Clinica Chimica Acta 436 (2014) 121–129

Contents lists available at ScienceDirect

Clinica Chimica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /c l inch im

S100A8 as potential salivary biomarker of oral squamous cell carcinomausing nanoLC–MS/MS

Yu-Jen Jou a,b, Chun-Hung Hua c, Chia-Der Lin c, Chih-Ho Lai d, Su-Hua Huang e, Ming-Hsui Tsai c,Jung-Yie Kao b,⁎, Cheng-Wen Lin a,e,⁎⁎a Department of Medical Laboratory Science and Biotechnology, China Medical University, Taichung, Taiwanb Department of Biochemistry, College of Life Sciences, National Chung Hsing University, Taichung, Taiwanc Department of Otolaryngology, China Medical University Hospital, Taichung, Taiwand Department of Microbiology, School of Medicine, China Medical University, Taichung, Taiwane Department of Biotechnology, College of Health Science, Asia University, Wufeng, Taichung, Taiwan

Abbreviations:OSCC, oral squamous cell carcinoma; Nliquid chromatography tandem mass spectrometry.⁎ Corresponding author.⁎⁎ Correspondence to: C.-W. Lin, Department of MeBiotechnology, China Medical University, No. 91, HsuTaiwan. Tel.: +886 4 22053366x7210; fax: +886 4 2205

E-mail address: cwlin@mail.cmu.edu.tw (C.-W. Lin).

http://dx.doi.org/10.1016/j.cca.2014.05.0090009-8981/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 7 December 2013Received in revised form 1 April 2014Accepted 6 May 2014Available online 24 May 2014

Keywords:Oral squamous cell carcinomaSalivaLC–MS/MSS100A8Biomarker

Background: Oral squamous cell carcinoma (OSCC) shows low 5-year survival; early treatment greatly reducesmortality andmorbidity. Saliva is a non-invasive sample,with good potential to discover biomarkers for early de-tection.Methods:NanoLC–MS/MS served to analyze saliva proteome from control subjects (n=35) andOSCCpatients T1(n= 29), T2 (n= 36), T3 (n= 14) and T4 (n= 21) stages. Identified biomarkers were verified byWestern blotand ELISA assays.Results: NanoLC–MS/MS analysis of salivary proteins between 10 and 15 kDa identified S100A8, hemoglobindelta and gamma-G globin in T3 and T4 stage OSCC as well as S100A7 in T1 and T2 stage OSCC. Western blotand ELISA indicated positive correlation between salivary S100A8 increment and tumor size stage. High levelof S100A8 appeared in 3.4, 13.9, 92.9, and 100% of saliva OSCC patientswith T1, T2, T3, and T4 stages, respectively.Significant increase of salivary S100A7 was observed in 20.7% and 11.1% of those with T1 and T2, respectively.

AUROC curve indicated high sensitivity, specificity and accuracy of S100A8-based ELISA as a detector.Conclusions: NanoLC–MS/MS, Western blot and ELISA manifested salivary S100A8 as a specific and sensitivemarker for detection of OSCC patients. Salivary S100A8 protein could be applicable in developing OSCCdiagnostics.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Oral squamous cell carcinoma (OSCC) is predominant in oral cancer,with N300,000 cases annually worldwide that account for 3% of malig-nancies in men and 2% in women [1–3]. Tobacco, alcohol, and betelquid are common risk factors for causing large areas of mucosal change,synergistically triggering oral carcinogenesis, even developing a second-ary upper aerodigestive tract cancer [4–9]. Histological and clinical dataindicate multi-step changes: leukoplakia, erythroplakia, hyperkeratosis,dysplasia, even carcinoma [10]. Despite advances in surgery, radiothera-py, and chemotherapy, overall average 5-year survival rate for patientshas not improved significantly (still approximately 50%) over the past30 years, far lower than that of laryngeal or nasopharyngeal carcinoma

anoLC–MS/MS, automated nano

dical Laboratory Science andeh-Shih Road, Taichung 404,7414.

[11,12]. Appropriate treatment for those with pre-malignant oral lesionsproves more effective, significantly raising the survival rate to 80–90%[12]. Consequently, early diagnosis of oral cancer, distinguishing be-tweenmalignant or premalignant lesions, is crucial to reduce themortal-ity andmorbidity. Biopsy of suspicious lesions offers the gold specimensfor the discovery of molecular biomarkers, but non-uniform appearanceof cancerous and precancerous lesions allows the difficulty in choosingthe location of biopsy, affecting the accuracy of potential OSCC markers[13,14]. Developing credible, accurate, cost-effective, and noninvasivetechniques for early detection is essential.

