serum hmgb1 as a potential biomarker for patients with...

Research ArticleSerum HMGB1 as a Potential Biomarker for Patients withAsbestos-Related Diseases

Shibo Ying1 Zhaoqiang Jiang1 Xianglei He2 Min Yu1 Riping Chen1

Junqiang Chen1 Guoqing Ru2 Yuan Chen2 Wanyuan Chen2 Lijin Zhu1 Tao Li1

Yixiao Zhang1 Xinnian Guo1 Xianhong Yin3 Xing Zhang1 and Jianlin Lou1

1 Institute of Occupational Diseases Zhejiang Academy of Medical Sciences Hangzhou 310013 China2Department of Pathology Zhejiang Provincial Peoplersquos Hospital Hangzhou Zhejiang 310014 China3Jiading District Center for Disease Control and Prevention Shanghai 201800 China

Correspondence should be addressed to Jianlin Lou jianlinlou163com

Received 3 November 2016 Revised 11 January 2017 Accepted 19 January 2017 Published 1 March 2017

Academic Editor Vincent Sapin

Copyright copy 2017 Shibo Ying et al This is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

High-mobility group box 1 (HMGB1) functions as a proinflammatory cytokine and is one of the most intriguing molecules ininflammatory disorders and cancers Notably HMGB1 is a potential therapeutic target and novel biomarker in related diseasesHowever the diagnostic value of HMGB1 for benign andmalignant asbestos-related diseases (ARDs) remains unclear In this workwe detected preoperative serumHMGB1 levels in Chinese asbestos-exposed (AE) and ARDs populations and further evaluated thediagnostic value of HMGB1 in patients with certain types of ARDs including those with pleural plaques asbestosis or malignantmesothelioma (MM) The experimental data presented that the serum level of HMGB1 was significantly elevated in AE and ARDssubjects Our findings indicated that serum HMGB1 is a sensitive and specific biomarker for discriminating asbestosis- and MM-affected individuals from healthy or AE individuals In addition serum matrix metalloproteinases 2 and 9 are not correlated withHMGB1 inARDsThus our study provides supporting evidence forHMGB1 as a potential biomarker either for the clinical diagnosisof high-risk AE cohorts or for evaluating ARDs

1 Introduction

Asbestos exposure and asbestos-related diseases (ARDs) are aglobal health issueThe epidemic of ARDswill likely continueto increase for at least another decade in most industrializedcountries and for several decades in industrializing countriesbefore reaching its peak [1] Globally occupational exposureto asbestos causes an estimated 107000 deaths each yearMostdeaths result from certain types of ARDs such as asbestos-related lung cancer asbestosis and mesothelioma [1] InChina the estimated number of occupationally exposedworkers may be over one million while the prevalence andincidence of ARDs as well as the mortality from ARDsare expected to be substantial and are likely to increase[2] Malignant ARDs continue to present challenges in thearenas of occupational medicine public health and clinicalresearch and practice [3]These serious diseases require earlyprevention recognition and treatment in the future

High-mobility group box 1 (HMGB1) is a member ofthe high-mobility group superfamily which exists ubiq-uitously in the nuclei and cytosols of mammalian cells[4] HMGB1 is a DNA chaperone primarily found in thenucleus where it stabilizes nucleosomes and contributesto DNA transcription replication and recombination [4ndash6] Under certain physiological or pathological conditionsHMGB1 can translocate from the nucleus to the cytosol andthen be secreted into the extracellular environment [4 6]HMGB1 functions as a proinflammatory cytokine due to itsactive secretion by innate immune cells such as neutrophilsmonocytes macrophages and several tumor cells but it canalso be released passively during cell injury and death [7]Recent in vitro experimental evidence showed that asbestosfibers could induce programmed necrosis and inflammationin primary human mesothelial cells [8] Upon asbestosexposure HMGB1 translocates from the nucleus to the

HindawiDisease MarkersVolume 2017 Article ID 5756102 9 pageshttpsdoiorg10115520175756102

2 Disease Markers

cytosol and is released into the cell culture medium [8] Sub-sequently HMGB1 release induces macrophages to secretetumor necrosis factor-120572 (TNF-120572) which protects mesothelialcells from asbestos-induced cell death and triggers a chronicinflammatory response [8] Chronic inflammation is also thehallmark of asbestos deposition in tissue and contributes toits carcinogenesis [9]

As a possible mechanism in asbestos-induced inflam-mation HMGB1 may be secreted into the stroma andsubsequently makes its way into the systemic circulation[9] Recent clinical studies have shown that HMGB1 is apotential diagnostic or prognostic biomarker in a varietyof inflammatory disorders [10ndash12] and cancers [13ndash19] Onegroup from the USA provided the first evidence that HMGB1was highly expressed in both tissues and sera of mesothe-lioma patients [20] In a following report with similar resultsbased on the exposure of Japanese subjects to asbestos theHMGB1 protein levels in the sera from patients with diffusemalignant peritoneal mesothelioma were also reported to besignificantly higher than those of nonmesothelioma subjectswith a history of exposure to asbestos [21] Serum HMGB1was also a prognostic marker for Japanese patients withmalignant pleural mesothelioma (MPM) [22] However theclinical importance ofHMGB1 for other nonmalignantARDsis still unknown Little data is available for serum HMGB1levels from a Chinese asbestos-exposed (AE) population

In the present study we quantified serum HMGB1 levelsin a Chinese AE population and further evaluated the diag-nostic value of HMGB1 in patients with nonmalignant andmalignant ARDs including pleural plaques (PP) asbestosisand malignant mesothelioma (MM) We provide evidencesupporting the hypothesis that HMGB1 could be a candidatebiomarker in clinical diagnosis for MM and asbestosis

2 Materials and Methods

This study was conducted in accordance with the Declarationof Helsinki and was approved by theMedical Ethics Commit-tee of Zhejiang Academy of Medical Sciences

21 Study Population and Data Collection A total of 497subjects were recruited from several small neighboring townswith a long history of asbestos industry during the years1960 to 2012 in Southeast China [23] Chest X-rays weretaken for all of the subjects Nonmalignant asbestos-relateddiseases were diagnosed according to the AmericanThoracicSociety (ATS) criteria [24] MPM was diagnosed by at leasttwo independent pathologists according to the Guidelines forPathologic Diagnosis of Malignant Mesothelioma proposedby the International Mesothelioma Interest Group [25] Astandardized questionnaire focused on occupational historywas taken during a face-to-face interview The contents ofquestionnaires included age gender smoking and drinkinghabits occupation and jobs entailing past asbestos exposureFor those who were occupationally exposed to asbestosexposure duration was measured as the interval of yearsbetween the start and the end of AE jobs Potential expo-sure levels to asbestos were estimated according to theapproach in our previous report [26] Briefly we retrieved

periodic data of total dust concentrations in different work-shopsmanufactories available from 1984 to 2010 in thisarea The median dust concentrations were 07mgm3 in ahousehold textile workshop (small-scale) and 59mgm3 in atextile manufactory (large-scale) Based on the correspond-ing median dust concentration potential exposure levelswere calculated (in mgm3 times years) when applicable

