Circulating Haptoglobin Is an Independent Prognostic Factorin the Sera of Patients with Epithelial Ovarian Cancer*

Changqing Zhao*, Loganath Annamalai*, Changfa Guo y, Narasimhan Kothandaraman z,Stephen Chee Liang Koh*, Huoming Zhang*, Arijit Biswas* and Mahesh Choolani*

Departments of *Obstetrics and Gynecology, and ySurgery, Yong Loo Lin School of Medicine, National Universityof Singapore, Singapore, Singapore; zDepartment of Obstetrics and Gynecology, National University Hospital,Singapore, Singapore

Abstract

OBJECTIVE: This study was conducted to evaluate the

prognosticsignificanceofhaptoglobin levels in theover-

all survival of patients presenting with various stages of

epithelial ovarian cancer. MATERIALS AND METHODS:

We employed an in-house sandwich enzyme-linked

immunosorbent assay method to determine the con-

centrations of preoperative haptoglobin and C-reactive

protein (CRP) in sera samples obtained from 66 malig-

nant tumors, 60 benign tumors, and 10 normal healthy

women. RESULTS: Levels of serum haptoglobin signifi-

cantly correlated with tumor type (P < .001) and Interna-

tional Federation of Gynecology and Obstetrics stage

(P < .05). A significant correlation was observed be-

tween clinical stage and patient survival (r = 5.99, P =

.026). Our data also indicated that elevated serum

haptoglobin levels were associated with poor outcome

for overall survival using both univariate and multi-

variate analyses (P = .048 and P = .036 respectively).

Using Pearson’s correlation, we have noted that serum

CRP concentrations significantly correlated with hapto-

globin levels (r2 = 0.22, P < .001). Immunohistochemical

findings and Western blot analyses were compatible

with sera levels of haptoglobin in which a higher in-

tensity of staining occurred in late-stage epithelial

ovarian cancers. CONCLUSION: This study provides

evidence that preoperative serum levels of haptoglobin

could serve as an independent prognostic factor in

patients presenting with epithelial ovarian cancer.

Neoplasia (2007) 9, 1–7

Keywords: Haptoglobin, epithelial ovarian cancer, sandwich ELISAmethod, CRP levels, prognostic biomarker.

Introduction

Epithelial ovarian cancer is the most lethal of female genital

tract cancers. Owing to a paucity of symptoms and their

insidious onset, most patients presented with advanced

stages of the disease, with the 5-year survival rate being a

mere 30% because of failure in early diagnosis. The prog-

nosis of ovarian cancer is largely dependent on stage and

histologic grade on presentation. The 5-year survival rate for

stage I epithelial ovarian cancer was reported to be approxi-

mately 90% [1]. Previous studies have also indicated that

patients with well-differentiated tumors based on histologic

grade tend to have a survival rate better than those with poorly

differentiated tumors [2]. The current management of patients

with ovarian cancer involves optimal cytoreductive surgery

followed by chemotherapy. Despite an initial favorable reac-

tion to therapy, most patients, however, appear to have a re-

lapse within 5 years after the management of tumors. Although

CA-125 is, by far, themost useful biomarker for ovarian cancer, it

is only limited to the evaluation of therapeutic effectiveness

in patients diagnosed with ovarian cancer and in the detection

of recurrent disease during posttreatment surveillance [3].

Therefore, it is imperative to identify alternative prognostic

markers for improvement in patient survival, to formulate ade-

quate treatment, and to monitor this ominous disease.

The basic haptoglobin molecule is a tetrameric protein with

two a/b dimers. b-Chains are identical in all haptoglobin types,

and polymorphisms of haptoglobin arise from differences in a-

chains [4]. In a manner similar to other acute-phase proteins,

haptoglobin also originates mainly from the liver, and elevation

of this peptide could be observed in infections, inflammations,

and various malignant diseases, including lung and bladder

cancers [5,6], leukemia [7], breast cancer [8], malignant lym-

phoma [9], urogenital tumors [10], and esophageal squamous

cell carcinoma [11]. Elevation of haptoglobin in the sera and

ascitic fluid of ovarian cancer has been reported in previous

studies using enzyme-linked immunosorbent assay (ELISA)

[12,13]. In addition, recent studies using proteomics-based

methods have shown that haptoglobin levels were elevated in

Address all correspondence to: Dr. Mahesh Choolani, Department of Obstetrics and

Gynecology, Yong Loo Lin School of Medicine, National University Hospital, 5 Lower Kent

Ridge Road, Singapore 119074, Singapore. E-mail: obgmac@nus.edu.sg

*This article refers to supplementary material, which is designated by ‘‘W’’ (i.e., Figure W1)

and is available online at www.bcdecker.com.

