Pancreatic cancer is currently one of the leading causes of cancer-related deaths and the overall 5-year survival has been reported to be less than 5%.1, 2 One of the reasons for its poor prognosis is that an early diagnosis is quite difficult and a high-risk population for pancreatic cancer has not yet been identified. Carbohydrate Antigen 19-9 (CA19-9) and carcinoembrionic antigen (CEA) are commonly used as markers of pancreatic cancer, but false positives are a problem in the diagnosis.3 To increase the specificity of a diagnosis, a combination of tumor markers would be desirable. To this end, novel markers for pancreatic cancer, which have different characteristics from those of CA19-9 or CEA, are required. Oligosaccharides are known to be one of the most important post-translational modifications, and many studies have shown that changes in oligosaccharide structures occur during inflammation and tumorigenesis.4 This oligosaccharide heterogeneity has been applied to tumor markers for the differential diagnosis for Hepatocellular Carcinoma (HCC). Alpha-fetoprotein (AFP), a well-known tumor marker for HCC, contains 1 asparagine-linked oligosaccharide.5 However, serum levels of AFP also increase in certain patients with chronic hepatitis and liver cirrhosis. α 1-6 fucosylated AFP (AFP-L3 fraction) has been applied to the clinical diagnosis of HCC. α 1-6 fucosylated AFP, which is produced via α 1-6 fucosyltransferase (FUT8), is specifically found in the serum of patients with HCC and can be diagnosed by measuring the Lentil Lectin-(LCA) binding portion of AFP.6, 7 Moreover, AFP-L3 has been reported as a marker for a poor prognosis of HCC.8 Changes in fucosylation patterns, as the result of different levels of expression for various fucosyltransferases, have been reported in certain diseases including various types of cancers.9, 10, 11, 12
To identify potentially novel tumor markers of pancreatic cancer, we conducted a search for fucosylated proteins that are increased in the serum of patients with pancreatic cancer. The findings showed that the haptoglobin β chain was highly fucosylated and the oligosaccharide structures of haptoglobin purified from the serum of patients with pancreatic cancer were examined in detail. Furthermore, we investigated the mechanisms associated with the increased levels of fucosylated haptoglobin in pancreatic cancer.
Serum samples of patients with pancreatic cancer (n = 49, male 31, female 18, mean age 62 years), HCC (n = 23, male 17, female 6, mean age 69 years), liver cirrhosis (n = 12, male 9, female 3, mean age 63 years), gastric cancer (n = 10, male 5, female 5, mean age 59 years) and colon cancer (n = 17, male 10, female 7, mean age 61 years) were obtained from Osaka National Hospital, Osaka University Hospital and Osaka Medical Center for Cancer and CVD. The present project was approved by the ethics committees of the participating hospitals. Serum samples of healthy volunteers (n = 30, male 16, female 14, mean age 34 years) were obtained in our laboratory.
Human pancreatic carcinoma cell lines (PK8, PANC-1, PSN-1, KMP4, KLM-1 and MIAPaCa2) and a human hepatoma cell line, Hep3B, were grown in RPMI-1640 (Nacalai Tesq, Kyoto, Japan) supplemented with 10% fetal bovine serum (FBS), 50 U/ml penicillin and 100 μg/ml kanamycin at 37°C in 5% CO2. These cell lines were obtained from the ATCC (American Type Culture Collection), or the Institute of Development, Aging and Cancer, Tohoku University.
Identification of fucosylated proteins in the serum of patients with pancreatic cancer
A 0.5 μl aliquot of serum proteins from patients with pancreatic cancer and normal controls were electrophoresed on 8% polyacrylamide gels in duplicate. One gel was used for the aleuria aurantica lectin (AAL) blot analysis, which preferentially recognized α 1-3/α 1-6 fucosylated proteins.13 The other was stained with Coomassie Brilliant Blue (CBB) after transferring onto a PVDF membrane. All procedure of AAL lectin blot analyses was described previously.14 Bands strongly stained with AAL were subjected to N-terminal amino-acid sequence.