22 Grouping Sets The study subjects were divided into sixgroups (1) a healthy control group that consisted of 71 healthysubjects without any known history of asbestos exposure andwith normal chest X-ray examination results (2) a group of170 subjects occupationally exposed to asbestos for less than10 years but with normal chest X-ray examination results (3)a group of 129 subjects occupationally exposed to asbestosfor over 10 years but with normal chest X-ray examinationresults (4) a group of 81 subjectswith PP found in their chests(5) a group of 31 patients diagnosed with asbestosis and (6) agroup of 15 patients diagnosed with MPM

23 Detection Using Blood Samples Blood samples werecollected using standard blood procurement protocols fromthe enrolled subjects who provided written informed consentfor participation in this approved study Five milliliters ofblood was collected from the antecubital vein After clot-ting these samples were centrifuged at 4000 rpm for 5minThe subsequent separated sera were stored at minus80∘C untildetection assays for HMGB1 and matrix metalloproteinases(MMPs) were performed The concentrations of HMGB1MMP2 andMMP9 in the sera weremeasured using enzyme-linked immunosorbent assays (ELISAs) according to themanufacturerrsquos instructions (Cloud-Clone Corp HoustonTX USA) All assays were performed in duplicate

24 Statistical Analysis Statistical analyses were performedusing GraphPad Prism 5 software (GraphPad Software IncLa Jolla CA USA) or SPSS Statistics 170 (SPSS Inc ChicagoIL USA) Qualitative data are reported as frequencies andproportion distributions The chi-square test was used forcomparison of the proportions between groups and thesignificance level was adjusted using the Bonferroni methodQuantitative data with nonnormal distributions are pre-sented as medians and interquartile ranges A Kruskal-Wallis test followed by Dunnrsquos multiple comparisons test wasconducted to compare the median values between groupsStatistical significance was defined as a two-sided 119875 lt 005Univariate andmultivariate linear regression analyses were fitto explore predictors ofHMGB1The receiver operating curve(ROC) method was conducted for differentiating ARDs andAE groups from the control group Areas under ROC curves(AUCs) were presented with 95 confidence intervals (CIs)An AUC of 10 indicates perfect discrimination whereas anarea with a confidence interval of 05 indicates that the testdiscriminates no better than chance The optimum statisticalcut-off value was selected when Youdenrsquos index (sensitivity +specificity-1) reached the maximum value The relationshipsamong the correlations betweenHMGB1MMP2 andMMP9were identified using Spearmanrsquos rank correlation coefficient(119903119904)

Disease Markers 3

Table 1 Basic characteristics of the subjects

Control(119899 = 71)

AE lt 10 Y(119899 = 170)

AE ge 10 Y(119899 = 129)

PP(119899 = 81)

Asbestosis(119899 = 31)

MPM(119899 = 15)

Age ymedian (IQR) 67 (8) 66 (8) 67 (7) 68 (10) 73 (16) 66 (15)e

GenderMale n () 44 (620) 47 (277)lowast 21 (163) 33 (407) 8 (258)lowast 7 (467)Female n () 27 (380) 123 (724) 108 (837) 48 (593) 23 (742) 8 (533)

Smoker n () 10 (145) 16 (99) 7 (56) 13 (171) 1 (33) 4 (267)Drinker n () 5 (73) 9 (56) 5 (40) 6 (79) 0 (00) 4 (267)Potential exposure levelmgm3-years (median(IQR))

NA 177 (319) 708 (839) 502 (770) 998771 119 (833)998771 110 (131)

Exposure duration y(median (IQR)) NA 5 (30) 14 (50) 10 (85) 998771 10 (100)998771 14 (125)998771

PP pleural plague AE asbestos-exposed MPM malignant pleural mesothelioma IQR interquartile range NA not applicable Y yearslowast versus control group 119875 lt 005 versus AE lt 10 Y group 119875 lt 005 998771 versus AE ge 10 Y group 119875 lt 005 e versus asbestosis group 119875 lt 005

3 Results

31 Characteristics of the AE and ARD Populations Thedemographic and occupational characteristics of subjectsincluding age gender smoker drinker distribution potentialasbestos exposure level and exposure duration in studygroups are presented in Table 1 The mean ages among thesegroups were significantly different (1205942 = 1337 119875 = 002)Subjects with asbestosis were significantly older (median age73 years) than the AE lt 10 years group (median age 66years 119875 lt 005) and the MPM group (median age 66years 119875 lt 005) The gender distributions were significantlydifferent among these groups (1205942 = 5015 119875 lt 001) andthe AE PP and asbestosis groups had larger proportions offemale subjects than the control group The reason for themarked proportional difference is that most asbestos hand-spinning workers are female and tend to be more heavilyexposed than males [23] Both the potential AE level and theexposure duration were significantly different among thesegroups (1205942 = 8399 119875 lt 00001 1205942 = 122319 119875 lt 00001)Of these three ARDs groups the PP group had the highestpotential AE level (502mgm3) every year In contrast MPMpatients the only malignant ARDs group in this study hadthe lowest potential AE level (110mgm3) Furthermorethe mean asbestos exposure durations in both the PP andasbestosis groups were 10 yearsTheMPMgroup had a longerexposure duration (14 years) than these two nonmalignantgroups

32 Serum HMGB1 as a Potential Biomarker for ARDs Toinvestigate whether serum HMGB1 functions as a potentialbiomarker for identifyingAEARD individuals or cohorts wetested HMGB1 levels in sera from all six groups As shownin Figure 1 the median levels of serum HMGB1 in the AEgroups 5006 ngmL (AE lt 10 years) and 5042 ngmL (AEge 10 years) were similar Both of these values were signifi-cantly higher than the median level of 4168 ngmL observedin healthy controls However no significant differences in

n = 15n = 31n = 81n = 129n = 170n = 71

lowastlowast

lowastlowast

lowastlowast

lowastlowast

lowastlowast

lowastlowast

100

80

60

40

20

0

HM

GB1

(ng

ml)

MPMAsbestosisPPA ge 10 YA lt 10 YHealth

Figure 1 HMGB1 levels in serum from individuals with pleuralplaques (PP) asbestosis andMPM exposed to asbestos and healthycontrols ELISAs shown were performed in parallel and blindlyBars show the median of HMGB1 levels Statistical significance wasdefined as two-sided lowastlowast119901 lt 001