S100 A1-14 and B, a group of small acidic proteins, contain EF-handcalcium-binding motifs [15], modulating multiple biological propertiesin distinct cell- and tissue-types via binding with Ca2+, Zn2+, andCu2+. S100 proteins involve in calcium homeostasis and cytoskeletaldynamics, aswell as regulate cell proliferation and transcriptional factoractivity [16]. S100A4 protein regulates myosin dynamics by inhibitingprotein kinase C (PKC)-mediated phosphorylation on C-terminus ofmyosin heavy chain [17,18]. Secreted S100A4 is a candidate makerpredicting metastatic and prognostic potential in breast cancer [19].S100A7 is up-regulated in inflammatory epidermis, correlating with

122 Y.-J. Jou et al. / Clinica Chimica Acta 436 (2014) 121–129

epithelial malignancies: e.g., breast, skin, esophagus, head and neck [20].S100A8 and S100A9 can be synthesized and secreted by granulocytes,monocytes, and macrophages, identified as cytokine-like and transcrip-tional factor-like molecules affecting expression of tumor necrosisfactor-α, interleukin-1, and matrix metalloproteinases [21–24]. Up-regulation of S100A8 and S100A9 is found in gastric, colorectal, breast,lung, pancreatic, and prostate cancer, correlating with inflammationcell proliferation and metastatic processes in tumor development[25–27]. S100B also inhibits PKC-mediated phosphorylation on p53, re-ducing tumor suppressor activity by suppressing p53-dependent tran-scription activation [28,29]. Altered expression of S100 proteins isthereby associatedwith cancer development; secreted formof S100 pro-teins could act as potential cancer markers.

Tissue microarray indicated significant up-expression of S100A8 insevere oral dysplasias andOSCC tissue [30,31]. Proteomic analysis of nor-mal and OSCC tissues suggested S100A7 as a positive marker for OSCCcarcinogenesis and early tumor progression that can be confirmedby im-munofluorescence and quantitative RT-PCR analyses [32]. Sharp de-crease of S100A4 mRNA was evident in OSCC tissues [33]. Earlier weidentified S100A8, transferrin, and zincfinger protein 497 as salivary bio-markers, using two-dimensional gel electrophoresis (2DE) and massspectrometry (MS) [30]. Salivary samples from such patients showed el-evated S100A8, necessary for further probe or correlation with oral can-cer tumor grade. This study rated the potential of S100 proteins assalivarymarkers via proteomic analysis of lowmolecular weight salivaryproteins, using nanoLC–MS/MS. Protein profile indicated change ofS100A7 and S100A8 as unique markers in saliva of oral cancer patients.Western blot and direct binding ELISA further examined levels ofS100A7 and S100A8 in their saliva while evaluating potency of thesemarkers.

2. Materials and methods

2.1. Human subjects and saliva collection

In all, 35 subjects without and 100with OSCCwere enrolled for studyfrom February 2007 to March 2014; OSCC patients were diagnosed viabiopsy at China Medical University Hospital in Taichung. Subjects gaveinformed consent prior to saliva collection approved by the InstitutionalReview Board of China Medical University Hospital (permission number

Table 1Clinicopathological features of controls oral cancer patients.

Controls OSCC patients (100)

(35) T1 (29)

GenderMale 27 26Female 8 3

Mean age (year) 50.5 ± 13.6 51.2 ± 12.7Male 50.2 ± 14.5 50.8 ± 13.1Female 49.6 ± 10.3 54.7 ± 10.8

Cancer sitesOral cavity 27Mouth 1Oropharynx 1Hypopharynx 0Laryngeal 0

Node stageN0 25N1 1N2 3

Histology⁎WD 10MD 16PD 2Keratinizing 1

⁎ WD,Well-differentiated; MD, Moderately differentiated; PD, Poorly-differentiated.

DMR96-IRB-80). Exclusion criteria for OSCC patients and control individ-uals were followed, as in our prior studies [30,34]. Table 1 lists personalinformation (age, gender, and clinical features) of subjects. Tumor size(T) and nodal metastasis (N) staging for OSCC patients, as verified bypathological examination, was reviewed according to the universalTNM staging system of the International Union against Cancer (UICC)[35]. Collection protocol of salivary samples from patients and controlswere performed as in prior studies [30,34]: 5 ml collected in 15-ml cen-trifuge tube, mixed 5 μl Complete™ Protease inhibitor Cocktail (Roche),then centrifuged at 12,000 rpm (~13,400 × g) for 10 min at 4 °C. Theresulting supernatants were stored at−80 °C.

2.2. SDS-PAGE and Western blot

Preparation of each 500 μl sample for Western blot was performedaccording to the protocol of the 2-D clean up kit (Amersham). Salivaryproteins were precipitated by a combination of precipitant and co-precipitant, then rehydrated in 100 μl of rehydration buffer (8 mol/lurea, 4% CHAPS, 0.002% bromophenol blue). Afterward, 5 μg of salivaryproteins from each sample was mixed with sample buffer, heated at100 °C for 8 min, then loaded onto 12% SDS-PAGE gels stained withCoomassie blue after running. For Western blot, the dissolved proteinsin gels were transferred to nitrocellulose membranes. The resultantnitrocellulose membrane was blocked with 5% skim milk in Tris buff-ered saline (TBS) buffer containing 0.1% Tween 20 (TBST) at 4 °C for2 h, reacted with primary monoclonal antibodies anti-S100A7 and-S100A8 (Abnova) overnight, then incubated with HRP-conjugatedanti-mouse IgG antibodies (Invitrogen). Also, human sIgA as controlwas detected in salivary samples. Immune-reactive bands of interestwere detected by ECL™ Western Blotting Detection Reagents(Amersham).

