RESEARCH ARTICLE

Serum levels of macrophage migration-inhibitoryfactor (MIF) have diagnostic, predictive and prognosticroles in epithelial ovarian cancer patients

Faruk Tas & Senem Karabulut & Murat Serilmez &

Rumeysa Ciftci & Derya Duranyildiz

Received: 6 November 2013 /Accepted: 18 November 2013# International Society of Oncology and BioMarkers (ISOBM) 2013

Abstract Macrophagemigration-inhibitory factor (MIF) playsan important role in the pathogenesis of multiple malignancies,and its expression strongly also affects outcomes of cancerpatients. The objective of this study was to determine theclinical significance of serum levels ofMIF in epithelial ovariancancer (EOC) patients. A total of 50 patients with a patholog-ically confirmed diagnosis of EOC were enrolled into thisstudy. Serum MIF concentrations were determined using thesolid-phase sandwich ELISA method. Age- and sex-matched30 healthy controls were included in the analysis. Median ageof patients was 56.5 years old, range 22 to 83 years. Majority ofthe patients had an advanced disease (International Federationof Gynecologists and Obstetricians (FIGO) stages III and IV)(90 %). Baseline serum MIF levels were significantly higherthan those in the healthy control group (p =0.005). No knownclinical variables including histology, grade of histology, stageof disease, debulking surgery, and serum CA 125 levels werefound to be correlated with serum MIF levels (p >0.05). Onlythose chemotherapy-unresponsive patients had higher serumMIF levels compared with responsive ones (p =0.02). Patientswith elevated serum MIF concentrations had significantly un-favorable overall survival compared to those with lower levels(p =0.01). However, a serum MIF level was found to play noprognostic role for progression-free survival (p =0.09). In con-clusion, serum levels of MIF have diagnostic, predictive, andprognostic roles in EOC patients.

Keywords Serum .MIF . Ovarian cancer . Prognostic factor

Introduction

Macrophage migration-inhibitory factor (MIF) is a cytokinethat plays an important role in regulating inflammatory andimmunological functions [1–8]. It helps macrophages incarrying out functions including phagocytosis, adherence,spreading, and metabolism. Apart from these functions,MIF is considered to play a role in cell proliferation anddifferentiation. Moreover, it has also a potential to suppressthe tumor suppressor gene p53. MIF expression has beenreported in several malignant tumors. The expression ofMIF corresponds with tumor aggressiveness and metastaticpotential [1–3].

Although MIF expression has been reported in severalmalignant tumors, its role in epithelial ovarian cancer(EOC) remains unclear [1–8]. Increasing evidence hassuggested that MIF contributes to carcinogenesis byinactivating p53 and enhancing angiogenesis. Hagemannet al. showed that a normal ovarian surface epitheliumdoes not express MIF; however, borderline and EOCdoes [3]. They observed significant increased amountsof MIF protein and MIF activity in EOC, suggesting thatMIF may be a key protein in EOC. The production ofMIF through autocrine signals from EOC cells has shownto stimulate the production of cytokines, chemokines,and angiogenic factors which lead to growth of a tumor[1, 3].

The significance of serological levels of MIF in EOCpatients is not known yet [6–8]. Given the conflicting resultsfrom recent epidemiological studies examining the MIF andEOC, we conducted this study to test the diagnostic, predic-tive, and prognostic role of serumMIF concentrations in EOCpatients.

F. Tas (*) : S. Karabulut :M. Serilmez : R. Ciftci :D. DuranyildizInstitute of Oncology, University of Istanbul,Capa, 34390 Istanbul, Turkeye-mail: faruktas2002@yahoo.com

Tumor Biol.DOI 10.1007/s13277-013-1438-z

Materials and methods

Patients

A total of 50 EOC patients with histologically proven diag-nosis treated at the Institute of Oncology, Istanbul University,were enrolled into the study. The staging was established inaccordance with the International Federation of Gynecologistsand Obstetricians (FIGO) classification. Patients with stage IIIbulky disease or stage IV disease were initially treated withneoadjuvant chemotherapy and operated afterwards. Patientswith operable stage III disease who had undergone primarysurgery consisting of total abdominal hysterectomy, bilateralsalpingo-oophorectomy, appendectomy, omentectomy, andpelvic and/or para-aortic lymphadenectomy were treated withadjuvant chemotherapy. All patients received standard pacli-taxel–carboplatin containing chemotherapy regimen.