HMGB1 levels were observed between the two groups withdifferent exposure durations (AE lt 10 and AE ge 10 years)These results strongly suggest that HMGB1 may be a usefulbiomarker for evaluating AE individuals

The serum HMGB1 levels in the PP group (medianlevel 4977 ngmL) were significantly higher than that ofthe control group but were approximately equal to that ofthe two AE groups HMGB1 levels in the asbestosis group(median level 5877 ngmL) were significantly higher thanthose of the control group the two AE groups and thePP group Similarly HMGB1 levels in the MPM group(median level 6023 ngmL) were also significantly higherthan those of the control AE and PP groups Neverthelessthere was no significant difference in HMGB1 levels between

4 Disease Markers

Table 2 Factors influencing serum levels of HMGB1 in univariable and multiple linear regression

Predictor variableslowast Univariable analysis Multivariable analysis

119887 SE 119875 value 119887 SE 119875 valueAge minus0001 0002 066Gender (femalemale) 005 003 007Smoking (nonsmokersmoker) minus007 004 008Drinking (nondrinkerdrinker) minus005 005 035PP group versus control group 020 004 00001 020 004 lt00001AE lt 10 y group versus controlgroup 023 004 lt00001 023 004 lt00001

AE ge 10 y group versus controlgroup 022 004 lt00001 022 004 lt00001

Asbestosis group versus controlgroup 044 006 lt00001 044 006 lt00001

MPM group versus control group 048 007 lt00001 048 007 lt00001lowastThe dependent variable was natural logarithm form estimated with forward stepwise method 119887 regression coefficient SE standard error

Table 3 AUC and cut-off value of HMGB1 comparing healthy controls AE and ARDs

AUC (95 CI) 119875 value Cut-off (ngml) Sensitivity () Specificity ()MPM versus healthy 094 (089sim099) lt00001 5216 100 8028Asbestosis versus healthy 088 (082sim095) lt00001 4711 9355 7042MPM versus AE lt 10 Y 081 (073sim090) lt00001 5229 100 5765MPM versus AE ge 10 Y 080 (072sim089) 00001 5239 100 5736Asbestosis versus AE lt 10 Y 074 (066sim083) lt00001 5843 6129 7706Asbestosis versus AE ge 10 Y 074 (064sim083) lt00001 5648 6452 7209PP versus healthy 068 (060sim077) 00001 4701 5926 7042MPM versus asbestosis 056 (039sim073) 05043 5164 100 2903AUC area under the curve CI confidence interval AE asbestos-exposed subjects ARDs asbestos-related diseases

the asbestosis and MPM groups Our findings indicated thatHMGB1 is a potential biomarker to screen certain subsetsof patients with ARDs (asbestosis and MPM) from AE andhealthy individuals However the HMGB1 serum level is notuseful for discriminating patients with MM from those withasbestosis

33 Multiple Factors May Influence the Serum Level ofHMGB1 To test whether the predicted factors including agegender smoking asbestos exposure duration and potentialexposure levels influence HMGB1 serum levels we per-formedunivariate andmultivariate linear regression analysesAs shown in Table 2 the regression coefficients standarderrors of coefficients and 119875 values were analyzed by linearregression In the results of univariate regression analysisdummy variables of PP AE asbestosis and MPM groupsshowed significant influences on the serum level of HMGB1with a positive coefficient In addition the results from themultivariate regression analysis revealed that dummy vari-ables of AE PP asbestosis andMPM groups had significantlypositive influences on serum HMGB1 levels However wedid not find significant differences between different groupsof potential exposure levels (1205942 = 170 119875 = 043) andexposure duration (1205942 =034119875 = 084) Taken together these

findings support the notion that HMGB1 may be a valuableindependent index to differentiate AE and ADRs patientsfrom healthy individuals

34 HMGB1 Is Sensitive and Specific for Discriminating ARDsfrom Controls To further evaluate the diagnostic value ofHMGB1 to discriminate ARDs from AE individuals andhealthy controls we calculated the sensitivity and specificityof HMGB1 as a potential biomarker According to the sizesof the AUC the comparisons were subsequently ranked(Table 3) Due to the considerably high diagnostic valuefour ROCs with high AUC (gt0800) were selected and areshown in Figure 2 The AUC for HMGB1 values was 094(95 confidence interval 089ndash103) for distinguishingMPMfrom healthy controls An HMGB1 level of 5216 ngmL wasdetermined as the best cut-off value with a sensitivity of100 and a specificity of 8028 The AUCs for HMGB1values to distinguish MPM from AE lt 10 years and ge10 yearswere 081 (95 CI 073ndash090) and 080 (95 CI 072ndash089)respectively

The best cut-off value was 5229 ngmL (sensitivity 100specificity 5765) for distinguishing MPM from AE lt 10years Similarly the optimal cut-off value was 5239 ngmL(sensitivity 100 specificity 5736) for distinguishing

Disease Markers 5

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0938

100806040200

MPM vs healthy

100 minus Ｍpecificity ()

(a)

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0883

100806040200

Asbestosis vs healthy


(b)

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0812

100806040200

MPM vs A lt 10 Y


(c)

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0804

100806040200

MPM vs A ge 10 Y


(d)

Figure 2 ROC curves of HMGB1 for distinguishing patients with asbestos-related diseases (ADRs) from AE individuals or healthy controlsFour ROCs with high AUC (gt0800) were selected and shown in panels (a) (b) (c) and (d) ROC receive operation curve AE asbestos-exposed AUC area under the curve vs versus Y years See details in text

MPM from AE ge 10 years In addition the AUCs for HMGB1values to distinguish asbestosis from subjects of AE lt 10years AE ge 10 years and healthy controls were 074 (95 CI066ndash083) 074 (95CI 064ndash083) and 088 (95CI 082ndash095) respectively (Table 3 Figure 2) Overall HMGB1 is agood diagnostic index to identify MPM patients among AEand healthy cohorts In addition HMGB1 is also a valuablebiomarker in distinguishing patients with asbestosis fromhealthy control subjects

35 Secreted MMP29 Are Not Correlated with HMGB1in ARDs MMPs are expressed in a variety of connectivetissue types and proinflammatory cells and are consideredas promising therapeutic targets due to their strong involve-ment in key pathological events Among the MMPs bothMMP2 and MMP9 function as secreted gelatinases whichare responsible for tissue remodeling and the degradation ofextracellular matrix proteins such as collagen and gelatinHMGB1 has been reported to mediate MMP2 and MMP9

expression in some inflammatory disorders and cancers [27ndash31] To test whether their levels correlated with HMGB1we further measured the concentrations of MMP2 andMMP9 in sera from the studied population Serum levelsof MMP2 and MMP9 are shown in Supplementary FiguresS1 and S2 (see Supplementary Materials available online athttpsdoiorg10115520175756102) We found that MMP2was positively correlated with serum HMGB1 levels in sub-jects with AE lt 10 years (119903119904 = 02215 119875 = 00037) and AEge 10 years (119903119904 = 02345 119875 = 00075) Similarly MMP9 waspositively correlated with serum HMGB1 levels in healthycontrols (119903119904 = 02580 119875 = 00298) and AE lt 10 years (119903119904 =02721 119875 = 00003) Interestingly both MMP2 and MMP9serum levels were not correlated with HMGB1 in ARDsincluding PP asbestosis and MPM (Table 4) MMP2 andMMP9 were correlated with each other in all groups exceptfor the asbestosis group (Table 4) Although it was confirmedthat MMP2 and MMP9 expression levels were closely linkedwith HMGB1-mediated mechanisms our findings indicated