Received 28 September 2006; Revised 16 November 2006; Accepted 16 November 2006.

Copyright D 2007 Neoplasia Press, Inc. All rights reserved 1522-8002/07/$25.00

DOI 10.1593/neo.06619

Neoplasia . Vol. 9, No. 1, January 2007, pp. 1 – 7 1

www.neoplasia.com

BRIEF ARTICLE

the sera of patients with ovarian cancer [14,15]. Using con-

ventional two-dimensional gel electrophoresis, it was noted

that a decreased intensity of staining for this protein was

observed in samples from patients undergoing chemo-

therapy [16]. This change also correlated well with the pro-

file of CA-125 levels before and after chemotherapy [16]. The

evaluation of haptoglobin on survival rates and the out-

come of patients with ovarian cancer have not been exam-

ined as yet. In this study, we employed the in-house ELISA

method to measure serum haptoglobin levels in the sera

of patients presenting with epithelial ovarian cancer. Our

data indicate that haptoglobin levels could provide a use-

ful measure in the prognostic evaluation of ovarian cancer.

Materials and Methods

Patient and Clinical Samples

Women aged 20 to 72 years with malignant (n = 66) and

benign (n = 60) epithelial ovarian tumors were recruited for

open surgical or laparoscopic treatment at the Department

of Obstetrics and Gynecology, National University Hospital

(Singapore, Singapore), from 1998 to 2006. The histologic

type of ovarian cancer was classified as defined by Inter-

national Federation of Gynecology and Obstetrics (FIGO)

staging [17]. Blood and tissue samples from cysts of ovarian

carcinomas and benign ovarian cysts were collected during

surgery without intraoperative spillage. Thirteen patients di-

agnosed with ovarian cancer were treated with combination

therapy consisting of carboplatin (AUC 5)/Taxol (175 mg/m2

body weight) following surgery. Blood were drawn before

and after six cycles of chemotherapy. Venous blood samples

were collected in 8-ml Vacuette serum tubes containing clot-

activating factor (Greiner Bio-One, Kremsmuenster, Aus-

tria). All samples were transported on ice to the laboratory,

centrifuged immediately at 4jC for 10 minutes at 1500g, and

stored at �80jC until analysis, according to the method

described previously [18]. Tissue samples were snap-frozen

in liquid nitrogen and stored in a freezer at�80jC. Ten normal

healthy controls without any evidence of cancer or meta-

bolic disease were also included in this study. The sample

collection for this study was approved by the Institution

Review Board of the National University Hospital, and in-

formed consent was obtained from each patient. All patients

will be observed every 3 months for the first 2 years, then for

6 months for up to 5 years, and annually thereafter. For

Western blot and immunohistochemistry analyses, tumor

tissue sections that comprised > 60% of cancer cells in the

field were carefully selected by a histopathologist.

Quantitative Analysis by Sandwich ELISA Method

Total haptoglobin levels were quantified using the sand-

wich ELISA technique with two polyclonal antibodies against

this protein. Goat anti-human haptoglobin antibody (Bio-

design, Saco, ME) in an appropriate dilution (1:200) with

0.5% bovine serum albumin (BSA) in PBST buffer [0.05%

Tween-20 in phosphate-buffered saline (PBS) buffer] was

coated onto PVC plates (NUNC, Roskilde, Denmark) over-

night at 4jC. After blocking with 200 ml of 3% nonfat dry milk

in PBST for 2 hours at room temperature, individual serum

samples were mixed with 0.5% BSA in PBST buffer (1:100).

A diluted mixture of 200 ml was added and incubated for

2 hours at room temperature. Purified human haptoglobin

(Sigma-Aldrich, St. Louis, MO) was used as standard at

concentrations ranging between 10 ng/ml and 10 mg/ml.

The plate was then incubated with rabbit anti-human hapto-

globin antibody (Dakocytomation, Glostrup, Denmark) di-

luted in 0.5% BSA in PBST (1:400) at room temperature for

2 hours, followed by the addition of biotin-labeled anti-rabbit

secondary antibody (1:400; Vector Laboratories, Burlingame,

CA) and alkaline phosphatase streptavidin (1:400; Vector

Laboratories). p-Nitrophenylphosphate (Vector Laboratories)

was applied at 5 mM in 100 mM sodium bicarbonate solution

at 37jC for about 1 hour. To determine the concentration of

haptoglobin, the plate was read at 405 nm using an ELISA

reader (Tecan, Salzburg, Austria). Between each stage of

the assay, the plate was washed four times by immersion in

PBST buffer followed by aspiration.

