It has previously been shown that heparanase-1 (HPR1), an endoglycosidase, is up-regulated in pancreatic carcinoma. The purpose of this study was to test whether serum HPR1 levels in pancreatic carcinoma patients are elevated, and whether higher serum HPR1 levels are associated with a shortened survival.
Serum HPR1 levels in 40 healthy donors, 31 pancreatic carcinoma patients, and 11 patients treated with gemcitabine were measured by a novel enzyme-linked immunoadsorbent assay. HPR1 expression in tumors was analyzed by immunohistochemical staining. Patient overall survival time was determined according to the Kaplan–Meier method, and their difference was evaluated by the log-rank test. A P value < 0.05 was considered statistically significant.
The mean serum HPR1 activity in pancreatic carcinoma patients was 439 ± 14 units/mL, compared with 190 ± 4 units/mL in the control serum samples from healthy donors. Serum HPR1 levels were significantly higher in patients with HPR1-positive tumors (660 ± 62 units/mL) compared with those with HPR1-negative tumors (241 ± 14 units/mL). The mean survival of 19 pancreatic carcinoma patients with serum HPR1 activity > 300 units/mL was 7.9 ± 0.2 months, whereas the mean survival of 12 patients with serum HPR1 activity < 300 units/mL was 13.3 ± 0.6 months. A Kaplan–Meier plot of the patient survival curve followed by log-rank test revealed that patients in the high serum HPR1 group had a significantly shorter survival compared with those in the low serum HPR1 group. Mean serum HPR1 activity decreased by 64% in 11 pancreatic carcinoma patients after 2 weeks of treatment with gemcitabine.
Pancreatic adenocarcinoma is among the most aggressive tumors, causing approximately 30,000 deaths per year in the U.S. Approximately 80% of pancreatic carcinoma patients have regional and/or distant metastases at the time of diagnosis.1–4 These patients have a very poor prognosis, with an overall 5-year survival of < 2% and a mean survival of 5–6 months.1–4 Approximately 15–20% of pancreatic carcinoma patients have resectable disease, but only approximately 20% of these patients survive to 5 years. For patients with locally advanced, unresectable, and metastatic disease, 5-fluorouracil-based chemoradiation can provide some palliative benefits.5 Gemcitabine, a deoxycytidine analog with structural and metabolic similarities to cytarabine, is currently the standard therapy for metastatic pancreatic carcinoma.6 Gemcitabine treatment is reported to improve patient survival better than 5-fluorouracil (median survival of 5.6 mos vs. 4.4 mos). In addition, more clinically meaningful effects on disease-related symptoms (pain, performance status, and weight) were achieved with gemcitabine than with fluorouracil (24% vs. 4%).
Heparanase-1 (HPR1) is an endoglycosidase that specifically cleaves the heparan sulfate (HS) chains of proteoglycans (HSPGs),7–15 a main component required to form the complex network with other extracellular matrix (ECM) components such as fibronectin, laminin, Type IV collagen, and vitronectin in the basement membrane (BM) and ECM.16 Breakdown of the barriers provided by the ECM allows tumor cells to invade local tissues and to migrate to distant sites.10, 15 HPR1 may also promote tumor angiogenesis by releasing growth factors and angiogenic molecules such as basic fibroblastic growth factor (bFGF) that are sequestered in depot form within HSPG,10, 15 making them available to bind and activate their tyrosine kinase receptors and to promote endothelial cell proliferation and neovasculization.
HPR1 is synthesized as a 65-kilodalton (kD) inactive proenzyme but can be processed by cleavage to form a heterodimer of a 50- and a 8-kD polypeptide and to become active under acidic conditions with an optimal pH value of 4.8–5.5.17–20 Although both forms of HPR1 have been detected as soluble molecules12, 21 in the conditioned medium, the 65-kD HPR1 is predominantly present in the supernatant fluids of human HPR1-transfected Eb lymphoma cells and COS-7 cells. We and others have previously demonstrated that HPR1 expression is detected in approximately 70% of pancreatic adenocarcinomas, and that HPR1 expression is associated with a poorer survival.22–24 The current study examined whether HPR1 expression in pancreatic adenocarcinoma patients leads to elevated serum HPR1 levels, and whether serum HPR1 activity can be used to predict the prognosis of the tumor and its chemotherapeutic response.