6 Disease Markers

Table 4 Relationship between HMGB1 and MMPs covariates in study groups

Group Coefficient (HMGB1 MMP2) (HMGB1 MMP9) (MMP2 MMP9)

Healthy 119903119904 01998 02580 0586119875 00949 00298 lt0001

AE lt 10 Y 119903119904 02215 02721 0345119875 00037 00003 lt0001

AE ge 10 Y 119903119904 02345 01456 0269119875 00075 00997 0002

PP 119903119904 00677 00099 0494119875 05483 09303 lt0001

Asbestosis 119903119904 minus00343 minus01347 0018119875 08548 04701 0925

MPM 119903119904 00821 03464 0725119875 07710 02059 0002

AE asbestos-exposed PP pleural plaques Y years 119903119904 Spearmanrsquos rank correlation coefficient

that serum MMP29 levels may not represent helpful sup-plementary indexes for HMGB1 in the clinical diagnosis ofARDs

4 Discussion

In East Asia ARDs associated with occupational and envi-ronmental asbestos exposure and related products con-taminated with asbestos have gathered continuous publicattention This study provided exhaustive data of one proin-flammatory cytokine HMGB1 from Chinese populationswith AE and ARDs Our results indicated that there areassociations between serum HMGB1 levels and asbestos anddifferent ARDs Most ARDs are diagnosed in those whohave experienced occupational or environmental asbestosexposure It is possible that the serum level of HMGB1 isassociated with previous cumulative exposure to asbestoswhich in turnmay be strongly linked to the severity of inflam-matory disorders Increasing clinical evidence indicates thatextracellular HMGB1 contributes to inflammatory disordersand cancer development [5 6 32] As a potential biomarkerour results strongly support the clinical application value ofHMGB1 in screening for severeARDs To our knowledge thisstudy is the first to report serumHMGB1 levels fromChinesepatients with benign ARDs and MPM

As an alarmin HMGB1 signals cell damage in responseto injury and inflammation [5] In agreement with previousstudies in a variety of cell lines and animal model systemsour results support the notion that circulating HMGB1 iscaused by exposure to asbestos fibers HMGB1 production inmesothelioma cells in vitro such as H2052 (epithelioid) andH28 (sarcomatoid) produced higher HMGB1 protein levelsthan that of a normal human mesothelial cell line [22] Simi-larly in both mice and hamsters injected with asbestos fibersHMGB1 was observed in the nuclei cytoplasm and extra-cellular spaces of mesothelial and inflammatory cells aroundasbestos deposits in vivo [8] Serum levels of HMGB1 per-sisted long-term after exposure to crocidolite fibers but notto chrysotile in asbestos-injected mice However prolongedchrysotile exposure may induce sustained serum levels of

HMGB1 [33] Remarkably chrysotile is one type of asbestosthat accounts for approximately 90 of asbestos used com-merciallyworldwide [34] and is also themain rawmaterial forasbestos product manufacturers from Southeast China in ourstudy [23] Pulmonary receptor for advanced glycation end-products (RAGE) a well-characterized cell surface receptorof HMGB1 was down-regulated after asbestos injury in amouse model of pulmonary fibrosis [35] In most healthytissues RAGE is expressed at low to undetectable levelsHowever its expression in the lung is very high undernormal conditions and is depleted in fibrotic disease stateswhich may have the protective effects of accepting HMGB1signaling thus causing chronic inflammation [36]This resultsuggested that substantialHMGB1 is to the extracellular spaceduring asbestos-induced pulmonary fibrosis Combined withthese previous findings our results show that HMGB1 maybe a key inflammatory mediator in ARDs Novel strategiestargeting HMGB1 such as glycyrrhizin [37] gabexate mesi-late [37] ethyl pyruvate [38 39] and PPAR ligands [40] thatinterfere with asbestos-mediated inflammation may preventor delay ARDs onset and relieve the progress of malignantARDs

The possibility that HMGB1 may be a blood biomarkerfor MM or other ADRs is of particular interest HumanMM is an aggressive and highly lethal cancer predominantlycaused by chronic exposure to asbestos and erionite Theprognosis for this cancer is generally poor due to late-stagediagnosis and resistance to current conventional therapies[41] One study from Japan reported that the diagnosticsensitivity of serum HMGB1 for MPM measured on anROC curve was not high (344) but its specificity andpositive predictive value were extremely high (100 and100 resp) [22] Similarly one research group from the USAalso reported that the HMGB1 concentrations in the seraof patients with MM were significantly higher than thosein sera from healthy controls [20 42] Furthermore theydetermined that hyperacetylated HMGB1 a specific subset ofHMGB1 molecules reliably distinguished MM patients fromindividuals who were occupationally exposed to asbestoswith 100 sensitivity and specificity [42] Our results also

Disease Markers 7

support the notion that HMGB1 is a potential biomarker todiscriminate MM patients from asbestos-exposed patientswith benign ARDs (ie PP) or healthy individuals Howeverthere are no previous statistics of total serumHMGB1 levels toperform a comparison between patients with asbestosis andMM To our knowledge this study is the first to report serumHMGB1 levels from the patients with asbestosis Althoughtotal HMGB1 levels are sensitive and specific to discriminateARDs from unexposed individuals the very low AUC of056 observedwhen comparingMPMpatientswith asbestosisindividuals would limit its clinical utility for identifyingdifferent types of ARDs patients among large cohorts of AEor healthy individuals Notably we recently reported a highincidence of peritoneal mesothelioma in relatively youngwomen (mean age of 524 years) from Southeast China [23]Furthermore epidemiologic investigations have shown thatlocal women who worked at asbestos plants also performedhand spinning with their families including children athome Other studies have indicated that improper storageof asbestos products also may lead to household or environ-mental exposure [23 43 44] From our perspective HMGB1may be a suitable blood biomarker to monitor occupationalworkers and their families who have a history of residentialexposure to asbestos but accurately discriminating MM andasbestosis requires further investigation Nevertheless thecombination of serum HMGB1 and radiographic findingsshould be helpful to stratify the risk of MM in AE popula-tions