Western blot analysis of haptoglobin and immunoprecipitation
A 0.5 μl aliquot of serum was electrophoresed on an 8% polyacrylamide gel and transferred onto a nitrocellulose membrane (Schleicher & Schuell, Dassel, Germany). The membranes were incubated with 5% skim milk in phosphate buffered saline (PBS) overnight and then incubated with 1/1,000 diluted anti human haptoglobin antibody (Dako Cytomation Kyoto, Japan) for 2 hr. After washing 3 times with Tris-buffered saline-T (TBS) (136 mM NaCl, 2.6 mM KCl, 24 mM Tris, 0.05% Tween 20, pH 7.4) for 10 min each, the membrane was incubated with peroxidase-conjugated rabbit IgG for 1 hr. After washing the membrane 3 times with TBS-T for 10 min each, development was performed using an ECL™ Western Blotting Detection Reagents (Amersham Biosciences, Uppsala, Sweden), according to standard protocols. The same membrane was used in an AAL lectin blot analysis.14 In all AAL lectin blotting experiments, 1 pair of a negative control (a healthy control) and a positive control (a case of pancreatic cancer) was used in the same gel. For the immunoprecipitation of haptoglobin, 5 μl samples of serum from patients with pancreatic cancer and from controls were used. Serum samples were preincubated with normal rabbit serum and proteinG-sepahrose (Amersham Bioscience) followed by incubation with anti-human haptoglobin antibody for 2 hr. Immunoprecipitaed haptoglobin was analyzed by AAL lectin blot, as described earlier.
Purification of the haptoglobin β chain
To purify the haptoglobin β chain, 80 μl of sera in which albumin was depleted by a Montage Albumin Deplete kit (Millipore Corp.) or 2.5 ml of 100-fold concentrated conditioned media from PSN-1 cells were applied to an anti-haptoglobin affinity column that was coupled with 300 μl of anti human haptoglobin antibody, according to standard protocols of HiTrap NHS-activated HP (Amersham Biosciences). The haptoglobin bound to the column was eluted with 5 ml of elution buffer (100 mM Glycine, 0.5 M NaCl, pH 3.0). Thirty microliters of 50-fold concentrated fraction was subjected to SDS-PAGE under reducing conditions and stained with CBB or blotted onto a nitrocellulose membrane followed by lectin blot analyses using AAL or aspergillus oryzae lectin (AOL).15
Mass spectrometry was used to identify the structure of the oligosaccharide in haptoglobins. The gels that contained purified haptoglobin were cut into smaller sizes and collected in a 1.5-ml microtube. To remove CBB, 50 mM NH4HCO3 (SIGMA, Tokyo Japan) in 30% acetonitrile (MERCK, Darmstadt Germany) was added, followed by washing at room temperature for 20 min using Bio shaker (TAITEC). The samples were then added with 300 μl of acetonitrile and incubated at room temperature for 10 min. After removing the extra acetonitrile, a reduction solution consisting of 10 mM DTT, 10 mM EDTA and 50 mM NH4HCO3 was added, followed by incubation at 65°C for 60 min. Samples were then alkylated in a solution consisting of 40 mM idoacetamide, 10 mM EDTA and 50 mM NH4HCO3 in the dark for 30 min. After washing twice with 50 mM NH4HCO3 for 10 min, an additional 300 μl of acetonitrile was added and the sample was then incubated at room temperature for 10 min. For trypsin digestion, the samples were incubated at 37°C overnight with 0.5 μg of sequencing grade-modified trypsin (Promega, Madison, WI USA) in 50 mM NH4HCO3. After the gels were removed, the sample was concentrated and taken to dryness with a Speed Vac (CENTRIFUGAL EVAPORATOR CVE-2000, EYELA). The residues were dissolved in 20 μl of water. A 2-μl aliquot of this solution was used in the Mascot research. The other sample was incubated at 100°C for 10 min with 32 μl of a 20 mM phosphate solution. The samples were then treated with N-Glycosidase F rec. [E. coli] (Roche, USA) and incubated at 37°C overnight. After boiling at 100°C for 10 min, PA (pyridylamino) modification was performed using a Glyco TAG™ Reagent Kit (TaKaRa, Otsu Japan), according to the standard protocols. The samples were filtered with Sephadex LH-20 and N-glycans derived with the PA fraction were collected. The samples were dried with a Speed Vac and then dissolved in 100 μl of water. Liquid chromatography-electrospray ionization mass spectrometry (LC-ESI-MS) and matrix-assisted laser desorption ionization-time of flight–mass spectrometry (MALDI-TOF-MS) were then performed.