MATERIALS AND METHODS
Patient Information, Tumor Specimens, and Blood Samples
The use of specimens from human subjects was approved by the Institutional Review Board (IRB) of Rush University Medical Center. Paraffin-embedded tissue blocks derived from patients with primary pancreatic carcinoma were obtained from the Department of Pathology. Peripheral blood samples were collected at the time of final diagnosis with IRB approval and patient consent from 31 pancreatic carcinoma patients (22 males and 9 females) by the medical oncology service over a 55-month period. Mean patient age at the time the blood samples were collected was 62.8 years (median age, 63 yrs; range, 38–80 yrs). Based on the American Joint Committee on Cancer (AJCC) Staging System, one patient had Stage I disease, 1 patient was diagnosed with Stage II disease, 3 patients were diagnosed with Stage III disease, and the remaining 26 patients were diagnosed with Stage IV disease. Among these 31 patients, 13 underwent surgical resection of their tumors that were later used for analysis of HPR1 expression using immunohistochemical (IHC) staining. Four patients received chemotherapy (two with gemcitabine, one with gemcitabine plus cisplatin, and one with cytoxan plus vincristine). Blood samples were also drawn after final diagnosis but before surgery or chemotherapy. Patient survival was calculated by counting the number of months from the time of diagnosis of pancreatic carcinoma to death. Forty blood samples drawn from healthy donors (16 males and 24 females, with a mean age of 35 yrs and a median age of 33 yrs [range, 21–60 yrs]) were used as controls.
Gemcitabine (Gemzar®; Eli Lilly, Indianapolis, IN) was purchased as a lyophilized powder in 200-mg and 1000-mg vials and stored at room temperature. It was reconstituted with saline and given to 11 patients at a dose of 1000 mg/m2 weekly by an intravenous infusion in 250 mL over 30 minutes. These subjects did not include any of 31 patients described earlier. Patients were treated for 3 weeks. Blood samples from 11 patients were drawn before and 2 weeks after drug administration. Serum samples were prepared, aliquoted, and stored at –80 °C until use.
The sections of tumor tissues from 13 of 31 patients whose serum HPR1 levels had been measured were dewaxed with xylene and rehydrated. Slides were then heated in 10 mM of sodium citrate (pH 6.0) in a microwave for 3 minutes for antigen retrieval. Cooled slides were rinsed with phosphate-buffered saline (PBS) and then incubated with 1% hydrogen peroxide in methanol for 30 minutes at room temperature. Sections were then blocked with 5% normal goat serum in PBS for 30 minutes at room temperature followed by a 1-hour incubation with an anti-HPR1 rabbit serum (1:500 dilution) in PBS. Slides were washed and then incubated with biotinated goat antirabbit antibody (PharMingen, San Diego, CA) diluted at 1:300 in PBS/5% human serum. Streptavidin-HRP conjugate (Zymed, San Francisco, CA) diluted at 1:200 in PBS with 5% normal human serum was added and incubated for 45 minutes at room temperature. Color development was done with diaminobenzidine (DAB) substrate (Sigma Chemical Company, St. Louis, MO) followed by DAB enhancer (Vector Laboratories, Burlingame, CA). Slides were counterstained with Mayer hematoxylin for 2 minutes, rinsed, dehydrated, and mounted. HPR1 expression was independently examined by a pathologist (P.G.) and two investigators (R.M.Q. and X.X.).