2.3. In-gel digestion of low molecular weight salivary proteins

Since S100 were low molecular weight proteins, salivary proteinbands between approximate 10–15 kDa solved by 12% SDS-PAGEwere excised, washed twice with buffer (25 mM ammonia bicarbonate(ABC) in 50% acetonitrile (ACN)) for 5min and 100%ACN for 5min, thendried by speed vacuum concentrator. Proteins embedded in gels werereduced by reduction buffer (10 mmol/l dithiothreitol (DTT), 25 mM

T2 (36) T3 (14) T4 (21)

36 13 210 1 052.9 ± 11.0 56.9 ± 10.3 52.1 ± 11.152.9 ± 11.0 56.9 ± 10.7 52.1 ± 11.1

57.0

30 11 165 0 00 0 01 2 40 1 1

17 8 88 2 211 4 11

7 4 925 9 94 1 30 0 0

10

15

10

15

10

15

10

15

10

15

Control group

T1 group

T2 group

T3 group

T4 group

kDa subjects

K G T N Y L A D V F E K b1 b2 b4b3 b6b5 b8b7 b9 b10

y1y9 y8 y7 y6 y5 y4 y3 y2y10

b11

y11

A

B

Fig. 1. Peptide mass fingerprinting and peptide sequencing of salivary protein using nanoLC–MS/MS. Salivary samples from controls and OSCC patients were dissolved in 12% SDS PAGE(A), protein bands between 10 and 15 kDa subjected to nanoLC–MS/MS for identification. Representative LC–MS/MS spectra of S100A7 (B), S100A8 (C), and S100A9 (D) are shownwiththe calculated molecular weight (m/z values) along x-axis and relative intensity along y-axis. Amino acid sequence (Upper panel) was determined from mass differences in y- and b-fragment ions series.

123Y.-J. Jou et al. / Clinica Chimica Acta 436 (2014) 121–129

A L N S I I D V Y N Kb1 b2 b4b3 b6b5 b8b7 b9 b10

y1y9 y8 y7 y6 y5 y4 y3 y2y10

N I E T I I N T F H Q Y S V Kb1 b2 b4b3 b6b5 b8b7 b9 b10

y1y9 y8 y7 y6 y5 y4 y3 y2y10

b11 b12 b13 b14

y13 y12 y11y14

C

D

Fig. 1 (continued).

124 Y.-J. Jou et al. / Clinica Chimica Acta 436 (2014) 121–129

ABC) at 56 °C for 15 min, alkylated by solution (55 mmol/liodoacetamide (IAA), 25 mM ABC) at room temperature for 20 min inthe dark, and digested with fresh trypsin solution (2 ng/μl trypsin,

25 mmol/l ABC) at 37 °C overnight. Peptides were extracted by sonica-tion with 50–100% ACN and 0.1% formic acid (FA); released peptides insupernatants were dried by speed vacuum concentrator.

Table 2NanoLC–MS/MS identification of low molecular weight salivary proteins in samples from controls and OSCC patients at T1, T2, T3 and T4 stages.

Accession no. Protein name mol.weight⁎

Control T1 T2 T3 T4

Score Peptidesmatched

Score Peptidesmatched

Score Peptidesmatched

Score Peptidesmatched

Score Peptidesmatched

gi|235948 Cystatin SA-III = potential precursor ofacquired enamel pellicle

14,181 87 5 162 6 93 4

gi|183817 Beta-globin 15,870 80 3 104 3 103 7gi|4505821 Prolactin-inducible protein precursor 16,562 72 7 133 5 112 6 65 7 171 5gi|4506773 Protein S100-A9 13,234 71 12 145 8 249 9 50 9 134 5gi|190668 S100-A7 (psoriasin) 11,450 165 8 76 3gi|229172 Hemoglobin delta 15,897 123 5 197 7gi|31725 Gamma-G globin 16,959 58 3gi|957232 S100-A8 (oncodevelopmental protein) 13,086 57 2 110 5

⁎ mol. weight refers to molecular weight given in database, as inferred from gene sequence.