Serum concentrations of human C-reactive protein (CRP)

were determined by a specific ELISA kit (Chemicon Inter-

national, Temecula, CA) according to the manufacturer’s

instructions.

Western Blot Analysis

Ten samples of late-stage ovarian cancer and benign

ovarian tumors were selected for further Western blot and

immunohistochemical analyses. All tissues used were lysed

using Tris buffer (pH 8.0; 50 mM Tris, 150 mM NaCl, and

5 mM EDTA with protease inhibitor cocktail; Roche Diagnos-

tics, Mannheim, Germany). To performWestern blot analysis

for haptoglobin, 10 mg of protein from each sample was

separated by 12% sodium dodecyl sulfate–polyacrylamide

gel electrophoresis. The bands were then electrically

transferred onto 0.45-mm nitrocellulose membranes using

Semidry Transblot Cell (Bio-Rad Laboratories, Hercules,

CA). Membranes were soaked in 5% nonfat dry milk in TBST

(20 mM Tris base, 500 mM NaCl, and 0.05% Tween 20,

pH 7.5) for 1 hour at room temperature and incubated

overnight at 4jC with rabbit anti-human haptoglobin poly-

clonal antibody (1:5000; Dakocytomation). They were

washed thrice with TBST and incubated with horseradish

peroxidase–labeled anti-rabbit secondary antibody (Pierce

Biotechnology, Rockford, IL) diluted in 5% nonfat dry milk

in TBST (1:1000) for 1 hour at room temperature. The mem-

branes were then washed and developed with chemilumi-

nescent substrates (Pierce Biotechnology).

Semiquantitative Reverse Transcription–Polymerase

Chain Reaction (RT-PCR)

In addition, semiquantitative RT-PCR was employed to

confirm mRNA expression for haptoglobin in three ovarian

epithelial cancer cell lines TOV-112D, Caov-3, and SKOV-3

obtained from the American Type Culture Collection (Manas-

sas, VA). These cells were cultured at 37jC in 5% CO2 in

RPMI 1640medium (Invitrogen Corp., Carlsbad, CA) supple-

mented with 100 U/ml penicillin, 100 mg/ml streptomycin

2 Haptoglobin: Prognostic Factor in Ovarian Cancer Zhao et al.

Neoplasia . Vol. 9, No. 1, 2007

(Sigma-Aldrich), and 10% fetal calf serum (Hyclone, Logan,

UT). The cells were grown to 90% confluency and harvested

by treatment with 0.25% trypsin and 0.02% EDTA (GIBCO

Invitrogen, Carlsbad, CA).

Total RNA from cultured ovarian cancer cells (5� 106) was

extracted using an RNeasy Mini Kit (Qiagen GmbH, Hilden,

Germany). Specific mRNA for haptoglobin was analyzed by

RT-PCR, which was carried out using a one-step RT-PCR Kit

(Qiagen GmbH) in accordance with the manufacturer’s

instructions. Optimal PCR conditions for haptoglobin were

determined in preliminary experiments so that the number of

cycles selected was within the linear range of amplification

[19]. The following specific primers were used for the am-

plification of a 197-bp region of human haptoglobin cDNA

fragment: 5V-CTG TGC TGG CAT GTC TAAG-3V (5V up-

stream) and 5V-CAG CTATGG TCT TCT GAAC-3V (3V down-

stream) [20]. Amplification consisted of 40 cycles of 94jCfor 45 seconds, 55jC for 45 seconds, and 72jC for 60 sec-

onds. Finally, after an additional extension step at 72jC for

10 minutes, PCR products were analyzed by electrophoresis

on 2% agarose gels stained with ethidium bromide and

visualized under ultraviolet radiation. GAPDH was used as

positive control.

Immunohistochemistry

Rabbit anti-human haptoglobin polyclonal antibody

(1:5000; Dakocytomation) was used in standard indirect

immunoperoxidase procedures, as reported previously [21].

Statistical Analysis

Statistical significance was evaluated using ANOVA for

comparison of the means of haptoglobin between normal

healthy controls, benign ovarian tumors, and malignant ovar-

ian tumors. Pearson’s chi-square test was used to compare

demographic and clinical data. Overall cancer-specific sur-

vival was defined from the date of blood sampling to the

date of death due to cancer. Kaplan-Meier method was used

to estimate the probability of survival, and differences in sur-

vival rates were assessed by log-rank test. Multivariate

analysis of prognostic factors was performed using the Cox

regression model. All statistical analyses were performed

using SPSS 13.0 (SPSS, Inc., Chicago, IL). Results were

considered statistically significant at P < .05.