Ourwork also found that the protein levels of bothMMP2and MMP9 were not correlated with levels of HMGB1 insera of patients with asbestosis or MM It is well known thatMMPs are involved in the process of metastasis formationand that their overexpression correlates with the processesof tumor cell invasion and metastasis in human cancers bydegrading the extracellular matrix [45 46] Among MMPsthe activation ofMMP2 andMMP9 is also associatedwith thepathogenesis of airway inflammation and remodeling [47]Extracellular HMGB1 may activate the RAGE-Ras-MAPKpathway which results in MMP2 and MMP9 expression[27] For example in non-small-cell lung cancer (NSCLC)cells HMGB1 overexpression resulted in increased MMP9expression which was related to a higher metastasis rate [2730] HMGB1may also triggerMMP9upregulation in neuronsand astrocytes in mouse brains after cerebral ischemia [28]In patients with ischemic stroke increased plasma levels ofMMP9 and HMGB1 are associated with a poor functionaloutcome and are significantly correlated with each other[29] In our study of ADRs serum MMP2 and MMP9 levelswere not significantly positively correlated with HMGB1levels therefore the diagnostic value of serum MMP2 andMMP9 was not useful (Table 4) Another study also supportsthis notion [48] Combining these results it is likely thatextracellular HMGB1 may induce intracellular MMP2 andMMP9 expression in local tissue cells but not their secretionsinto the blood Hence the underlying mechanism of secretedMMPs involving their release and circulation is complex inADRs and remains to be elucidated in the future

To summarize serum HMGB1 levels were high in theAE population and particularly high in asbestosis and MM

patients HMGB1 may be a useful clinical biomarker forscreening severe ARDs and it is of potential diagnosticvalue for evaluating high-risk AE cohorts This evidence alsosuggests that HMGB1 is a potential therapeutic target inARDs

5 Conclusions

To our knowledge this is the first clinical study describingserumHMGB1 levels in asbestos-exposed and ARDs popula-tions inChinaOur study demonstrated that the expression ofHMGB1 was significantly elevated in AE and ARDs subjectsFor clinical diagnosis these results indicated that serumHMGB1 is a sensitive and specific biomarker to discriminateasbestosis and MM from healthy or AE individuals Thefindings in this work may also provide new insights into themolecular mechanisms of the progression and prognosis ofADRs andmay lead to new approaches for effective diagnosisand therapy

Competing Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

This study was supported by grants from Zhejiang ProvincialNatural Science Foundation (LY16H240001) the QianjiangTalents Project of Zhejiang Province (QJD1602027) theNational Natural Science Foundation of China (81502794)the Health and Family Planning Commission of ZhejiangProvince (2015KYA058 2016KYB068 2017PY015 and 11-ZC02) and Zhejiang Committee of Science and Technology(2014C03030)

References

[1] K Takahashi P J Landrigan and R Collegium ldquoThe globalhealth dimensions of asbestos and asbestos-related diseasesrdquoAnnals of Global Health vol 82 no 1 pp 209ndash213 2016

[2] M N Courtice S Lin and X Wang ldquoAn updated review onasbestos and related diseases in Chinardquo International Journal ofOccupational and Environmental Health vol 18 no 3 pp 247ndash253 2012

[3] E-K Park K Takahashi Y JiangMMovahed and T KamedaldquoElimination of asbestos use and asbestos-related diseases anunfinished storyrdquo Cancer Science vol 103 no 10 pp 1751ndash17552012

[4] H Naglova and M Bucova ldquoHMGB1 and its physiological andpathological rolesrdquo Bratislavske Lekarske Listy vol 113 no 3 pp163ndash171 2012

[5] R Kang Q Zhang H J Zeh III M T Lotze and D TangldquoHMGB1 in cancer good bad or bothrdquo Clinical CancerResearch vol 19 no 15 pp 4046ndash4057 2013

[6] G P Sims D C Rowe S T Rietdijk R Herbst and A J CoyleldquoHMGB1 and RAGE in inflammation and cancerrdquo AnnualReview of Immunology vol 28 pp 367ndash388 2010

[7] J ANogueira-MachadoCMDOVolpe CAVeloso andMM Chaves ldquoHMGB1 TLR and RAGE a functional tripod that

8 Disease Markers

leads to diabetic inflammationrdquo Expert Opinion on TherapeuticTargets vol 15 no 8 pp 1023ndash1035 2011

[8] H Yang Z Rivera S Jube et al ldquoProgrammednecrosis inducedby asbestos in human mesothelial cells causes high-mobilitygroup box 1 protein release and resultant inflammationrdquo Pro-ceedings of the National Academy of Sciences of the United Statesof America vol 107 no 28 pp 12611ndash12616 2010

[9] M Carbone and H Yang ldquoMolecular pathways target-ing mechanisms of asbestos and erionite carcinogenesis inmesotheliomardquo Clinical Cancer Research vol 18 no 3 pp 598ndash604 2012

[10] Y Albayrak A Albayrak M Celik et al ldquoHigh mobilitygroup box protein-1 (HMGB-1) as a new diagnostic markerin patients with acute appendicitisrdquo Scandinavian Journal ofTrauma Resuscitation and Emergency Medicine vol 19 article27 2011

[11] K Resman Rus L Fajs M Korva and T Avsic-ZupancldquoHMGB1 is a potential biomarker for severe viral hemorrhagicfeversrdquo PLoS Neglected Tropical Diseases vol 10 no 6 ArticleID e0004804 2016

[12] J A Nogueira-Machado and C M de Oliveira Volpe ldquoHMGB-1 as a target for inflammation controllingrdquo Recent Patents onEndocrine Metabolic amp Immune Drug Discovery vol 6 no 3pp 201ndash209 2012

[13] Y Li J Tian X Fu et al ldquoSerum high mobility group boxprotein 1 as a clinical marker for ovarian cancerrdquo Neoplasmavol 61 no 5 pp 579ndash584 2014

[14] J Paek M Lee E J Nam S W Kim and Y T Kim ldquoClinicalimpact of highmobility group box 1 protein in epithelial ovariancancerrdquoArchives of Gynecology andObstetrics vol 293 no 3 pp645ndash650 2016

[15] X Pang Y Zhang HWei et al ldquoExpression and effects of high-mobility group box 1 in cervical cancerrdquo International Journal ofMolecular Sciences vol 15 no 5 pp 8699ndash8712 2014

[16] G-H Shang C-Q Jia H Tian et al ldquoSerum high mobilitygroup box protein 1 as a clinical marker for non-small cell lungcancerrdquo Respiratory Medicine vol 103 no 12 pp 1949ndash19532009