125Y.-J. Jou et al. / Clinica Chimica Acta 436 (2014) 121–129

2.4. NanoLC–MS/MS analysis of digested salivary proteins

Eluted peptides were subjected to nanoLC–MS/MS analysis, an inte-grated system (QSTAR XL) comprising LC Packings NanoLC systemwithautosampler and QSTAR XL Q-Tof mass spectrometer (AB Sciex) fittedwith nano-LC sprayer. The eluted samples were desalted on a LC-Packings PepMap™ C18 μ-Precolumn™ Cartidge (5 μm, 30 μm I.D. ×5 mm; Dionex, Sunnyvale, CA), separated on an LC-Packings PepMapC18 column (3 μm,15 cm×75 μm i.D.,) at 200 nl/min, using 45min gra-dient of 5–60% ACN in 0.1% FA and analyzed by connecting inline to amass spectrometer. NanoESI–MS and CID MS/MS sequencing of pep-tides were fully automated, synchronized with nanoLC runs underAnalystQS software. To identify proteins, 1 s survey scans were set atmass range m/z 400–1600 and 6 s MS/MS spectral acquisitions ofmultiple-charged precursor ions were detected at intensity abovepredefined threshold. Acquired individual MS/MS spectra within a sin-gle LC run were combined as a single Mascot-searchable peak list file.Peak list files served to query Swiss-Prot database via Mascot programwith these parameters: peptide mass tolerance of 150 ppm, MS/MSion mass tolerance of 0.15 Da, and allowing up to one missed cleavage.Minimum score above 20 was randomly set as acceptance threshold.

2.5. Relative quantitative analysis of salivary markers by ELISA

Each sample containing 5 μg/ml of protein was quadruplicate pro-coated in 96-well plates at 4 °C overnight, unbound substrates washedout with TBST thrice. After blocking with 5% skim milk in TBST, relativeamount of salivary markers was detected by monoclonal antibodiesanti-S100A7 and anti-S100A8. Anti-mouse IgG conjugated HRP was

S100A7

S100A8

sIgA

Fig. 2.Western blot analysis of S100A7 and S100A8 in saliva from controls and OSCC patients. Satransferred onto nitrocellulose membrane probed withmonoclonal antibodies to S100A7, S100cent HRP substrates. Lane 1: control group; Lane 2: T1 stage; Lane 3: T2 stage; Lane 4: T3 stage

added into each well for 2-h incubation at room temperature,immune-reactive complexesmeasuredwithABTS/H2O substrates. Opti-cal absorbance was recorded at 405 nm by ELISA plate reader (ELX 808,BioTek).

2.6. Immunohistochemical staining of S100A7 and S100A8 in OSCC tissues

Tissue sections from control, T1, and T3 cases underwent immuno-histochemical staining. Sections were deparaffined and rehydrated.After blockingwith 1% normal goat serum, dilutedmonoclonal antibod-ies anti-S100A7 or anti-S100A8 were added to tissue sections and incu-bated for 1 h, followed by HRP-conjugated anti-mouse IgG antibodiesand then 3,3′-diaminobenzidine (Sigma-Aldrich) as substrate.

3. Results

3.1. Clinical parameters of OSCC patients and control subjects

A total of 35 controls and 100 OSCC cases were recruited, the lattergrouped as T1, T2, T3 and T4, based on tumor size stage of UICC TNMstaging system (Table 1). Males formed a majority in all groups; meanage of OSCC cases was over 50 years, with controls (50.2 years) slightlyyounger. Among OSCC cases, tongue and buccal sites showed most fre-quent oral cancer lesions; nearly half were N0 stagewithout tumor cellsfrom regional lymph nodes, 13% N1 and 29% N2 stage. Histologic exam-ination of lesions showed moderately differentiated OSCC in 59% andwell differentiated OSCC in 30% of the cases. Table 1 lists other clinicalparameters of patients and controls.

OSCC

livary samples from each groupwere analyzed by 12% SDS-PAGE, then electrophoreticallyA8 and sIgA, and developed with HRP-conjugated secondary antibody and chemilumines-; Lane 5: T4 stage.

Table 3Number of positive ELLISA reaction of S100A7 and S100A8 in salivary samples from con-trols and OSCC patients with T1, T2, T3 and T4 stages.

S100 family⁎ Controls OSCC patients (100)

(35) T1 (29) T2 (36) T3 (14) T4 (21)

S100A7 0 6 (20.7%) 4 (11.1%) 0 0S100A8 1†(2.9%) 1 (3.4%) 5 (13.9%) 13 (92.9%) 21 (100.0%)S100A7/S100A8# 1 (2.9%) 7 (24.1%) 9 (25%) 13 (92.9%) 21 (100.0%)

⁎ Cut-off value set at OD405nm of 0.3.† Case diagnosed as OSCC patient 1 year post sampling.# A positive from either one or both markers.

126 Y.-J. Jou et al. / Clinica Chimica Acta 436 (2014) 121–129

3.2. NanoLC–MS/MS analysis of low molecular weight salivary proteins

To examine potential of salivary S100 as OSCC markers, samplesfrom control subjects and patients were analyzed by 12% SDS-PAGE,protein bands between 10 and 15 kDa excised and identified byNanoLC–MS/MS analysis system (Fig. 1). Table 2 details proteins fromin-gel trypsin digestion of low molecular weight salivary proteins insamples of each group by LC–MS/MS. Fig. 1B–D show peptide mass fin-gerprinting and peptide sequencing of S100A7, S100A8 and S100A9.Comparison of protein profile among groups indicated S100A7 as aunique marker of T1 and T2 stages and S100A8 as a potential markerof T3 and T4 stage in OSCC patients (Table 2). S100A9 appeared in sali-vary samples from control and all patient groups. Western blot analysisindicated significant rise of S100A8 in salivary samples from T3 and T4groups, S100A7 elevated in T1 but falling dramatically in T3 and T4stages (Fig. 2).