Quantitation of Serum HPR Activity
A novel enzyme-linked immunoadsorbent assay (ELISA) was developed in this laboratory to measure HPR1 activity.25 Briefly, Matrigel (BD Biosciences, San Diego, CA), an artificial basement membrane that contains abundant HSPG, was dissolved in ice-cold PBS (0.1 M [pH 7.4])/carbonate buffered saline (0.1 M [pH 9.6]) (volume:volume of 50:50) at a concentration of 20 μg/mL and used to coat ELISA plates (25 μL/well) at 4 °C overnight. The plates then were washed 3 times with PBS containing 0.05% Tween-20 and blocked with 5% bovine serum albumin (BSA) in PBS at room temperature for 1 hour. Serum samples were serially diluted at 1:5 in HPR1 assay buffer (0.1 M sodium acetate [pH 5.0], 0.1 mg/mL BSA, 0.01% Triton X-100, 0.5 mM phenylmethylsulfonyl fluoride, and 10 μg/mL leupeptin and aprotinin each) and added to each well and incubated at 37 °C overnight. Anti-HS-specific MoAb (clone HepSS; Seikagaku, Tokyo, Japan) (1:1,000, diluted in PBS containing 5% BSA) was added and incubated at room temperature for 1 hour. After washing, horseradish peroxidase-conjugated goat antimouse immunoglobulin M (IgM) antibody (1:2,000 diluted in PBS with 5% BSA) was added and incubated at room temperature for 1 hour, followed by addition of 50 μL ABTS (2,2-azino-bis-(3-ethylbenzthiazoline-6-sulfonic acid) substrate. The OD405 absorbance was read in an ELISA plate reader (Bio-Rad, Hercules, CA). HPR1 activity in serum samples was calculated based on the standard curve of serially diluted purified platelet HPR1 (starting at 1:200) at a concentration of 1 μL HPR1 with the activity of degrading 0.133 μg heparan sulfate per hour at 37 °C in HPR1 buffer. HPR1 purification and characterization from human platelets were conducted as previously reported.17 HPR1 activity was designated as per 100 units capable of degrading 1 ng heparan sulfate at 37 °C per hour in HPR1 buffer. To test the specificity of serum HPR1 activity, serum samples from four patients (three with high activity and one with low activity) were diluted at 1:10 and added to an ELISA plate in the presence of various concentrations of PI-88 (0, 0.5, 1, and 2 μg/mL), an HPR1 inhibitor that has been previously reported to inhibit HPR1 activity with a 50% inhibitory concentration (IC50) value of approximately 2 μg/mL26 (Progen Industries, Darra, Queensland, Australia), and incubated overnight as described earlier. To confirm that the effect of gemcitabine on serum HPR1 levels in pancreatic carcinoma patients was not due to its direct effect on HPR1 activity, gemcitabine at various concentrations (0 nM, 20 nM, 75 nM, 320 nM, and 1300 nM) was incubated with purified 2 μL of platelet HPR1 followed by incubation overnight in a Matrigel-coated ELISA plate. HPR1 activity was measured as described earlier.
Flow Cytometric Analysis
MPANC-96 cells, an HPR1-low cell line (data not shown), were grown in a 6-well plate in RPMI 1640 complete medium containing 10% fetal bovine serum. At the time of 80% confluence, cells were grown in the same medium with the addition of an HPR1-negative or HPR1-positive serum from pancreatic carcinoma patients. Purified platelet HPR1 (2 μL per well) was included as a positive control. Single cell suspensions of MPANC-96 cells were prepared by using cell dissociation solution. Cells (5 × 105/sample) were stained by incubation for 30 minutes at 4 °C with an anti-HS MoAb (0.5 μg/sample) or mouse IgM as a negative control. Cells were incubated with fluorescein isothiocyanate (FITC)-labeled goat antimouse IgM (5 μL/sample) for 30 minutes at 4 °C followed by FITC-labeled rabbit antigoat IgG (5 μL/sample) for 30 minutes at 4 °C. Cell surface HS expression was analyzed in a Becton Dickinson (San Jose, CA) flow cytometer.
A Mann-Whitney U-test was used to compare serum HPR1 levels between the control group and pancreatic carcinoma patients. A Student t test for paired data was used to determine whether serum HPR1 levels were significantly changed before and 2 weeks after gemcitabine treatment. The chi-square test was used to analyze significant differences in the clinicopathologic parameters of pancreatic carcinoma patients with serum HPR1 levels > 300 and < 300 units/mL as listed in Table 1. The differences in the survival of pancreatic carcinoma patients with serum HPR1 activity at a cutoff of 300 units/mL were determined according to the Kaplan–Meier method and these differences were evaluated by the log-rank test. A P value < 0.05 was considered statistically significant. All statistics were determined by using SigmaStat 3 software (Systat Software Inc.,Richmond, CA).
Table 1. Serum Heparanase-1 Levels Are an Independent Predictor of Patient Survival
< 300 units/mL (n = 12)
> 300 units/mL (n = 19)
AJCC: American Joint Committee on Cancer; SED: standard error of difference.