RNA extraction and RT-PCR
A human hepatoma cell line (Hep3B) and human pancreatic carcinoma cell lines (PSN-1, KLM-1, MiaPaCa-2, PK8, PK59, PANC-1) were cultured as described earlier. Trizol (1 ml) was added to each 10-cm dish and collected in a 2-ml microtube. After 15 min, 200 μl of chloroform was added to the samples followed, by vortexing for 15 sec. After standing at room temperature for 10 min, the samples were centrifuged at 15,000 rpm for 15 min, and an equal amount of 2-propanol was then added to the supernatant. After an additional 15 min, the samples were centrifuged at 15,000 rpm at 4°C for 15 min and the pellets were washed with 0.5 ml of 75% ethanol twice. The pellets were dried and dissolved in 50 μl of DEPC (diethylpyrocarbonate) treated water. The concentration of RNAs was measured at an absorbance of 260 nm.
According to the SuperScript™(III) Reverse Transcriptase (Invitrogen Corp. Carlsbad, CA UA) protocol, 5 μg of total RNA was incubated with 1 μl of Oligo dT at 70°C for 10 min. The samples were incubated at 42°C for 5 min with a 1st strand cDNA synthesis buffer consisting of 10 μl of 5× First Strand Buffer, 10 μl of dNTP Mixture, 5 μl of 0.1 M DTT and 13 μl of DEPC, at 42°C for 50 min with 1 μl of Reverse Transcriptase, at 99°C for 5 min and at 37°C for 20 min with 1 μl of RNaseH.
The samples served as a template DNA for 30 rounds of amplification using the GeneAmp PCR System 2700. PCR was performed in a standard 100 μl reaction mixture consisting of 10 μl of 10× Ex taq Buffer, 8 μl of dNTP Mixture, 1 μl of sense and antisense primer, 0.5 μl of Ex taq (TaKaRa), 1 μl of cDNA, PCR primers for haptoglobin cDNA were as follow, forward primer, 5′-TTCCCTGGCAGGCTAAGATG-3′(position 562-581); and reverse primer, 5′-GCACCCATCAGCTTCAAACC-3′(position 1363-1382). Amplification was performed at 95°C for 30 sec, at 66°C for 30 sec, at 72°C for 1 min. Finally, an additional extension step was performed at 72°C for 10 min. The amplified PCR products were run on a 1.5% agarose gel containing 0.005% ethidium bromide. To estimate the amount of total cDNA, glyceraldehydes-3-phosphate dehydrogenase with the same cDNA was used as an internal control under identical conditions.
Induction of production of fucosylated haptoglobin in hep3B cells
A human hepatoma cell line (Hep3B) and human pancreatic carcinoma cell lines (PSN-1, MiaPaCa-2) were grown in low glucose D-MEM (Nacalai Tesq, Kyoto, Japan) supplemented with 10% FBS, 50 U/ml penicillin and 100 μg/ml kanamycin at 37°C in 5% CO2. The cells, at sub confluent conditions in 10-cm dishes, were washed twice with PBS and cultured in 10 ml of serum free D-MEM. After 2 days, the media from each run were collected and added to the other cells. After an additional incubation for 2 days, the media were collected. The media were concentrated 100 times and used in a western blot analysis of haptoglobin.
Haptoglobin, as a target protein for fucosylation in the serumof patients with pancreatic cancer
A preliminary study suggested that pancreatic cancer cells produce a variety of fucosylated proteins into the condition medium. To identify fucosylated proteins in the serum of patients with pancreatic cancer, AAL blot analyses were performed. The total binding of serum proteins to AAL was increased in pancreatic cancer as compared with healthy controls. In these proteins, increases in the fucosylation of the ∼40 kD band were observed with a high frequency in the serum of patients with pancreatic cancer. The N-terminal amino-acid sequences revealed that the sequence was ILGGHLDAKG, corresponding to the haptoglobin β chain (Fig. 1). A similar approach was performed in 4 cases of pancreatic cancer and all of the fucosylated proteins of 40 kD identified were the haptoglobin β chain (data not shown). Furthermore, a western blot analysis of haptoglobin was performed to confirm the position of haptoglobin molecular size (Fig. 1).