3.3. Relative levels of S100A7 and S100A8 in saliva from OSCC cases andcontrols

To correlate between S100A7/A8 and OSCC stage, relative levels ofS100A7 and S100A8 in the saliva of OSCC patients were measured via

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

OD

405

nm

S100A7

Control T1 T2 T3 T4

Control T1 T2 T3 T4

Stage

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

OD

405

nm

S100A8

Stage

A

B

Fig. 3.Direct binding ELISA to salivary S100A7 and S100A8 proteins. 2.5 μg of salivary pro-tein from each individual of control, T1, T2, T3, and T4 groupswere pre-coated ontomicro-titer plates, then incubatedwith 2000-fold dilution of monoclonal antibodies anti-S100A7or anti-S100A8 at room temperature for 1 h. Immune complexes were analyzed by ELISAprocedures.

binding ELISA, compared to those in the saliva of controls (Fig. 3, Sup-plemental Figs. 1 and 2). Binding ELISA indicated the significant changefor S100A7 and S100A8 in tumor size stage (Fig. 3), but no meaningfuldifferences in lymph node status and age (Supplemental Figs. 1 and2). Themean of OD405 nm indicated that S100A8 is significantly higherin the saliva of patients with T2-4 stages compared to T1 and controlsubjects (Fig. 3B). By contrast, salivary S100A7 peaked at T1 stage andreached its nadir in T3 and T4 stages (Fig. 3A). When cutoff OD indexof 0.3was recommended, positive cases of S100A8were 1 (2.9%) in con-trols, 1 (3.4%) in T1, 5 (13.9%) in T2, 13 (92.9%) in T3, and 21 (100%) inT4 (Table 3).Meanwhile, S100A7-positive cases tallied 6 (20.7%) in T1, 4(11.1%) in T2, but negative in other groups. Combination of the

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 1

Sens

itivi

ty

1-Specificity

S100A7

T1

T2

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 1

Sens

itivi

ty

1-Specificity

S100A8

T3

T4

B

A

Fig. 4.Receiver-operating characteristic curves of salivary biomarkers for predicting tumorsize stages. Salivary samples were analyzed via direct binding ELISA of S100A7 (A) andS100A8 (B).

127Y.-J. Jou et al. / Clinica Chimica Acta 436 (2014) 121–129

positivity from either one or both markers significantly improved thesensitivity in T1 and T2 stages.

3.4. ROC curve analysis of S100A8 as the potential salivary marker of OSCC

To rate sensitivity and specificity of S100A8 for OSCC detection, thearea under receiver-operating characteristic curves (AUROC) discrimi-nates between OSCC-free subjects and OSCC cases at each stage, usingELISA (Fig. 4). AUROC of S100A7 for predicting OSCC was 0.71 for T1(95% CI: 0.88–1.04), and 0.68 for T2 (95% CI: 0.89–1.01) (Fig. 4A), thatof salivary S100A8 0.99 for T3 (95% CI: 0.58–1.00), and 0.98 for T4(95% CI: 0.63–1.06), respectively (Fig. 4B). This indicated S100A8-based ELISA as highly accurate in detecting OSCC at T3 and T4 stages.

3.5. Overexpression of S100A8 in OSCC tissues

Immunohistochemical staining further validated levels of S100A7and S100A8 in oral tissues from OSCC and control groups (Fig. 5). Up-regulation of S100A7 was observed in OSCC tissues from T1, but notcontrol and T3 groups. By contrast, strong positive reaction for S100A8was observed in tissues from the patients, not the controls. The resultsindicated that S100A8 is moderately expressing in T1 and overexpress-ing in T3 stage tissue, in agreement withWestern blot and ELISA analy-ses of the saliva from OSCC patients and controls.

4. Discussion

This study was the first report that nanoLC–MS/MS utilized to iden-tify salivary protein filing of OSCC patients and controls. NanoLC–MS/MS served as powerful tools to discover potential serum, urine, and tis-sue biomarkers for colon, lung, breast, colorectal, and gastric cancers[36–39]. In the present study, nanoLC–MS/MS analysis of salivary pro-teins between 10 and 15 kDa identified S100A8 as a potential salivary

Control T1 sta

S100A7

S100A8

Fig. 5. Immunohistochemical analysis of S100A7 and S100A8 expression in OSCC. Tissue sectiotochemical staining withmonoclonal antibodies anti-S100A7 and anti-S100A8. Immunoreactivnobenzidine as substrate, figures photographed under ×100 magnification.