For comparison of serumMIF levels, age- and sex-matched30 healthy controls were included in the analysis. Informedconsent was obtained from all patients, and the study wasreviewed and approved by a local ethical committee.

Measurement of serum MIF levels

Serum samples were obtained on the first admission beforeany adjuvant and metastatic treatment was given or duringpatients’ follow-up. Blood samples were obtained from pa-tients and healthy controls by venipuncture and clotted atroom temperature. The sera were collected following centri-fugation and frozen immediately at −20 °C until analysis.

Serum MIF (Hangzhou Eastbiopharm Co., Ltd., People’sRepublic of China) levels were determined using the solid-phase sandwich ELISA method. The MIF ELISA used adouble-antibody sandwich enzyme-linked immunosorbent as-say to determine the level of human macrophage MIF in sam-ples. Serum samples and standards were added to the wellswhich were pre-coated with human MIF-1 monoclonal anti-body, and streptavidin–HRP was added to form immune com-plex and allowed to incubate for 1 h. Unbound material waswashed away. Chromogen solution was added and incubatedfor 10 min (protected from light) for the conversion of thecolorless solution to a blue solution, and the intensity of whichwas proportional to the amount of MIF in the sample. As aneffect of the acidic stop solution, the color has become yellow.The colored reaction product was measured using an automatedELISA reader (Rayto, RT-1904C Chemistry Analyzer, AtlantaGA, USA) at 450 nm. The results were expressed as nanogramsper milliliter.

Statistical analysis

Continuous variables were categorized using median valuesas cutoff point. Assessment of relationships and comparisons

between various clinical/laboratory parameters was accom-plished using Mann–Whitney U test. Overall survival (OS)was calculated from the date of the first admission to clinics todisease-related death or date of the last contact with the patientor any family member. Progression-free survival (PFS) wascalculated from the date of admission to the date of the firstradiologic progression with/without an elevated serum tumormarker. Kaplan–Meier method was used for the estimation ofsurvival distribution, and differences in survival were assessedby the log-rank statistics. A p value of <0.05 was considered

Table 1 Characteristicsof the patient and disease Variables No.

No. of patients 50

Age of patients (years)

<55 24

≥55 26

Histology

Serous papillary 21

Endometrioid 5

Mixed 4

Clear cell 2

Mucinous 1

Undifferentiated 17

Histological grade

I 2

II 6

III 13

Stage of disease

I and II 5

III 33

IV 12

Debulking surgery (optimal)

Yes 22

No 28

Serum hemoglobin level (12 g/dL)

Low 23

Normal 27

Serum LDH level (480 U/L)

Normal 31

High 13

CA 125 level (35 IU/mL)

Normal 10

High 40

Response to chemotherapy

Yes 25

No 12

Platinum sensitivity in relapse

Sensitive 8

Resistant 15

Tumor Biol.

significant. Statistical analysis was carried out using SPSS16.0 software.

Results

A total of 50 patients with a pathologically confirmeddiagnosis of EOC were enrolled into this study. Baselinehistopathological characteristics and demographic featuresof patients are listed in Table 1. Median age of patientswas 56.5 years old, range 22 to 83 years. Majority of thepatients had advanced disease (FIGO stages III and IV)(90 %).

Baseline serum MIF levels were significantly higherthan those in the healthy control group (6.8 versus0.4 ng/mL, p =0.005) (Table 2, Fig. 1). Table 3 showsthe correlations between the serum MIF concentrationsand the known clinicopathological variables. The knownclinical variables including histology, grade of histology,stage of disease, debulking surgery, and serum CA 125levels were not found to be correlated with serum MIF con-centrations (p >0.05). Only chemotherapy-unresponsive pa-tients had higher serum MIF levels compared with responsiveones (p =0.02).