Results and Discussion

Tetrameric protein haptoglobin has three phenotypes

(Hp 1-1, Hp 2-1, and Hp 2-2) [22]. All phenotypes are com-

posed of different combinations of b, a1, and a2 subunits.

Results from Western blot analysis showed that the two

antibodies used for the in-house sandwich ELISA method

possessed the capacity of binding both a and b subunits

(Figure W1). Based on this finding, the ELISA method used

in this study measured the total serum haptoglobin concen-

trations. Three samples of different concentrations were

tested eight times on the same day (intra-assay) and were

repeated on five consecutive days (interassay). The coeffi-

cients of variance were 3.8% and 9.2% for intra-assay and

interassay for haptoglobin, respectively. The sensitivity of

this test is 1 mg/ml. We observed that serum haptoglobin con-

centrations were significantly higher in late-stage ovarian

cancer patients (4.6 ± 0.6 mg/ml) than in benign ovarian

tumors and healthy controls (3.12 ± 0.5 and 1.5 ± 0.5 mg/ml,

respectively; P < .05 and P < .01). Moreover, a significant

difference was observed in haptoglobin levels between

early-stage cancer (3.6 ± 0.5 mg/ml) and healthy controls

(P < .01). This observation appears to be in agreement with

previous reports [23,24].

Because CRP is a serological marker of inflammation that

can be used to investigate the association between inflam-

mation and risk of cancer, the levels of this acute-phase

reactant were measured in our patients. CRP concentrations

were observed to be significantly elevated in both early-

stage and late-stage ovarian cancers (13.9 ± 3.0 and 12.7 ±

2.7 mg/ml, respectively) when compared with normal healthy

women and benign tumors (1.28 ± 0.4 and 2.67 ± 0.9 mg/ml;

P < .01). The distribution of haptoglobin and CRP levels in

benign, early, and late epithelial ovarian cancers, as well as in

normal women, is presented in a boxplot, which is bounded

by 75th and 25th percentiles (Figure 1). A significant ele-

vation in serum haptoglobin and CRP levels was observed

in the sera of early-stage and late-stage epithelial ovarian

cancer when compared with those from normal healthy

women and benign tumors, as illustrated in Figure 1.

The association of haptoglobin levels with clinicopatho-

logical variables, including age, FIGO stage, tumor type,

grade, and pathology, is given in Table 1. In this study, we

have used a cutoff value of 4 mg/ml for haptoglobin concen-

tration because this value corresponds to better diagnostic

accuracy in our series of patients. Hence, using the cutoff

value of 4 mg/ml, we were able to minimize false-positive and

false-negative rates in differentiating normal and benign

cases from malignant ovarian tumors. As indicated in the

receiver operating characteristic (ROC) curve (Figure W2),

by adding a series of sensitivity and specificity at different

cutoff coordinate points, we found that at the 4.0-mg/ml

point, the largest value of the sum of sensitivity and specificity

could be achieved. Accordingly, following this criterion, the

ovarian cancer patients were divided into two groups based

on haptoglobin concentrations. Tumor staging/grading ac-

cording to FIGO classification is known to be the most

important prognostic factor for ovarian cancer. In this study,

we noted a significant association between the serum hap-

toglobin level and FIGO stage on diagnosis (P < .05; Table 1).

However, no significant association was observed between

the serum haptoglobin level and age, histology, or tumor

grade on diagnosis (P > .05).

In addition, tumor haptoglobin expression levels were

also examined for overall survival rates using Kaplan-Meier

analysis with log-rank statistics for determining significance.

The mean follow-up time for the entire cohort of 66 patients

was 37.7 months (range, 4–100 months). Using the 4-mg/ml

haptoglobin concentration as a cutoff point, as mentioned

earlier, the median and mean survival times of patients

with enhanced serum haptoglobin levels were observed

to be significantly less than those with low serum levels for

Haptoglobin: Prognostic Factor in Ovarian Cancer Zhao et al. 3

Neoplasia . Vol. 9, No. 1, 2007

this acute-phase reactant protein (38 vs 79, and 46.7 vs

84.7 months; P = .048). Kaplan-Meier survival curves gener-

ated for high versus low concentrations of haptoglobin are

illustrated in Figure 2. It is noteworthy that significantly

elevated haptoglobin concentrations were associated with

poor outcome for survival in women with epithelial ovarian

cancer (Figure 2; P = .048).