biomarker for oral cancer (Fig. 1 and Table 2), in accordance with ourprior report: up-regulation of S100A8 in OSCC patients' saliva identifiedby two-dimensional gel electrophoresis (2DE) andmatrix-assisted laserdesorption/ionization time-of-flight mass spectrometry (MALDI-TOFMS) analyses [30]. Western blot showed that salivary S100A8 signifi-cantly elevated in patients compared to controls (Fig. 2). Further ELISAevaluation indicated the mean amount of S100A8 in the saliva of OSCCpatients as higher than in healthy controls (Fig. 3), while AUROC ofS100A8-based ELISA for the diagnosis of OSCC revealed that salivaryS100A8 serves as OSCC biomarker (Fig. 4). The results indicated highsensitivity, specificity, and accuracy of S100A8 in detecting OSCC. Inthe healthy control group, one case (2.9%) with high level of salivaryS100A8 was followed up and diagnosed as OSCC 1 year post sampling.Therefore, S100A8 proves valuable as the target of OSCC diagnostics.

Comparison of salivary protein profiling between OSCC and controlindicated increment of S100A8 and hemoglobin delta in T3 and T4 aswell as that of S100A7 in T1 and T2 groups. S100A8, a calcium-binding and pro-inflammatory protein, was overexpressed in prostate,breast, lung, gastric, pancreatic and colorectal cancers [40], hence sug-gested as one critical tissue marker for aggressive breast cancer pheno-type [41], discriminatory ovarian cyst fluid marker for ovarian tumors[42], and serological marker for colorectal and gastric cancers [43,44].In this study, quantity of S100A8 in salivary samples was validated bydirect binding ELISA (Fig. 3). NanoLC–MS/MS and ELISA analyses re-vealed positive correlation between tumor size stage of OSCC and sali-vary S100A8 protein level. Comparison of salivary S100A8 with thereported candidate marker transferrin [30] indicated that S100A8 hadhigher specificity but lower sensitivity in detecting T1 and T2 stageOSCC than transferrin. Salivary transferrin elevated in patients with pa-rotid mixed tumor, alcoholic liver cirrhosis, nasal polyposis and bron-chial asthma [30]. The combination of salivary S100A8 and transferrincould rule out the false positive and confirm the diagnosis of OSCC.Meanwhile, AUROC curve indicated that specificity, sensitivity, and

ge T3 stage

ns from healthy controls and OSCC patients with T1 or T3 stage, analyzed by immunohis-ity complex was developed by HRP-conjugated anti-mouse IgG antibodies and 3,3′-diami-

128 Y.-J. Jou et al. / Clinica Chimica Acta 436 (2014) 121–129

accuracy of salivary S100A8 in detecting oral cancer were similar tothose of the reported candidate OSCC markers [13,14]. Immunohisto-chemical staining also demonstrated that S100A8 moderately up-regulated in OSCC tissues from T1 and overexpressed in T3 stage(Fig. 5). Immunohistochemical staining implied increment of salivaryS100A8 derived from secreted proteins of OSCC tissues in OSCCpatients.

Previous reports demonstrated S100A7 overexpressing in tissues oforal lesions [45] as well as regulating invasion and growth of breastand prostate cancers [46,47]. Our study showed low increment of sali-vary S100A7 in OSCC cases at T1 stage as well as low expression of it inT1 stage tissues (Figs. 3 and 5). Also, hemoglobin delta and gamma glo-bin were extracted from T3- and/or T4-stage patients' saliva (Table 2).Embryonic and fetal hemoglobinwere expressed in human glioblastomamultiform cells [48]. The role of S100A7, hemoglobin delta, and gammaglobin in OSCC diagnosis merits further examination. Salivary bio-markers have great potential for detection and surveillance of OSCC pro-gression or recurrence. This study explored the profiling of low-weightsalivary proteins between 10 and 15 kDa, using nanoLC–MS/MS; eluci-dating whole saliva proteome can identify potential biomarkers for de-tection or progression of OSCC as well as pathophysiology.

In summary, nanoLC–MS/MS analysis of the saliva from OSCC pa-tients and controls identified S100A8 as a potential salivary biomarkerfor the diagnosis of human oral cancer in humans. Salivary proteinlevel of S100A8 in OSCC patients and controls was confirmed by West-ern blot and ELISA. In addition, salivary S100A8 levels strongly correlat-ed with tumor size stage. AUROC curve exhibited high specificity,sensitivity, and accuracy of S100A8-based ELISA for detecting oralcancer.

Acknowledgments

This work was supported by the National Science Council ofTaiwan (NSC101-2320-B-039-036-MY3) and China Medical Uni-versity (CMU101-ASIA-05 and CMU101-S-24).

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.cca.2014.05.009.

References

[1] Greenlee RT, Murray T, Bolden S, Phyllis A. Cancer statistics, 2000. CA Cancer J Clin2000;50:7–33.

[2] Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics, 2002. CA Cancer J Clin2005;55:74–108.

[3] Cancer facts and figures 2006. Atlanta: American Cancer Society; 2006.[4] Ho PS, Yang YH, Shieh YT, et al. Consumption of areca quid, cigarettes, and alcohol

related to the comorbidity of oral submucous fibrosis and oral cancer. Oral SurgOral Med Oral Pathol Oral Radiol Endod 2007;104:647–52.