The median follow-up time was 12 months (range 1–78 months). Median PFS for all patients was 8±2.8 months(95 % confidence interval (CI)=3–14 months). One-year PFSrate was 27.3 % (95 % CI=8.7–45.9). Median OS for allpatients was 52±7.5 months (95 % CI=37–67 months), and

1-, 2-, and 3-year overall survival rates were 84.9 % (95 %CI=73.5–96.2), 77 % (95 % CI=62.3–91.7), and 55 % (95 %CI=27–83), respectively. Patients who are unresponsive tochemotherapy and with platinum resistance in a relapseddisease had worse outcome in both PFS and OS analyses(Table 4). The patients with elevated serum MIF concen-trations had significantly unfavorable OS than thosewith lower levels (p =0.01) (Table 4, Fig. 2). However,a serum MIF level was found to play no prognostic role forPFS (p =0.09).

Table 2 The values of serumMIF levels in patients with EOCand in healthy controls

Assay Patients (n =50) Controls (n =30) p

Median Range Median Range

Serum MIF level (ng/mL) 6.8 0.2–36.3 0.4 0.07–36.9 0.005

Fig. 1 The values of serum MIF assays in EOC patients and healthycontrols

Table 3 Results (median and range) of comparisons between serumMIFlevels and various clinical/laboratory parameters

Variables MIF (ng/mL)

Age, years (p) 0.13

Young (<55) 7.2 (0.3–27.7)

Older (Q55) 6.0 (0.2–36.3)

Histology (p) 0.42

Serous papillary 6.4 (0.2–36.3)

Others 7.1 (0.3–30.2)

Grade (p) 0.80

I + II 6.9 (0.3–16.5)

III 6.8 (0.2–36.3)

Stage (p) 0.70

Nonmetastatic 6.7 (0.2–36.3)

Metastatic 6.7 (0.3–30.2)

Debulking surgery (p) 0.40

Yes 7.0 (0.2–6.3)

No 6.4 (0.3–30.2)

Hemoglobin level (p) 0.78

Low 6.8 (0.3–27.7)

Normal 6.7 (0.2–36.3)

Serum LDH level (p) 0.14

Normal 7.2 (0.2–36.3)

High 6.1 (0.3–15.6)

Serum CA 125 level (p) 0.42

Normal 7.0 (0.3–36.3)

High 6.7 (0.2–30.2)

Response to chemotherapy (p) 0.02

Yes 5.8 (0.3–8.7)

No 7.2 (0.3–30.2)

Platinum sensitivity in relapse (p) 0.29

Sensitive 6.4 (0.2–30.2)

Resistant 5.6 (0.3–9.9)

Tumor Biol.

Discussion

In literature, only a limited number of studies have investigatedtissue MIF expression in human EOC [2–5]. Hagemann et al.presented evidences that the enhanced production of MIF bymalignant EOC cells support ovarian cancer carcinogenesis [2,3]. They showed that the autocrine secretion of MIF by EOCcells regulates the tumor microenvironment. The interactionwith other cytokines, angiogenic factors, and chemokinesmight act in an autocrine/paracrine manner to promote tumor

growth, progression, and neovascularization of EOC deposits.The mechanisms of MIF action might include direct effects onspread and cell survival of malignant EOC cells via othercytokines and angiogenic factors. Thus, they reported signifi-cant increased amounts of MIF protein and MIF activity inEOC samples. Moreover, the normal ovarian surface epitheli-um did not express MIF instead of borderline and EOC. Thismeans that MIF might be a key protein in EOC tumorigenesis.In other study, they reported a role for MIF in macrophage-induced EOC cell invasiveness [3].