Age in yrs
Mean ± SED
13.3 ± 0.57
7.9 ± 0.24
Increased Serum HPR1 Activity in Pancreatic Carcinoma Patients
We quantitated serum HPR1 activity using a novel ELISA developed in our laboratory.25 Purified platelet HPR1 was serially diluted and used as a standard to determine the concentration of HPR1 in serum. As shown in Figure 1A, platelet HPR1 degraded HS proteoglycan in a dose-dependent manner, as reflected by a gradual decrease of OD405 values. We next measured serum HPR1 activity from 31 patients with pancreatic adenocarcinomas and 40 healthy donors. As shown in Figure 1C, the mean HPR1 activity in the serum samples of 40 healthy donors was 190 units/mL, whereas the mean serum HPR1 activity in pancreatic carcinoma patients was 439 units/mL. The mean serum HPR1 activity in pancreatic carcinoma patients was elevated by 2.3-fold. Statistical analysis revealed that serum HPR1 activity from pancreatic carcinoma patients was significantly higher than that in normal healthy donors (P < 0.001). Although the mean age of healthy subjects was significantly younger than that of pancreatic carcinoma patients (P < 0.001), it is not likely that it adversely affected the finding of increased serum HPR1 levels in pancreatic carcinoma patients because a recent study27 demonstrated that serum HPR1 activity in young healthy adults is actually higher than old adults. We speculate that the difference in serum HPR1 activity between age-matched controls and pancreatic carcinoma patients might be even larger.
To further test the specificity of HPR1 activity in serum samples, PI-88 was employed to examine its ability to inhibit HPR1 activity in four serum samples from pancreatic carcinoma patients. As shown in Figure 1B, PI-88 inhibited HPR1 activity in all samples in a dose-dependent manner. The IC50 value of PI-88 to inhibit HPR1 activity in serum samples as well as in purified platelet HPR1 was at approximately 1 μg/mL. These observations suggest that degradation of HS proteoglycans in Matrigel by purified and serum HPR1 is specific.
We next tested whether patients with HPR1-positive tumors had significantly higher serum HPR1 levels than those with HPR1-negative tumors. Tumor sections from 13 patients who had received surgical resection of their tumors were analyzed for HPR1 expression by IHC. As shown in Figure 2, HPR1 was not expressed in normal ductal cells and acinar cells in a normal pancreas (Fig. 2A). However, HPR1 was abundantly expressed in the cytoplasm of ductal cells of a carcinoma (Fig. 2B). All six specimens graded as HPR1-positive demonstrated HPR1 expression in greater than 50% of pancreatic carcinoma cells, with moderate or strong HPR1 signals. An HPR1-negative pancreatic adenocarcinoma is shown in Figure 2C. Seven HPR1-negative tumors were defined as no or minimal HPR1 signals in < 10% of tumor cells. Normal rabbit serum included as a negative control (Fig. 2D) did not give any signal. Mean serum HPR1 levels in seven patients with HPR1-negative tumors were 241 ± 14 units/mL, whereas the mean serum HPR1 levels in six patients with HPR1-positive tumors were 660 ± 62 units/mL (Fig. 1D). Serum HPR1 levels in patients with HPR1-positive tumors were significantly higher than those with HPR1-negative tumors (P = 0.007) (Fig. 1D). These observations suggest that increased serum HPR1 levels may be due to increased HPR1 expression in the tumors.
Gemcitabine is a chemotherapeutic drug that is currently being evaluated for treatment of pancreatic carcinoma in the Phase II clinical trials.28 We tested whether gemcitabine therapy led to a decrease of serum HPR1 levels. Blood samples were taken from 11 patients before and 2 weeks after drug administration. As shown in Figure 1E, serum HPR1 activity was decreased in all 11 patients. The decrease of serum HPR1 levels was more profound in those with high serum HPR1 levels than those with low HPR1 levels. Statistical analysis revealed that the decrease of serum HPR1 activity after gemcitabine therapy was significant (P < 0.05). A decrease in HPR1 activity did not correlate with changes in serum CA19-9 levels in patients (data not shown). The in vitro HPR1 enzymatic assay revealed that gemcitabine up to 1.3 μm did not affect purified platelet HPR1 activity (Fig. 1F).