The appearance of fucosylation of haptoglobin in serum of patients with various cancers
To evaluate the levels of fucosylation of haptoglobin in the serum of patients with pancreatic cancer compared with those of other various cancers, an AAL blot analysis was performed (Fig. 2a). The results showed that fucosylated haptoglobin was also increased in HCC, liver cirrhosis (date not shown), gastric cancer and colon cancer. Interestingly, the appearance of the fucosylation was not correlated with total amount of haptoglobin. The immunoprecipitation of haptoglobin, followed by an AAL lectin blot showed more clearly that haptoglobin was strongly fucosylated in patients with pancreatic cancer. The incidence of increase in fucosylated haptoglobin in the serum of patients with various diseases is summarized (Table I). Appearance of fucosylated haptoglobin in the case of pancreatic cancer was significantly higher compared with that of healthy controls and patients with HCC, liver cirrhosis and gastric cancer.
Table I. Fucosylation of Haptoglobin in Serum of Patients with Various Diseases
Statistic analysis was performed according to the program for Stat-view software.
Relationship between fucosylation of haptoglobin and the clinical stage in pancreatic cancer
The appearance of fucosylation in pancreatic cancer was investigated in terms of the clinical background of the subjects (Table II). The fucosylation of haptoglobin was observed in 4/12 cases at stage I and II, and 15/22 cases at stage III and IV, respectively, suggesting that the incidence of haptoglobin fucosylation tended to increase in advanced stages. Interestingly, fucosylated haptoglobin disappeared after an operation in 2 cases in which it was possible to follow-up.
Table II. Relationship between the Incidence of Fucosylated Haptoglobin and the Clinical Stage of Pancreatic Cancer
p = 0.05, compared with stage I, II and stage III, IV (χ2 test). Statistic analysis was performed according to the program of Stat-view software.
Stage I, II
Stage III, IV
Analysis of oligosaccharide structures of haptoglobin by lectin blot and mass spectrometry
To determine the oligosaccharide structures, haptoglobin was purified from 80 μl of sera of pancreatic cancer and healthy individuals. The detailed procedure is described in “Material and methods”. Five microliters of a 50-fold concentrated fraction was subjected to SDS-PAGE and then stained with CBB. A major band was detected at the expected molecular weight (Fig. 3a). To determine the oligosaccharide structures in the purified haptoglobin β chain, lectin blot analyses using AAL and AOL were performed (Fig. 3b). AOL specifically interacts with core fucosylation. The results indicate that α 1-3 fucosylation as well as core fucosylation were both increased in the haptoglobin β chain of pancreatic cancer patients.
Thirty microliters of 50-fold concentrated haptoglobin was subjected to SDS-PAGE and the 40 kD band was excised from the gel. This purified protein was confirmed to be the haptoglobin β chain by MALDI-TOF mass spectrometry (data not shown). To determine the oligosaccharide structures of haptoglobin β chain in more detail, LC-ESI-MS was performed (Figs. 4a–4c). A high level of fucosylation was observed in the case of haptoglobin associated with pancreatic cancer. Furthermore, biantennary chains with disialic acid, which are considered to be the major oligosaccharide structures, were analyzed by MS/MS. This analysis showed that a high level of core fucosylation is associated with pancreatic cancer (data not shown).
To determine the oligosaccharide structures of triantennary structures with trisialic acid, MALDI-TOF-MS was performed. In this experiment, fucose was found to be attached to the α 1-3/α 1-4 position to GlcNAc or α 1-2 position to Galactose. As the result of the mass spectrometry analysis, core fucosylation as well as α 1-3/α 1-4 fucosylation was confirmed to be increased in the haptoglobin β chain purified from serum of patients with pancreatic cancer (Fig. 4d).