Using the Cox proportional hazards model multivariate

survival analysis of 66 patients, we observed enhanced

haptoglobin levels [relative risk = 5.054; 95% confidence

interval (95% CI) = 1.11–22.98] and tumor stage on diag-

nosis (relative risk = 5.99; 95% CI = 1.24–29.0), which ap-

peared to be significant predictor variables of overall survival

in these patients. In contrast, however, age, tumor type, his-

tology, and grade on diagnosis were not significantly corre-

lated with survival rates, as illustrated in Table 2.

Western blot analysis and immunohistochemistry tech-

niques were performed to demonstrate tumor tissue expres-

sion of haptoglobin. Results of Western blot analysis indicated

that the intensity of bands corresponding to haptoglobin in

tissue extracts was higher in advanced ovarian cancer than

in benign ovarian tumors (Figure 3). Data from RT-PCR have

convincingly demonstrated that there occurs a strong ex-

pression of haptoglobin mRNA in ovarian epithelial cancer

cells in culture (Figure 4). Moreover, immunohistochemical

analysis also indicated intense staining of haptoglobin in the

cytoplasm of ovarian cancer cells obtained from late-stage

ovarian cancers when compared with those from benign

ovarian tumors (Figure 5). This elevation in haptoglobin there-

fore provides strong evidence that malignant ovarian tumor

tissues possess the capacity to synthesize this protein.

Table 1. Association of Serum Haptoglobin Levels with Clinicopathological

Variables.

Variable n Haptoglobin Concentration P *

Low (<4 mg/ml) High (>4 mg/ml)

Age (years)

< 52 31 13 18 .691

z 52 35 13 22

Tumor type

Benign 60 41 19 < .001

Malignant 66 26 40

FIGO stage

Early (I/II) 19 11 8 < .05

Late (III/IV) 47 15 32

Histology

Serous 27 11 16 .254

Mucinous 20 8 12

Endometrioid 11 5 6

Clear cell 6 0 6

No record 2

Differentiation

Well 24 8 16 .918

Moderate 14 5 9

Poor 23 9 14

No record 5

*Pearson’s chi-square test was used to compare demographic and clinical data.

Figure 2. Overall survival of 66 patients with epithelial ovarian cancer in

relation to haptoglobin level; a cutoff value of 4 mg/ml was used to divide the

patients into two groups. Patients with higher haptoglobin concentrations had

a significantly worse survival probability.

Figure 1. Serum haptoglobin and CRP concentrations obtained from 66 pa-

tients with epithelial ovarian cancer, 60 benign ovarian tumors, and 10 healthy

controls. Serum levels of haptoglobin and CRP were significantly higher in pa-

tients with early-stage and/or late-stage ovarian cancer than in control subjects.

4 Haptoglobin: Prognostic Factor in Ovarian Cancer Zhao et al.

Neoplasia . Vol. 9, No. 1, 2007

Coupled to our data on survival rates, a prominent role for

haptoglobin in predicting the survival outcome of patients with

ovarian cancer is indicated (Figure 2).

It is envisaged that inflammation could serve as a potential

contributory factor in the development of a variety of cancers.

CRP is a serological marker of inflammation that could be

used to investigate the association between inflammation

and risk of cancer [25]. It is noteworthy that, in a manner

similar to that of haptoglobin, serum CRP levels were also

significantly elevated in early-stage and late-stage ovarian

cancers. Moreover, a significant correlation existed between

CRP and haptoglobin levels (r2 = 0.22, P < .001; Figure W3),

indicating that inflammation occurs in the early and late

stages of ovarian cancer development. Based on this ob-

servation, it seems plausible to suggest that CRP could serve

as an adjunct to haptoglobin in the evaluation of ovarian

cancers. Because haptoglobin is among the most abundant

glycoproteins secreted by the liver and because circulating

hepatic haptoglobin accounts for the major elevation of

this protein in the serum of ovarian cancer patients [16], it is

reasonable to suggest that enhanced hepatic synthesis of

this protein occurs due to an acute-phase response in these

ovarian cancer patients.