[17] S Sun W Zhang Z Cui et al ldquoHigh mobility group box-1 andits clinical value in breast cancerrdquoOncoTargets andTherapy vol8 pp 413ndash419 2015

[18] D Suren M Yildirim O Demirpence et al ldquoThe role of HighMobility Group Box 1 (HMGB1) in colorectal cancerrdquo MedicalScience Monitor vol 20 pp 530ndash537 2014

[19] M Ueda Y Takahashi Y Shinden et al ldquoPrognostic signifi-cance of High Mobility Group Box 1 (HMGB1) expression inpatients with colorectal cancerrdquoAnticancer Research vol 34 no10 pp 5357ndash5362 2014

[20] S Jube Z S RiveraM E Bianchi et al ldquoCancer cell secretion ofthe DAMP protein HMGB1 supports progression in malignantmesotheliomardquo Cancer Research vol 72 no 13 pp 3290ndash33012012

[21] C Tabata S Kanemura R Tabata et al ldquoSerum HMGB1 asa diagnostic marker for malignant peritoneal mesotheliomardquoJournal of Clinical Gastroenterology vol 47 no 8 pp 684ndash6882013

[22] C Tabata E Shibata R Tabata et al ldquoSerum HMGB1 as aprognostic marker for malignant pleural mesotheliomardquo BMCCancer vol 13 article 205 2013

[23] Z Gao K Hiroshima XWu et al ldquoAsbestos textile productionlinked to malignant peritoneal and pleural mesothelioma in

women analysis of 28 cases in Southeast Chinardquo AmericanJournal of Industrial Medicine vol 58 no 10 pp 1040ndash10492015

[24] D D Smith ldquoDiagnosis and initial management of non-malignant diseases related to asbestosrdquo American Journal ofRespiratory and Critical Care Medicine vol 171 no 6 pp 691ndash715 2005

[25] A N Husain T Colby N Ordonez et al ldquoGuidelines forpathologic diagnosis of malignant mesothelioma 2012 updateof the consensus statement from the International Mesothe-lioma Interest Grouprdquo Archives of Pathology and LaboratoryMedicine vol 137 no 5 pp 647ndash667 2013

[26] M Yu Y Zhang Z Jiang et al ldquoMesothelin (MSLN) methyla-tion and soluble mesothelin-related protein levels in a Chineseasbestos-exposed populationrdquo Environmental Health and Pre-ventive Medicine vol 20 no 5 pp 369ndash378 2015

[27] P-L Liu J-R Tsai J-J Hwang et al ldquoHigh-mobility groupbox 1-mediated matrix metalloproteinase-9 expression in non-small cell lung cancer contributes to tumor cell invasivenessrdquoAmerican Journal of Respiratory Cell andMolecular Biology vol43 no 5 pp 530ndash538 2010

[28] J Qiu J Xu Y Zheng et al ldquoHigh-mobility group box 1promotes metalloproteinase-9 upregulation through toll-likereceptor 4 after cerebral ischemiardquo Stroke vol 41 no 9 pp2077ndash2082 2010

[29] N Sapojnikova T Kartvelishvili N Asatiani et al ldquoCorrela-tion between MMP-9 and extracellular cytokine HMGB1 inprediction of human ischemic stroke outcomerdquo Biochimica etBiophysica Acta vol 1842 no 9 pp 1379ndash1384 2014

[30] J-L Wang D-W Wu Z-Z Cheng W-Z Han S-W Xuand N-N Sun ldquoExpression of High Mobility Group Box-B1(HMGB-1) and Matrix metalloproteinase-9 (MMP-9) in non-small cell lung cancer (NSCLC)rdquoAsian Pacific Journal of CancerPrevention vol 15 no 12 pp 4865ndash4869 2014

[31] H-M Chen S-F Liou J-H Hsu et al ldquoBaicalein inhibitsHMGB1 release and MMP-2-9 expression in lipopolysac-charide-induced cardiac hypertrophyrdquo American Journal ofChinese Medicine vol 42 no 4 pp 785ndash797 2014

[32] D Tang R Kang H J Zeh III andM T Lotze ldquoHigh-mobilitygroup box 1 and cancerrdquoBiochimica et Biophysica Acta vol 1799no 1-2 pp 131ndash140 2010

[33] F Qi G Okimoto S Jube et al ldquoContinuous exposureto chrysotile asbestos can cause transformation of humanmesothelial cells via HMGB1 and TNF-120572 signalingrdquoThe Amer-ican Journal of Pathology vol 183 no 5 pp 1654ndash1666 2013

[34] T A Sporn ldquoMineralogy of asbestosrdquo Recent Results in CancerResearch vol 189 pp 1ndash11 2011

[35] J M Englert L E Hanford N Kaminski et al ldquoA role forthe receptor for advanced glycation end products in idiopathicpulmonary fibrosisrdquoAmerican Journal of Pathology vol 172 no3 pp 583ndash591 2008

[36] J M Englert C R Kliment L Ramsgaard et al ldquoParadoxicalfunction for the receptor for advanced glycation end productsin mouse models of pulmonary fibrosisrdquo International Journalof Clinical and Experimental Pathology vol 4 no 3 pp 241ndash2542011

[37] D Musumeci G N Roviello and D Montesarchio ldquoAnoverview on HMGB1 inhibitors as potential therapeutic agentsin HMGB1-related pathologiesrdquo Pharmacology and Therapeu-tics vol 141 no 3 pp 347ndash357 2014

[38] P Cheng W Dai F Wang et al ldquoEthyl pyruvate inhibitsproliferation and induces apoptosis of hepatocellular carcinoma

Disease Markers 9

via regulation of the HMGB1-RAGE and AKT pathwaysrdquoBiochemical andBiophysical ResearchCommunications vol 443no 4 pp 1162ndash1168 2014

[39] J-H Shin I-D Kim S-W Kim et al ldquoEthyl pyruvate inhibitsHMGB1 phosphorylation and release by chelating calciumrdquoMolecular Medicine vol 20 pp 649ndash657 2014

[40] S Ying X Xiao T Chen and J Lou ldquoPPAR ligands functionas suppressors that target biological actions of HMGB1rdquo PPARResearch vol 2016 Article ID 2612743 10 pages 2016

[41] A Bononi A Napolitano H I Pass H Yang and M CarboneldquoLatest developments in our understanding of the pathogenesisof mesothelioma and the design of targeted therapiesrdquo ExpertReview of Respiratory Medicine vol 9 no 5 pp 633ndash654 2015

[42] A Napolitano D J Antoine L Pellegrini et al ldquoHMGB1 andits hyperacetylated isoform are sensitive and specific serumbiomarkers to detect asbestos exposure and to identifymesothe-lioma patientsrdquo Clinical Cancer Research vol 22 no 12 pp3087ndash3096 2016