In order to investigate the in vivo expression of MIF inEOC, Krockenberger et al. [5] used the paraffin sections fromEOC (n =11), borderline tumors, and healthy control tissuestained by MIF-specific Mab 289. Average scores in EOCcells were significantly higher than those in borderline tumorsand normal ovaries (p <0.001). However, no significant dif-ferences were found between various EOC histological sub-types. Moreover, MIF concentrations were found to be similarin solid tissue from the primary tumor and in purified ascites-derived tumor cells, suggesting that MIF expression remainsunchanged when cells dissociate from the primary tumor andmigrate towards the peritoneum. Thus, it was the first todemonstrate that MIF expression in EOC increases with thegrade of malignancy. In the other study, Wu et al. [4] foundthat the positive expression of MIF protein was detected in53.5 % of ovarian carcinoma tissues and was positively cor-related to clinical stages of patients (p <0.01). They supposedthat MIF might play an important role in the pathogenesis andprogression of EOC. Thus, it might be used as a potentialtherapeutic target in EOC patients.

Similar to the MIF expression trials studied in tissue, only alimited number of studies have investigated serum MIF con-centrations in human EOC in literature [6–8]. Evaluating the

Table 4 Univariate analyses of progression-free and overall survivals

Variables Survival (month)

Progression-freesurvival

Overallsurvival

Median (±SE) Median (±SE)

Age, years (p) 0.60 0.12

Young (<55) 29.3 (9.9) 61.2 (8.9)

Older (Q55) 12.6 (2.6) 24.1 (2.3)

Histology (p) 0.85 0.74

Serous papillary 10.4 (2.2) 26.8 (2.6)

Others 27.3 (7.1) 52.6 (9.5)

Grade (p) 0.78 0.17

I + II 10.0 (0.0) NR

III 9.4 (3.1) NR

Stage (p) 0.76 0.34

Nonmetastatic 25.0 (7.0) 55.4 (8.3)

Metastatic 14.5 (3.6) 24.2 (4.3)

Debulking surgery (p) 0.47 0.09

Yes 26.1 (6.3) 45.4 (9.0)

No 9.8 (3.7) 30.3 (1.7)

Hemoglobin level (p) 0.80 0.21

Low 14.1 (3.2) 28.2 (1.8)

Normal 23.8 (7.8) 52.3 (8.6)

Serum LDH level (p) 0.11 0.01

Normal 34.5 (7.0) 62.9 (7.6)

High 9.6 (3.8) 21.7 (4.1)

Serum CA 125 level (p) 0.06 0.11

Normal 52.0 (0.0) NR

High 8.3 (1.4) NR

Response to chemotherapy (p) <0.001 <0.001

Yes 34.2 (7.0) 60.2 (7.9)

No 3.3 (0.5) 14.5 (4.5)

Platinum-sensitivity in relapse (p) 0.001 0.003

Sensitive 46.2 (5.3) 66.0 (10.4)

Resistant 6.9 (1.5) 15.1 (3.2)

Serum MIF level (p) 0.09 0.01

High (>median) 4.0 (1.4) 22.8 (2.6)

Low (<median) 11.0 (2.8) 72.8 (4.9)

NR not reached

Fig. 2 Overall survival curves in EOC patients according to MIF levels(p =0.01)

Tumor Biol.

hypothesis that EOC patients have significantly higher levelsof serum MIF, Agarwal et al. studied MIF levels by ELISA in54 EOC patients [6]. In this trial, they found that cancer cellssecrete significant amounts of MIF unlike normal cells. Thisresult correlated that serum MIF concentrations are signifi-cantly elevated in EOC patients, in vivo. Moreover, serumconcentrations ofMIF assay in EOC patients were significant-ly higher than those measured in healthy controls (p <0.0013).When analyzing the correlation between MIF levels and dis-ease stage, it showed that MIF levels were not significantlydifferentiated between patients with early stage disease (stagesI and II) compared to patients with advanced stage disease(stages III and IV).

