteomics 2014;14:322e31. [23] Shekhawat SD, Jain RK, Gaherwar HM, et al. Heat shock proteins: possible biomarkers in pulmonary and extrapulmonary tuberculosis. Hum Immunol 2014;75:151e8. [24] Wallis RS, Pai M, Menzies D, et al. Biomarkers and diagnostics for tuberculosis: progress, needs, and translation into practice. Lancet 2010;375:1920e37. [25] Perez-Guzman C, Vargas MH, Quinonez F, Bazavilvazo N, Aguilar A. A cholesterol-rich diet accelerates bacteriologic sterilization in pulmonary tuberculosis. Chest 2005;127:643e51. [26] Martens GW, Arikan MC, Lee J, Ren F, Vallerskog T, Kornfeld H. Hypercholesterolemia impairs immunity to tuberculosis. Infect Immun 2008;76: 3464e72. [27] Lahiry P, Cao H, Ban MR, et al. Apoc1 t45s polymorphism is associated with reduced obesity indices and lower plasma concentrations of leptin and apolipoprotein c-i in aboriginal canadians. J Lipid Res 2010;51:843e8. [28] Cottrell HN, Karnitis JV, Mobasseri MN. Sex hormone-binding globulin levels as a predictor for developing gestational diabetes mellitus. Obstet Gynecol 2014;123(Suppl. 1):143S. [29] Lonning PE, Bakke P, Thorse P, et al. Plasma sulfate levels of estradiol, estrone, estrone and sex hormone binding globulin in patients receiving rifampicin. J Steroid Biochem 1989;33:631e5. [30] Back DJ, Breckenridge AM, Crawford Francesca E, et al. The effect of rifampicin on the pharmacokinetics of ethynylestradiol in women. Contraception 1980;21:135e43.

Please cite this article in press as: Li C, et al., Discovery and verification of serum differential expression proteins for pulmonary tuberculosis, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.06.001

33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64