[43] S X Cai C H Zhang X Zhang and K Morinaga ldquoEpi-demiology of occupational asbestos-related diseases in ChinardquoIndustrial Health vol 39 no 2 pp 75ndash83 2001

[44] E Yano Z-M Wang X-R Wang et al ldquoMesothelioma in aworker who spun chrysotile asbestos at home during child-hoodrdquo American Journal of Industrial Medicine vol 52 no 4pp 282ndash287 2009

[45] R P Verma and C Hansch ldquoMatrix metalloproteinases(MMPs) chemical-biological functions and (Q)SARsrdquo Bioor-ganic and Medicinal Chemistry vol 15 no 6 pp 2223ndash22682007

[46] J Cathcart A Pulkoski-Gross and J Cao ldquoTargeting matrixmetalloproteinases in cancer bringing new life to old ideasrdquoGenes and Diseases vol 2 no 1 pp 26ndash34 2015

[47] P Sivakumar P Ntolios G Jenkins and G Laurent ldquoInto thematrix targeting fibroblasts in pulmonary fibrosisrdquo CurrentOpinion in Pulmonary Medicine vol 18 no 5 pp 462ndash4692012

[48] M Amati M Tomasetti M Scartozzi et al ldquoProfiling tumor-associated markers for early detection of malignant mesothe-lioma an epidemiologic studyrdquo Cancer Epidemiology Biomark-ers and Prevention vol 17 no 1 pp 163ndash170 2008

Submit your manuscripts athttpswwwhindawicom

Stem CellsInternational

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014


MEDIATORSINFLAMMATION

of


Behavioural Neurology

EndocrinologyInternational Journal of



Disease Markers


BioMed Research International

OncologyJournal of



Oxidative Medicine and Cellular Longevity


PPAR Research

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Immunology ResearchHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

ObesityJournal of



Computational and Mathematical Methods in Medicine

OphthalmologyJournal of


Diabetes ResearchJournal of



Research and TreatmentAIDS


Gastroenterology Research and Practice


Parkinsonrsquos Disease

Evidence-Based Complementary and Alternative Medicine

Volume 2014Hindawi Publishing Corporationhttpwwwhindawicom

2 Disease Markers

cytosol and is released into the cell culture medium [8] Sub-sequently HMGB1 release induces macrophages to secretetumor necrosis factor-120572 (TNF-120572) which protects mesothelialcells from asbestos-induced cell death and triggers a chronicinflammatory response [8] Chronic inflammation is also thehallmark of asbestos deposition in tissue and contributes toits carcinogenesis [9]

As a possible mechanism in asbestos-induced inflam-mation HMGB1 may be secreted into the stroma andsubsequently makes its way into the systemic circulation[9] Recent clinical studies have shown that HMGB1 is apotential diagnostic or prognostic biomarker in a varietyof inflammatory disorders [10ndash12] and cancers [13ndash19] Onegroup from the USA provided the first evidence that HMGB1was highly expressed in both tissues and sera of mesothe-lioma patients [20] In a following report with similar resultsbased on the exposure of Japanese subjects to asbestos theHMGB1 protein levels in the sera from patients with diffusemalignant peritoneal mesothelioma were also reported to besignificantly higher than those of nonmesothelioma subjectswith a history of exposure to asbestos [21] Serum HMGB1was also a prognostic marker for Japanese patients withmalignant pleural mesothelioma (MPM) [22] However theclinical importance ofHMGB1 for other nonmalignantARDsis still unknown Little data is available for serum HMGB1levels from a Chinese asbestos-exposed (AE) population

In the present study we quantified serum HMGB1 levelsin a Chinese AE population and further evaluated the diag-nostic value of HMGB1 in patients with nonmalignant andmalignant ARDs including pleural plaques (PP) asbestosisand malignant mesothelioma (MM) We provide evidencesupporting the hypothesis that HMGB1 could be a candidatebiomarker in clinical diagnosis for MM and asbestosis

2 Materials and Methods

This study was conducted in accordance with the Declarationof Helsinki and was approved by theMedical Ethics Commit-tee of Zhejiang Academy of Medical Sciences

21 Study Population and Data Collection A total of 497subjects were recruited from several small neighboring townswith a long history of asbestos industry during the years1960 to 2012 in Southeast China [23] Chest X-rays weretaken for all of the subjects Nonmalignant asbestos-relateddiseases were diagnosed according to the AmericanThoracicSociety (ATS) criteria [24] MPM was diagnosed by at leasttwo independent pathologists according to the Guidelines forPathologic Diagnosis of Malignant Mesothelioma proposedby the International Mesothelioma Interest Group [25] Astandardized questionnaire focused on occupational historywas taken during a face-to-face interview The contents ofquestionnaires included age gender smoking and drinkinghabits occupation and jobs entailing past asbestos exposureFor those who were occupationally exposed to asbestosexposure duration was measured as the interval of yearsbetween the start and the end of AE jobs Potential expo-sure levels to asbestos were estimated according to theapproach in our previous report [26] Briefly we retrieved

periodic data of total dust concentrations in different work-shopsmanufactories available from 1984 to 2010 in thisarea The median dust concentrations were 07mgm3 in ahousehold textile workshop (small-scale) and 59mgm3 in atextile manufactory (large-scale) Based on the correspond-ing median dust concentration potential exposure levelswere calculated (in mgm3 times years) when applicable

22 Grouping Sets The study subjects were divided into sixgroups (1) a healthy control group that consisted of 71 healthysubjects without any known history of asbestos exposure andwith normal chest X-ray examination results (2) a group of170 subjects occupationally exposed to asbestos for less than10 years but with normal chest X-ray examination results (3)a group of 129 subjects occupationally exposed to asbestosfor over 10 years but with normal chest X-ray examinationresults (4) a group of 81 subjectswith PP found in their chests(5) a group of 31 patients diagnosed with asbestosis and (6) agroup of 15 patients diagnosed with MPM

23 Detection Using Blood Samples Blood samples werecollected using standard blood procurement protocols fromthe enrolled subjects who provided written informed consentfor participation in this approved study Five milliliters ofblood was collected from the antecubital vein After clot-ting these samples were centrifuged at 4000 rpm for 5minThe subsequent separated sera were stored at minus80∘C untildetection assays for HMGB1 and matrix metalloproteinases(MMPs) were performed The concentrations of HMGB1MMP2 andMMP9 in the sera weremeasured using enzyme-linked immunosorbent assays (ELISAs) according to themanufacturerrsquos instructions (Cloud-Clone Corp HoustonTX USA) All assays were performed in duplicate