Besides, Krockenberger et al. reported that the levels of theMIF protein which is known to be secreted in the serum ofEOC patients seems to correlate with common prognosticparameters such as tumor stage or platinum sensitivity [7].They believed that MIF may play a suppressive role in thehost antitumor immune response, which might have a nega-tive impact on the course of the disease. The fact that MIFlevels in the serum of patients at primary diagnosis correlatewith platinum sensibility supporting the hypothesis that serumMIF concentrations should be evaluated as a parameterreflecting tumor sensibility towards chemotherapy in earlystages of the disease. In another trial, to investigate biomarkersand clinical parameters to distinguish ovarian cancers frombenign ovarian tumors, He et al. also studied MIF in additionto other markers in 37malignant EOC patients [8]. In this pilotstudy, a serum MIF concentration was found to be allowingdifferentiation between EOC and non-cancer cases.

In this trial, we studied serum MIF concentrations in 50patients with a pathologically confirmed diagnosis of EOCand age- and sex-matched 20 healthy controls determined bythe solid-phase sandwich ELISA methods. The baseline serumMIF levels were significantly higher than those in the healthycontrol group (p =0.005). The known clinical variables such ashistology, grade of histology, stage of disease, debulking sur-gery, and serum CA 125 levels were not found to be correlatedwith serum MIF concentrations. Only chemotherapy-unresponsive patients had higher serum MIF levels comparedwith responsive ones (p =0.02). The patients with elevatedserum MIF concentrations had significantly unfavorable OSthan those with lower levels (p =0.01), whereas a serum MIFlevel was found to play no prognostic role for PFS (p =0.09).

In conclusion, although little is known, evidence to datesuggests that the MIF may be involved in the etiology andprogression of EOC. However, there are much conflictingevidences in the literature regarding patterns of expression ofthese gene products; therefore, the precise functional rele-vance of these alterations is not yet well understood. Similarcomments were true for patterns of quantification of thecirculating serum MIF concentrations. The small sample sizeof our study could be considered as a significant limitation andmight have influenced these results. However, our study con-tributes to the literature. A standardized method remains to beestablished and validated in larger series of patients in pro-spective studies to determine the potential clinical significanceof these assays in EOC patients.

Conflicts of interest None

References

1. Babu SN, Chetal G, Kumar S. Macrophage migration inhibitoryfactor: a potential marker of cancer diagnosis and therapy. Asian JCancer Prev. 2012;13:1737–44.

2. Hagemann T, Robinson SC, Thompson RG, Charles K, Kulbe H,Balkwill FR. Ovarian cancer cell-derived migration inhibitory factorenhances tumor growth, progression, and angiogenesis. Mol CancerTher. 2007;6:1993–2002.

3. Hagemann T,Wilson J, Kulbe H, Li NF, Leinster DA, Charles K, et al.Macrophages induce invasiveness of epithelial cancer cells via NF-κBand JNK. J Immunol. 2005;175:1197–205.

4. Wu HM, Zhu SL, He LJ, Liu YH, Xie D. Clinical significance ofmacrophage migration inhibitory factor in invasion of ovarian cancer.Ai Zheng. 2009;28:1054–60 (article in Chinese).

5. Krockenberger M, Dombrowski Y, Weidler C, Ossadnik M, Hönig A,Hausler S, et al. Macrophage migration inhibitory factor (MIF) con-tributes to the immune escape of ovarian cancer by downregulatingNKG2D. J Immunol. 2008;180:7338–48.

6. Agarwal R, Whang DH, Alvero AB, Visintin I, Lai Y, Segal EA, et al.Macrophage migration inhibitory factor expression in ovarian cancer.Am J Obstet Gynecol. 2007;196:348.e1–348e5.

7. KrockenbergerM, Kranke P, Hausler S, Engel JB, Horn E, NürnbergerK, et al. Macrophage migration-inhibitory factor levels in serum ofpatients with ovarian cancer correlates with poor prognosis.Anticancer Res. 2012;32:5233–8.

8. He G, Holcroft CA, Beauchamp MC, Yasmeen A, Ferenczy A,Kendall-Dupont J, et al. Combination of serum biomarkers to differ-entiate malignant from benign ovarian tumours. J Obstet GynaecolCan. 2012;34:567–74.

Tumor Biol.