Functional Role of Serum HPR1 in Degrading HS
To confirm that the presence of HPR1 in serum from pancreatic carcinoma patients was enzymatically active, we tested whether HPR1 in patient's serum was able to degrade cell surface HS expression in MPANC-96 pancreatic tumor cell line. As shown in Figure 3, cell surface HS expression was abrogated by purified platelet HPR1 or by an HPR1-positive serum but not affected by an HPR1-negative serum. These observations further suggest that HPR1 secreted into the circulation from pancreatic adenocarcinomas was functional and capable of degrading cell surface HS and the basement membrane HS.
Elevated Serum HPR1 Levels are Associated with Shortened Patient Survival
We and others have previously demonstrated that HPR1 expression in pancreatic carcinomas correlates with patient survival.22–24 Hence, we tested whether serum HPR1 levels correlate with survival. Pancreatic carcinoma patients were divided into 2 groups according to a cutoff value of 300 units/mL of serum HPR1 activity measured by ELISA. Serum HPR1 levels in 12 of 31 (38%) pancreatic carcinoma patients were < 300 units/mL. This percentage is in agreement with the previously reported frequency of HPR1 expression in pancreatic adenocarcinomas.22–24 As shown in Figure 4, the mean survival of 12 pancreatic carcinoma patients with serum HPR1 activity < 300 units/mL was 13.3 ± 0.6 months (median survival, 10 mos; range, 6–28 mos), whereas the mean survival of 19 pancreatic carcinoma patients with serum HPR1 activity > 300 units/mL was 7.9 ± 0.2 months (median, 7 mos; range, 1–15 mos). Elevated serum HPR1 activity was associated with a poor overall survival in pancreatic carcinoma patients (P = 0.02), as shown by the log-rank test.
We next tested whether the difference of serum HPR1 levels in pancreatic carcinoma patients was due to the difference in their demographic and clinicopathologic parameters. As shown in Table 1, the Fisher exact test or chi-square test revealed that there was no significant difference with regard to the patients' age, gender, tumor stage, metastasis, differentiation, chemotherapy, and tumor resection between 12 patients with serum HPR1 levels < 300 units/mL and 19 patients with serum HPR1 levels > 300 units/mL. However, patients with serum HPR1 levels < 300 units/mL were found to survive significantly longer than those 19 patients with serum HPR1 levels > 300 units/mL. This suggests that patient survival predicted by serum HPR1 levels is not influenced by other clinicopathologic parameters.
Prior studies revealed that HPR1 expression is detected in a variety of gastrointestinal malignancies, including esophageal, gastric, and colon carcinoma.29–33 Using in situ hybridization, we previously detected HPR1 mRNA in 78% of primary pancreatic adenocarcinomas.23 Koliopanos et al.22 reported that HPR1 mRNA levels detected by real-time reverse transcriptase–polymerase chain reaction (RT-PCR) were elevated in 25 of 33 pancreatic carcinomas (83%) by an average of 30-fold, compared with that in normal pancreatic tissues. Rohloff et al.,24 using immunohistochemical staining, showed that HPR1 is expressed in 76% of pancreatic adenocarcinomas at moderate or high levels. All three studies consistently showed that HPR1 expression in pancreatic adenocarcinomas is associated with a short survival.22–24 The current study extends previous findings by demonstrating that serum HPR1 levels in pancreatic carcinoma patients were elevated compared with that in the healthy controls, and that higher serum HPR1 levels were associated with a shortened survival. A recent study by Kelly et al.34 did not find that serum HPR1 activity in patients with HPR1-positive myeloma was elevated, although HPR1 activity was elevated in bone marrow aspirates in 60% of multiple myeloma patients. It is not clear whether the discrepancy of serum HPR1 activity in pancreatic carcinoma patients and myeloma patients is due to the difference in tumor types or other factors. Nevertheless, to our knowledge, the current study is the first report demonstrating that serum HPR1 activity was elevated in patients with a malignant disease.