[5] Zhang X, Reichart PA. A review of betel quid chewing, oral cancer and precancer inMainland China. Oral Oncol 2007;43:424–30.

[6] Haughey BH, Gates GA, Arfken CL, Harvey J. Meta-analysis of second malignant tu-mors in head and neck cancer: the case for an endoscopic screening protocol. AnnOtol Rhinol Laryngol 1992;101:105–12.

[7] Jones AS, Morar P, Phillips DE, Field JK, Husband D, Helliwell TR. Second primary tu-mors in patients with head and neck squamous cell carcinoma. Cancer1995;75:1343–53.

[8] Jovanovic A, van der Toll G, Kostense JP, et al. Second respiratory and upper digestivetract cancer following oral squamous cell carcinoma. Eur J Cancer B Oral Oncol1994;30B:225–9.

[9] Tepperman BS, Fitzpatrick PJ. Second respiratory and upper digestive tract cancersafter oral cancer. Lancet 1981;2:547–9.

[10] Tsantoulis PK, Kastrinakis NG, Tourvas AD, Laskaris G, Gorgoulis VG. Advances in thebiology of oral cancer. Oral Oncol 2007;43:523–34.

[11] Canto MT, Devesa SS. Oral cavity and pharynx cancer incidence rates in the UnitedStates, 1975–1998. Oral Oncol 2002;38:610–7.

[12] Silverman SJ. Demographics and occurrence of oral and pharyngeal cancers. The out-comes, the trends, the challenge. J Am Dent Assoc 2001;132:7S–11S.

[13] Yen CY, Huang CY, HouMF, et al. Evaluating the performance of fibronectin 1 (FN1),integrin α4β1 (ITGA4), syndecan-2 (SDC2), and glycoprotein CD44 as the potentialbiomarkers of oral squamous cell carcinoma (OSCC). Biomarkers 2013:63–72.

[14] Yen CY, Chen CH, Chang CH, et al. Matrix metalloproteinases (MMP) 1 and MMP10but not MMP12 are potential oral cancer markers. Biomarkers 2009;14:244–9.

[15] Hsieh HL, Schäfer BW, Sasaki N, Heizmann CW. Expression analysis of S100 proteinsand RAGE in human tumors using tissue microarrays. Biochem Biophys ResCommun 2003;307:375–81.

[16] Salama I, Malone PS, Mihaimeed F, Jones JL. A review of the S100 proteins in cancer.EJSO 2008;34:357–64.

[17] Kriajevska M, Tarabykina S, Bronstein I, et al. Metastasis-associated Mts1 (S100A4)protein modulates protein kinase C phosphorylation of the heavy chain ofnonmuscle myosin. J Biol Chem 1998;273:9852–6.

[18] Murakami N, Kotula L, Hwang YW. Two distinct mechanisms for regulation ofnonmuscle myosin assembly via the heavy chain: phosphorylation for MIIB andmts 1 binding for MIIA. Biochemistry 2000;39:11441–51.

[19] Rudland PS, Platt-Higgins A, Renshaw C, et al. Prognostic significance of themetastasis-inducing protein S100A4 (p9Ka) in human breast cancer. Cancer Res2000;60:1595–603.

[20] Sneh A, Deol YS, Ganju A, et al. Differential role of psoriasin (S100A7) in estrogen re-ceptor α positive and negative breast cancer cells occur through actin remodeling.Breast Cancer Res Treat 2013;138:727–39.

[21] Odink K, Cerletti N, Brüggen J, et al. Two calcium-binding proteins in infiltrate mac-rophages of rheumatoid arthritis. Nature 1987;330:80–2.

[22] Roth J, Teigelkamp S, Wilke M, Grün L, Tümmler B, Sorg C. Complex pattern of themyelo-monocytic differentiation antigens MRP8 and MRP14 during chronic airwayinflammation. Immunobiology 1992;186:304–14.

[23] van Lent PL, Grevers LC, Blom AB, et al. Stimulation of chondrocyte-mediated carti-lage destruction by S100A8 in experimental murine arthritis. Arthritis Rheum2008;58:3776–87.

[24] Youssef P, Roth J, FroschM, et al. Expression of myeloid related proteins (MRP) 8 and14 and the MRP8/14 heterodimer in rheumatoid arthritis synovial membrane. JRheumatol 1999;26:2523–8.

[25] Koon N, Zaika A, Moskaluk CA, et al. Clustering of molecular alterations in gastro-esophageal carcinomas. Neoplasia 2004;6:143–9.

[26] Sy SM,Wong N, Lai PB, To KF, Johnson PJ. Regional over-representations on chromo-somes 1q, 3q and 7q in the progression of hepatitis B virus-related hepatocellularcarcinoma. Mod Pathol 2005;18:686–92.

[27] Moinzadeh P, Breuhahn K, Stützer H, Schirmacher P. Chromosome alter-ations in human hepatocellular carcinomas correlate with aetiology and his-tological grade-results of an explorative CGH meta-analysis. Br J Cancer2005;92:935–41.