Mechanisms responsible for the increases in fucosylated haptoglobin
While most haptoglobin is secreted from the liver, the expression of FUT8 in a normal liver is quite low.16 Therefore, a normal liver does not produce α 1-6 fucosylated haptoglobin. There are 2 possible mechanisms underlying the increased levels of fucosylated haptoglobin in the serum of patients with pancreatic cancer. One is that pancreatic cancer cells, themselves, produce fucosylated haptoglobin. To investigate this possibility, we performed a RT-PCR analysis of haptoglobin using 6 types of pancreatic cancer cells (Fig. 5a). The expression of haptoglobin mRNA was observed only in PSN-1 cells. After the purification of haptoglobin from conditioned media of these cells, the oligosaccharide structures were analyzed. Expectedly, binding to AAL and AOL was increased in haptoglobin purified from PSN-1 cells (Fig. 5b). Moreover, an LC-ESI-MS analysis indicated that core fucosylation as well as the α 1-3/α 1-4 fucosylation of haptoglobin were observed in the conditioned media of PSN-1 cells (data not shown). The other possibility is that pancreatic cancer produces a factor which induces the production of fucosylated haptoglobin from the liver. To examine this hypothesis further, media from pancreatic cancer cells such as PSN-1 and MIA PaCa-2 were added to a hepatoma cell line, Hep3B. Increase of haptoglobin production was observed in Hep3B cells after addition of the conditioned media of these pancreatic cancer cells (Fig. 6).
To find a novel marker for cancers, it is important to identify a protein that is secreted exclusively from cancer cells or to identify a specific modification of a protein that is produced by cancer cells. The best way for the former analysis would be a DNA micro array, and for the latter analysis would be of the detection of modified sugar chains. In the present study, we found that fucosylated haptoglobin was a good serum marker for pancreatic cancer, analyzed the oligosaccharide structure in detail and investigated the mechanism underlying why fucosylated haptoglobin is increased in the serum of patients with pancreatic cancer.
As a result of analyses of various serum samples, we found that fucosylated haptoglobin was observed at high levels in the serum of patients with pancreatic cancer. Haptoglobin is heterotetramer consisting of 2 α subunits and 2 β subunits joined by inter-chain disulfide bonds.17 There are 4 distinct asparagines residues (Asn 23, 46, 50, 80) in each β-chain and they display oligosaccharide heterogeneity. Recent studies of haptoglobin showed that certain oligosaccharide structures predominate in different diseases. For example, a highly-fucosylated structure is found in breast cancer and ovarian cancer, highly-sialylated structures in Crohn's disease and highly branched structures in alcoholic liver disease.18, 19, 20, 21, 22, 23 Furthermore, the aberrant glycosylation of haptoglobin was found to increase during mouse hepatocarcinogenesis, by our group.24 In our study, we reported that increases in core fucosylation as well as α 1-3/α 1-4 fucosylation was found in the haptoglobin β chain purified from serum of patients with pancreatic cancer compared to normal controls by LC-ESI-MS and MALDI-TOF-MS. Furthermore, we described 2 possibilities for the fucosylation of haptoglobin found in the serum of patients with pancreatic cancer. A pancreatic cancer cell line, PSN-1 actually produced fucosylated haptoglobin, suggesting that pancreatic cancer itself produces fucosylated (especially α 1-6 fucosylated) haptoglobin. To prove this possibility, the immunohistochemistry of haptoglobin was undertaken. Infiltrating lymphocytes could express ectopic haptoglobin in pancreatic cancer tissues. Secondly, pancreatic cancer produces a factor that induces the production of fucosylated (especially α 1-3 fucosylated) haptoglobin from the liver. To demonstrate this hypothesis, it will be necessary to identify a factor produced by pancreatic cancer. As shown in Figure 6, it would be difficult to know whether or not fucosylated haptoglobin is increased in a normal liver, because Hep3B is a cancer cell line and secretes high levels of fucosylated haptoglobin. Fucosylated haptoglobin disappeared after an operation, indicating that both of these 2 possibilities could exist in vivo.
In conclusion, we reported on the potential use of haptoglobin as a target protein for fucosylation in the serum of patients with pancreatic cancer. We also found that the α 1-3/α 1-4 fucosylation as well as the α 1-6 fucosylation of haptoglobin was specifically detected in pancreatic cancer, as evidenced by mass spectrometry. We conclude that there are 2 possibilities for the fucosylation of haptoglobin in pancreatic cancer. Further studies will be required to verify the clinical use of fucosylated haptoglobin as a tumor marker in terms of comparison with inflammatory diseases such as chronic pancreatitis (a preliminary study in the cases of chronic pancreatitis showed 25% positive (1/4 cases) for fucosylated haptoglobin in their serum).