It has been previously reported that abnormally elevated

glycosylated haptoglobin was detected in the blood of women

presenting with ovarian and breast cancers [26]. The oligo-

saccharide involved in this binding was identified as the

fucose-specific lectin Lotus tetragonolobus, which was ele-

vated seven-fold in the sera of patients with ovarian tumors

[27]. The serum concentration of fucosylated haptoglobin

closely correlated with tumor burden and levels of a-3-fucosyl

transferase—an enzyme that adds fucose toN-acetylglucos-

amine on glycoproteins [28,29]. Similar changes in oligo-

saccharide structures have been reported in other types of

malignant transformations, including pancreatic cancers

[30]. The origin of fucosylated haptoglobin is, however, un-

known at present, but it could be due to either aberrant

glycosylation by malignant ovarian cancer cells or a factor

produced by cancer cells that could enhance its liver synthe-

sis in afflicted women. Disturbances in fucose metabolism

with respect to haptoglobin binding may be implicated in the

pathophysiology of ovarian cancer.

Haptoglobin levels in the sera of 13 patients who were

subjected to chemotherapy were also measured to deter-

mine whether chemotherapeutic treatment had any effect

on the expression of this acute-phase reactant protein. Its

concentration was significantly reduced after six cycles of

chemotherapeutic treatment (2.2 ± 0.8 mg/ml) when com-

pared with those before treatment (5.2 ± 0.7 mg/ml; P <

.001). This result is in accordance with a previous study

[16] indicating a role for haptoglobin in monitoring patients

undergoing chemotherapy.

Recent investigations revealed that haptoglobin is a nat-

ural inhibitor of collagen degradation and is locally expressed

by fibroblasts in arterial walls [31]. This glycoprotein there-

fore plays an important role in cell migration and arterial

restructuring. Moreover, collagen turnover is an important

feature in many physiological processes such as growth

Table 2. Multivariate Survival Analysis By Cox Regression.

Variables Hazard Ratio (95% CI) P

Age 1.032 (0.357–2.981) .954

FIGO stage 5.99 (1.237–29.005) .026

Histology 0.839 (0.522–1.346) .466

Grade 1.162 (0.457–3.035) .896

Haptoglobin 5.054 (1.111–22.983) .036

Figure 3. Western blot analysis from ovarian tumor tissue samples. Ten micrograms of samples from benign (A) and late-stage ovarian cancer (B) was selected.

Expression of haptoglobin in cancer was significantly higher than that in benign ovarian tumor. Actin was used as loading control.

Figure 4. Analysis of ovarian cancer cell lines for haptoglobin mRNA ex-

pression by RT-PCR. A 197-bp fragment was amplified by RT-PCR from

samples of total RNA extracted from the cultured ovarian cancer cell lines TOV-

112D, Caov-3, and SKOV-3. RT control was performed without Sensiscript

reverse transcriptase, and PCR control was conducted without RNA template.

Haptoglobin: Prognostic Factor in Ovarian Cancer Zhao et al. 5

Neoplasia . Vol. 9, No. 1, 2007

and wound healing, and enhanced collagen degradation is

causally related to severe tissue destruction andmalfunction,

which are often encountered in pathological processes such

as arthritis and cancer [32]. The importance of haptoglobin

in extracellular matrix degradation and cell migration sug-

gests a role for this polypeptide in cancer and also provides

evidence supporting haptoglobin as a molecular target for

therapy in epithelial ovarian cancer.

Our retrospective analysis therefore provides evidence

that circulating haptoglobin levels could be a useful indepen-

dent prognostic indicator in patients with epithelial ovarian

cancer. Overall, our data suggest that lower haptoglobin

levels in preoperative sera are associated with better survival

outcome. More studies are, however, warranted to confirm

these findings.

References[1] Young RC, Walton LA, Ellenberg SS, Homesley HD, Wilbanks GD,

Decker DG, Miller A, Park R, and Major F Jr (1990). Adjuvant therapy

in stage I and stage II epithelial ovarian cancer. Results of two prospec-

tive randomized trials. N Engl J Med 322 (15), 1021–1027.

[2] Swenerton KD, Hislop TG, Spinelli J, LeRiche JC, Yang N, and Boyes

Figure 5. Differential expression of haptoglobin in ovarian tumor tissue samples examined by immunohistochemistry. Haptoglobin is expressed in both late-stage

ovarian cancer (A, serous; B, mucinous) and benign ovarian tumor (C, serous; D, mucinous). Expression level is higher in ovarian cancer than in benign tumors.

Hematoxylin –eosin staining of corresponding sections is also included (E–H). Original magnification, �200.

6 Haptoglobin: Prognostic Factor in Ovarian Cancer Zhao et al.

Neoplasia . Vol. 9, No. 1, 2007

DA (1985). Ovarian carcinoma: a multivariate analysis of prognostic

factors. Obstet Gynecol 65 (2), 264–270.