[28] Grigorian M, Andresen S, Tulchinsky E, et al. Tumor suppressor p53 protein is a newtarget for the metastasis-associated Mts1/S100A4 protein: functional consequencesof their interaction. J Biol Chem 2001;276:22699–708.

[29] Lin J, Blake M, Tang C, et al. Inhibition of p53 transcriptional activity by the S100Bcalcium-binding protein. J Biol Chem 2001;276:35037–41.

[30] Jou YJ, Lin CD, Lai CH, et al. Proteomic identification of salivary transferrin as a bio-marker for early detection of oral cancer. Anal Chim Acta 2010;681:41–8.

[31] Driemel O, Escher N, Ernst G, Melle C, von Eggeling F. S100A8 cellular distribution innormal epithelium, hyperplasia, dysplasia and squamous cell carcinoma and its con-centration in serum. Anal Quant Cytol Histol 2010;32:219–24.

[32] Kesting MR, Sudhoff H, Hasler RJ, et al. Psoriasin (S100A7) up-regulation in oralsquamous cell carcinoma and its relation to clinicopathologic features. Oral Oncol2009;45:731–6.

[33] Sapkota D, Bruland O, Bøe OE, et al. Expression profile of the S100 gene family mem-bers in oral squamous cell carcinomas. J Oral Pathol Med 2008;37:607–15.

[34] Jou YJ, Lin CD, Lai CH, et al. Salivary zinc finger protein 510 peptide as a novel bio-marker for detection of oral squamous cell carcinoma in early stages. Clin ChimActa 2011;412:1357–65.

[35] Sobin LH, Wittekind CH. TNM classification of malignant tumors. New York: Wiley;2002.

[36] Li Y, Wen T, ZhuM, et al. Glycoproteomic analysis of tissues from patients with coloncancer using lectin microarrays and nanoLC–MS/MS. Mol Biosyst 2013;9:1877–87.

[37] Schaaij-Visser TB, deWit M, Lam SW, Jiménez CR. The cancer secretome, current sta-tus and opportunities in the lung, breast and colorectal cancer context. BiochimBiophys Acta 1834;2013:2242–58.

[38] Opstal-vanWinden AW, Krop EJ, Kåredal MH, et al. Searching for early breast cancerbiomarkers by serum protein profiling of pre-diagnostic serum; a nested case–con-trol study. BMC Cancer 2011;11:381.

[39] Zhang Z, Li M, Zhang G, et al. Identification of human gastric cacinoma biomarkersby differential protein expression analysis using 18O labeling and nanoLC–MS/MScoupled with laser capture microdissection. Med Oncol 2010;27:296–303.

[40] Grebhardt S, Müller-Decker K, Bestvater F, Hershfinkel M, Mayer D. Impact ofS100A8/A9 expression on prostate cancer progression in vitro and in vivo. J CellPhysiol 2013. http://dx.doi.org/10.1002/jcp.24489.

[41] Parris TZ, Kovács A, Aziz L, et al. Additive effect of the AZGP1, PIP, S100A8, andUBE2C molecular biomarkers improves outcome prediction in breast carcinoma.Int J Cancer 2013. http://dx.doi.org/10.1002/ijc.28497.

[42] Skaggs HS, Saunders BA, Miller RW, et al. Ovarian cyst fluids are a cache of tumorbiomarkers that include calgranulin A and calgranulin B isoforms. Cancer Invest2013;31:433–53.

[43] Kim HJ, Kang HJ, Lee H, et al. Identification of S100A8 and S100A9 as serologicalmarkers for colorectal cancer. J Proteome Res 2008;8:1368–79.

[44] Wang L, Chang EW, Wong SC, Ong SM, Chong DQ, Ling KL. Increasedmyeloid-derived suppressor cells in gastric cancer correlate with cancer

129Y.-J. Jou et al. / Clinica Chimica Acta 436 (2014) 121–129

stage and plasma S100A8/A9 proinflammatory proteins. J Immunol2013;190:794–804.

[45] Kaur J, Matta A, Kak I, et al. S100A7 overexpression is a predictive marker for highrisk of malignant transformation in oral dysplasia. Int J Cancer 2013. http://dx.doi.org/10.1002/ijc.28473.

[46] Liu H,Wang L, Wang X, Cao Z, Yang Q, Zhang K. S100A7 enhances invasion of humanbreast cancer MDA-MB-468 cells through activation of nuclear factor-κB signaling.World J Surg Oncol 2013;11:93.

[47] Ye L, Sun PH, Martin TA, Sanders AJ, Mason MD, Jiang WG. Psoriasin (S100A7) is apositive regulator of survival and invasion of prostate cancer cells. Urol Oncol2013;31:1576–83.

[48] Emara M, Turner AR, Allalunis-Turner J. Adult, embryonic and fetal hemoglobin areexpressed in human glioblastoma cells. Int J Oncol 2013. http://dx.doi.org/10.3892/ijo.2013.2186.