HPR1 is not expressed in epithelial cells but is expressed in many hematopoietic cells such as platelets,14, 17, 35, 36 neutrophils,37, 38 activated T-lymphocytes,39 and monocytes.40 It is likely that HPR1 present in the blood of healthy controls is mainly from hematopoietic cells. Although we could not rule out the possibility that immune cell activation in response to tumors also may contribute to increased serum HPR1 levels in pancreatic carcinoma patients, current study data demonstrated that serum HPR1 levels in patients with HPR1-positive tumors were significantly higher than in those with HPR1-negative tumors (Fig. 1D). This suggests that elevation of serum HPR1 activity in pancreatic carcinoma patients is due largely to the release of HPR1 from HPR1-expressing tumor cells. If this is the case, one may expect that serum HPR1 levels should correlate with the tumor burden in HPR1-positive tumors. Unfortunately, this is not clear from the results of the current present study because the number of patients with Stages I–III pancreatic carcinoma is small (only 5 patients), and HPR1 expression was analyzed only in 13 surgically removed specimens.
HPR1 plays an important role in tumor metastasis and angiogenesis. B16 murine melanoma and Eb lymphoma cell lines stably transfected with HPR1 become more metastatic,12 whereas suppression of HPR1 expression by siRNA or ribozyme in MDA-435 cells (a human breast carcinoma cell line) and B16 cells inhibits their invasive and metastatic potentials.41 HPR1 expression in multiple myeloma and hepatocellular carcinoma is reportedly correlated with elevated microvessel density.29, 34 It is likely that HPR1 expressed in pancreatic adenocarcinomas will promote angiogenesis and stimulate tumor growth, leading to a shorter survival of pancreatic carcinoma patients. The current study results suggest that HPR1 inhibitors such as PI-88, laminarin sulfate, and 4-alkyl-RK-486, which have been shown to be able to inhibit tumor cell invasion in vitro and metastasis in animal tumor xenograft models,26, 42, 43 can be developed as novel therapeutic agents in combination with gemcitabine to treat pancreatic carcinoma. Indeed, a previous study by Lapierre et al.44 demonstrated that a modified, non-anticoagulant heparin was able to inhibit the growth of CAPAN-2 pancreatic tumor cell lines implanted in nude mice. A recent study by Joyce et al.45 demonstrated that PI-88 is able to suppress tumorigenesis at the early stage and inhibit tumor growth at a later stage in the RIP1-Tag2 mouse model of pancreatic islet β-cell carcinoma caused by the SV40T antigen transgene expression in insulin-producing β-cells. The antitumor effect of PI-88 is mediated by induction of tumor cell apoptosis and inhibition of tumor cell proliferation, angiogenesis, and metastasis.45
Pancreatic carcinoma is a devastating disease, with a 5-year survival rate of < 5%. Approximately 80% of pancreatic carcinomas have already progressed to an advanced stage at the time of diagnosis.4 Most patients have an unresectable tumor.4, 46 Gemcitabine has recently emerged as a standard treatment for pancreatic carcinoma.28 The results of the current study demonstrated that serum HPR1 activity declined in all 11 patients during the second week of treatment. The decline in patients with high basal serum HPR1 levels was more profound than in those with low basal serum HPR1 levels. It is likely that the decline in serum HPR1 activity is due to its cytotoxic or cytostatic effect on pancreatic carcinoma cells. Serum HPR1 levels rebounded at the fourth week in some patients whose blood samples were available. Because the number of patients in this pilot drug study was small, we were unable to determine whether a decline in serum HPR1 activity leads to prolonged survival. Nevertheless, these results should prompt further study of the feasibility of using HPR1 levels as a marker to monitor the response to chemotherapy in pancreatic carcinoma and other malignancies in clinical trials. This is particularly true in those patients with HPR1-positive tumors and elevated serum HPR1 levels.
The results of the current study demonstrated that serum HPR1 levels were elevated in pancreatic carcinoma patients, serum HPR1 levels were significantly higher in patients with HPR1-positive tumors compared with those with HPR1-negative tumors, and gemcitabine treatment led to a decline in serum HPR1 activity in pancreatic carcinoma patients. The results further demonstrated that pancreatic carcinoma patients with high serum HPR1 levels had a significantly shorter survival. These observations suggest that serum HPR1 activity can be measured and used as a potential marker of prognosis and response to therapy.
The authors thank Dr. Robert L. Heinrikson (Pharmacia & Upjohn, Kalamazoo, MI) for kindly providing anti-HPR1 antibodies, and Progen Industries for kindly providing PI-88.