[3] Meyer T and Rustin GJ (2000). Role of tumour markers in monitoring

epithelial ovarian cancer. Br J Cancer 82 (9), 1535–1538.

[4] Gerner-Smidt P, Friedrich U, Petersen GB, and Tischfield JA (1978). A

balanced translocation t(11;16)(q13;p11), a cytogenetic study and an

attempt at gene localization. Hum Genet 42 (1), 61–66.

[5] Beckman G, Eklund A, Frohlander N, and Stjernberg N (1986). Hapto-

globin groups and lung cancer. Hum Hered 36 (4), 258–260.

[6] Benkmann HG, Hanssen HP, Ovenbeck R, and Goedde HW (1987).

Distribution of alpha-1-antitrypsin and haptoglobin phenotypes in blad-

der cancer patients. Hum Hered 37 (5), 290–293.

[7] Mitchell RJ, Carzino R, and Janardhana V (1988). Associations be-

tween the two serum proteins haptoglobin and transferrin and leukae-

mia. Hum Hered 38 (3), 144–150.

[8] Awadallah SM and Atoum MF (2004). Haptoglobin polymorphism in

breast cancer patients form Jordan. Clin Chim Acta 341 (1–2), 17–21.

[9] Epelbaum R, Shalitin C, Segal R, Valansi C, Arselan I, Faraggi D,

Leviov M, Ben-Shahar M, and Haim N (1998). Haptoglobin-related pro-

tein as a serum marker in malignant lymphoma. Pathol Oncol Res 4 (4),

271–276.

[10] Dunzendorfer U, Jung K, and Ohlenschlager G (1980). Transferrin, C3

complement, haptoglobin, plasminogen and alpha 2-microglobulin in

patients with urogenital tumors. Eur Urol 6 (4), 232–236.

[11] An JY, Fan ZM, Zhuang ZH, Qin YR, Gao SS, Li JL, and Wang LD

(2004). Proteomic analysis of blood level of proteins before and after

operation in patients with esophageal squamous cell carcinoma at high-

incidence area in Henan Province. World J Gastroenterol 10 (22),

3365–3368.

[12] Dobryszycka W, Gerber J, Zuwala-Jagiello J, and Ujec M (1991). Acute

phase reactants and circulating immune complexes in patients with

ovarian carcinoma. Arch Immunol Ther Exp (Warsaw) 39 (1 – 2),

41–50.

[13] Elg SA, Carson LF, Fowler JM, Twiggs LB, Moradi MM, and

Ramakrishnan S (1993). Ascites levels of haptoglobin in patients with

ovarian cancer. Cancer 71 (12), 3938–3941.

[14] Ye B, Cramer DW, Skates SJ, Gygi SP, Pratomo V, Fu L, Horick NK,

Licklider LJ, Schorge JO, Berkowitz RS, et al. (2003). Haptoglobin-

alpha subunit as potential serum biomarker in ovarian cancer: iden-

tification and characterization using proteomic profiling and mass

spectrometry. Clin Cancer Res 9 (8), 2904–2911.

[15] Ahmed N, Barker G, Oliva KT, Hoffmann P, Riley C, Reeve S, Smith AI,

Kemp BE, Quinn MA, and Rice GE (2004). Proteomic-based identifica-

tion of haptoglobin-1 precursor as a novel circulating biomarker of ovar-

ian cancer. Br J Cancer 91 (1), 129–140.

[16] Ahmed N, Oliva KT, Barker G, Hoffmann P, Reeve S, Smith IA, Quinn

MA, and Rice GE (2005). Proteomic tracking of serum protein iso-

forms as screening biomarkers of ovarian cancer. Proteomics 5 (17),

4625–4636.

[17] FIGO Cancer Committee (1986). Staging announcement: FIGO Cancer

Committee. Gynecol Oncol 25, 383–385.

[18] Maccio A, Madeddu C, Massa D, Mudu MC, Lusso MR, Gramignano G,

Serpe R, Melis GB, and Mantovani G (2005). Hemoglobin levels cor-

relate with interleukin-6 levels in patients with advanced untreated epi-

thelial ovarian cancer: role of inflammation in cancer-related anemia.

Blood 106 (1), 362–367.

[19] Agarwal R, Loganath A, Roy AC, Wong YC, and Ng SC (2001). Expres-

sion profiles of interleukin-15 in early and late gestational human pla-

centa and in pre-eclamptic placenta. Mol Hum Reprod 7 (1), 97–101.