24 Statistical Analysis Statistical analyses were performedusing GraphPad Prism 5 software (GraphPad Software IncLa Jolla CA USA) or SPSS Statistics 170 (SPSS Inc ChicagoIL USA) Qualitative data are reported as frequencies andproportion distributions The chi-square test was used forcomparison of the proportions between groups and thesignificance level was adjusted using the Bonferroni methodQuantitative data with nonnormal distributions are pre-sented as medians and interquartile ranges A Kruskal-Wallis test followed by Dunnrsquos multiple comparisons test wasconducted to compare the median values between groupsStatistical significance was defined as a two-sided 119875 lt 005Univariate andmultivariate linear regression analyses were fitto explore predictors ofHMGB1The receiver operating curve(ROC) method was conducted for differentiating ARDs andAE groups from the control group Areas under ROC curves(AUCs) were presented with 95 confidence intervals (CIs)An AUC of 10 indicates perfect discrimination whereas anarea with a confidence interval of 05 indicates that the testdiscriminates no better than chance The optimum statisticalcut-off value was selected when Youdenrsquos index (sensitivity +specificity-1) reached the maximum value The relationshipsamong the correlations betweenHMGB1MMP2 andMMP9were identified using Spearmanrsquos rank correlation coefficient(119903119904)

Disease Markers 3


Control(119899 = 71)

AE lt 10 Y(119899 = 170)

AE ge 10 Y(119899 = 129)

PP(119899 = 81)


MPM(119899 = 15)

Age ymedian (IQR) 67 (8) 66 (8) 67 (7) 68 (10) 73 (16) 66 (15)e



NA 177 (319) 708 (839) 502 (770) 998771 119 (833)998771 110 (131)



3 Results



n = 15n = 31n = 81n = 129n = 170n = 71

lowastlowast

lowastlowast

lowastlowast

lowastlowast

lowastlowast

lowastlowast

100

80

60

40

20

0

HM

GB1

(ng

ml)





4 Disease Markers














Disease Markers 5

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0938

100806040200

MPM vs healthy


(a)

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0883

100806040200



(b)

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0812

100806040200

MPM vs A lt 10 Y


(c)

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0804

100806040200

MPM vs A ge 10 Y


(d)





6 Disease Markers



Healthy 119903119904 01998 02580 0586119875 00949 00298 lt0001

AE lt 10 Y 119903119904 02215 02721 0345119875 00037 00003 lt0001

AE ge 10 Y 119903119904 02345 01456 0269119875 00075 00997 0002

PP 119903119904 00677 00099 0494119875 05483 09303 lt0001


MPM 119903119904 00821 03464 0725119875 07710 02059 0002



4 Discussion





Disease Markers 7





5 Conclusions


Competing Interests


Acknowledgments


References








8 Disease Markers


































Disease Markers 9

















of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of
















Disease Markers 3


Control(119899 = 71)

AE lt 10 Y(119899 = 170)

AE ge 10 Y(119899 = 129)

PP(119899 = 81)


MPM(119899 = 15)

Age ymedian (IQR) 67 (8) 66 (8) 67 (7) 68 (10) 73 (16) 66 (15)e



NA 177 (319) 708 (839) 502 (770) 998771 119 (833)998771 110 (131)



3 Results



n = 15n = 31n = 81n = 129n = 170n = 71

lowastlowast

lowastlowast

lowastlowast

lowastlowast

lowastlowast

lowastlowast

100

80

60

40

20

0

HM

GB1

(ng

ml)





4 Disease Markers














Disease Markers 5

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0938

100806040200

MPM vs healthy


(a)

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0883

100806040200



(b)

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0812

100806040200

MPM vs A lt 10 Y


(c)

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0804

100806040200

MPM vs A ge 10 Y


(d)





6 Disease Markers



Healthy 119903119904 01998 02580 0586119875 00949 00298 lt0001

AE lt 10 Y 119903119904 02215 02721 0345119875 00037 00003 lt0001

AE ge 10 Y 119903119904 02345 01456 0269119875 00075 00997 0002

PP 119903119904 00677 00099 0494119875 05483 09303 lt0001


MPM 119903119904 00821 03464 0725119875 07710 02059 0002



4 Discussion





Disease Markers 7





5 Conclusions


Competing Interests


Acknowledgments


References








8 Disease Markers


































Disease Markers 9

















of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of
















4 Disease Markers














Disease Markers 5

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0938

100806040200

MPM vs healthy


(a)

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0883

100806040200



(b)

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0812

100806040200

MPM vs A lt 10 Y


(c)

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0804

100806040200

MPM vs A ge 10 Y


(d)





6 Disease Markers



Healthy 119903119904 01998 02580 0586119875 00949 00298 lt0001

AE lt 10 Y 119903119904 02215 02721 0345119875 00037 00003 lt0001

AE ge 10 Y 119903119904 02345 01456 0269119875 00075 00997 0002

PP 119903119904 00677 00099 0494119875 05483 09303 lt0001


MPM 119903119904 00821 03464 0725119875 07710 02059 0002



4 Discussion





Disease Markers 7





5 Conclusions


Competing Interests


Acknowledgments


References








8 Disease Markers


































Disease Markers 9

















of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of
















Disease Markers 5

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0938

100806040200

MPM vs healthy


(a)

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0883

100806040200



(b)

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0812

100806040200

MPM vs A lt 10 Y


(c)

100

80

60

40

20

0

Sens

itivi

ty (

)

AU = 0804

100806040200

MPM vs A ge 10 Y


(d)





6 Disease Markers



Healthy 119903119904 01998 02580 0586119875 00949 00298 lt0001

AE lt 10 Y 119903119904 02215 02721 0345119875 00037 00003 lt0001

AE ge 10 Y 119903119904 02345 01456 0269119875 00075 00997 0002

PP 119903119904 00677 00099 0494119875 05483 09303 lt0001


MPM 119903119904 00821 03464 0725119875 07710 02059 0002



4 Discussion





Disease Markers 7





5 Conclusions


Competing Interests


Acknowledgments


References








8 Disease Markers


































Disease Markers 9

















of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of
















6 Disease Markers



Healthy 119903119904 01998 02580 0586119875 00949 00298 lt0001

AE lt 10 Y 119903119904 02215 02721 0345119875 00037 00003 lt0001

AE ge 10 Y 119903119904 02345 01456 0269119875 00075 00997 0002

PP 119903119904 00677 00099 0494119875 05483 09303 lt0001


MPM 119903119904 00821 03464 0725119875 07710 02059 0002



4 Discussion





Disease Markers 7





5 Conclusions


Competing Interests


Acknowledgments


References








8 Disease Markers


































Disease Markers 9

















of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of
















Disease Markers 7





5 Conclusions


Competing Interests


Acknowledgments


References








8 Disease Markers


































Disease Markers 9

















of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of
















8 Disease Markers


































Disease Markers 9

















of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of
















Disease Markers 9

















of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of





















of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of
















serum hmgb1 as a potential biomarker for patients with...

Documents