[20] Wang H, Gao XH, Wang YK, Li P, He CD, Xie Y, and Chen HD (2005).

Expression of haptoglobin in human keratinocytes and Langerhans

cells. Br J Dermatol 153 (5), 894–899.

[21] Zhang D, Salto-Tellez M, Putti TC, Do E, and Koay ES (2003). Reli-

ability of tissue microarrays in detecting protein expression and gene

amplification in breast cancer. Mod Pathol 16 (1), 79–84.

[22] Tseng CF, Huang HY, Yang YT, and Mao SJ (2004). Purification of

human haptoglobin 1-1, 2-1, and 2-2 using monoclonal antibody affinity

chromatography. Protein Expr Purif 33 (2), 265–273.

[23] Warwas M, Dobryszycka W, Gerber J, and Pietkiewicz A (1981). Clini-

cal usefulness of serum acute-phase reactants in patients with ovarian

tumors. Neoplasma 28 (4), 485–490.

[24] Dobryszycka W, Katnik-Prastowska I, Gerber J, Lemanska K, Utko K,

and Rozdolski K (1999). Serum haptoglobin, CA 125 and interleukin

6 levels in malignant and non-malignant tumors of the ovary. Arch Im-

munol Ther Exp (Warsaw) 47 (4), 229–236.

[25] Helzlsouer KJ, Erlinger TP, and Platz EA (2006). C-reactive protein

levels and subsequent cancer outcomes: results from a prospective

cohort study. Eur J Cancer 42 (6), 704–707.

[26] Thompson S and Turner GA (1987). Elevated levels of abnormally-

fucosylated haptoglobins in cancer sera. Br J Cancer 56 (5), 605–610.

[27] Thompson S, Dargan E, and Turner GA (1992). Increased fucosylation

and other carbohydrate changes in haptoglobin in ovarian cancer. Cancer

Lett 66 (1), 43–48.

[28] Thompson S, Cantwell BM, Cornell C, and Turner GA (1991). Abnormally-

fucosylated haptoglobin: a cancer marker for tumour burden but not

gross liver metastasis. Br J Cancer 64 (2), 386–390.

[29] Thompson S, Cantwell BM, Matta KL, and Turner GA (1992). Parallel

changes in the blood levels of abnormally-fucosylated haptoglobin and

alpha 1,3 fucosyltransferase in relationship to tumour burden: more

evidence for a disturbance of fucose metabolism in cancer. Cancer Lett

65 (2), 115–121.

[30] Okuyama N, Ide Y, Nakano M, Nakagawa T, Yamanaka K, Moriwaki K,

Murata K, Ohigashi H, Yokoyama S, Eguchi H, et al. (2006). Fuco-

sylated haptoglobin is a novel marker for pancreatic cancer: a detailed

analysis of the oligosaccharide structure and a possible mechanism for

fucosylation. Int J Cancer 118 (11), 2803–2808.

[31] Smeets MB, Pasterkamp G, Lim SK, Velema E, van Middelaar B, and

de Kleijn DP (2002). Nitric oxide synthesis is involved in arterial hapto-

globin expression after sustained flow changes. FEBS Lett 529 (2–3),

221–224.

[32] Westermarck J and Kahari VM (1999). Regulation of matrix metallo-

proteinase expression in tumor invasion. FASEB J 13 (8), 781–792.

Haptoglobin: Prognostic Factor in Ovarian Cancer Zhao et al. 7

Neoplasia . Vol. 9, No. 1, 2007

Figure W1. Western blot analysis showing the binding capacity of antibodies used in ELISA (A, rabbit anti-human haptoglobin polyclonal antibody; B, goat anti-

human haptoglobin antibody). Haptoglobin standard (5 �g; Lanes 1 and 6) and 10 �g of protein from ovarian cancer (Lanes 2, 3, 7, and 8) and benign ovarian

tumors (Lanes 4, 5, 9, and 10) were loaded into each well.

Figure W2. ROC curves for serum haptoglobin. Using different cutoff

coordinate points, a series of sensitivity and specificity would be generated. It

is noted that at the 4.0-mg/ml point, the greatest value of the sum of sen-

sitivity and specificity could be achieved.

Figure W3. Single regression analysis between haptoglobin and CRP.

Haptoglobin was significantly correlated to CRP (P < .001). The central

straight line corresponds to the best-fit linear regression line. The two curved

lines surrounding the best-fit line define the 95% CI of regression line. The

graphs also show prediction interval (the curves defining the prediction

interval are further from the regression line than from confidence lines).