Gut Peptide Receptor Expression in Human Pancreatic Cancers

Abstract

Objective

To determine the prevalence of gastrointestinal (GI) peptide receptor expression in pancreatic cancers, and to further assess signaling mechanisms regulating neurotensin (NT)-mediated pancreatic cancer growth.

Summary Background Data

Pancreatic cancer remains one of the leading causes of GI cancer death; novel strategies for the early detection and treatment of these cancers is required. Previously, the authors have shown that NT, an important GI hormone, stimulates the proliferation of an NT receptor (NTR)-positive pancreatic cancer.

Methods

A total of 26 human pancreatic adenocarcinomas, obtained after resection, and 5 pancreatic cancer xenografts were analyzed for expression of NTR, vasoactive intestinal peptide receptor (VIPR), substance P receptor (SPR), and gastrin-releasing peptide receptor (GRPR). In addition, NTR expression, [Ca²⁺]_i mobilization, and growth in response to NT was determined in L3.6, a metastatic pancreatic cancer cell line.

Results

Neurotensin receptor was expressed in 88% of the surgical specimens examined and all five of the pancreatic cancer xenografts. In contrast, VIPR, SPR, and GRPR expression was detected in 31%, 27%, and 8% of pancreatic cancers examined, respectively. Expression of NTR, functionally coupled to the Ca²⁺ signaling pathway, was identified in L3.6 cells; treatment with NT (10 μmol/L) stimulated proliferation of these cells.

Conclusions

The authors demonstrated NTR expression in most of the pancreatic adenocarcinomas examined. In contrast, VIPR, SPR, and GRPR expression was detected in fewer of the pancreatic cancers. The expression of NTR and other peptide receptors suggests the potential role of endocrine manipulation in the treatment of these cancers. Further, the presence of GI receptors may provide for targeted chemotherapy or radiation therapy or in vivo scintigraphy for early detection.

Adenocarcinoma of the pancreas remains one of the leading causes of cancer death in the United States. ¹ In 1999, there were an estimated 29,000 new cases diagnosed and almost as many deaths from the disease. As with other gastrointestinal (GI) malignancies, our ability to intervene surgically in pancreatic cancer is predicated on its early detection. However, the lack of specific signs and symptoms of the disease and its relative infrequency make early detection by mass screening difficult and hamper efforts to identify subsets of patients at high risk for the disease. Even in cases detected early and treated with currently optimal therapy, the outcome remains poor compared with other malignancies such as colorectal and anal cancer. ¹ Although new chemotherapeutic agents for pancreatic cancer are being evaluated in clinical trials, ^2,3 more effective adjuvant therapies are needed in the treatment of this deadly disease, as well as novel techniques for its early detection.

In a manner analogous to breast and prostate cancers, which may possess peptide receptors and may be amenable to therapy with peptide receptor antagonists, ⁴ various GI cancers possess peptide receptors, and in experimental studies the proliferation of these receptor-positive cancers is increased with peptide treatment or inhibited by receptor blockade. ^5–7 For example, the hormone gastrin stimulates the growth of certain gastrin receptor-positive colon and gastric cancers both in vitro and in vivo. ⁷ Conversely, treatment of the gastrin receptor-positive murine colon cancer, MC-26, with a gastrin receptor antagonist, proglumide, or enprostil, an agent that blocks gastrin release, significantly inhibits tumor growth. ^7–9 Similarly, we ^10,11 and others ¹² have shown that the growth of cholecystokinin (CCK) receptor-positive pancreatic cancers is stimulated with CCK treatment and inhibited by CCK receptor blockade. Collectively, these studies have clearly shown that modulation of cancer growth can occur in various experimental models of receptor-positive GI cancers. In addition to the effects of peptides on cancer growth, other investigators have focused on the use of GI peptide receptors as a means of early detection and diagnosis of malignancies. Nuclear medicine techniques using radiolabeled vasoactive intestinal peptide (VIP), for example, have been studied for detection of colorectal and pancreatic cancer. ^13–15 The possibility of using gut peptide receptors as a focus for treatment or a means of early detection of pancreatic cancer is quite appealing. However, data are required regarding the expression patterns of GI peptide receptors in human pancreatic cancers so that specific receptor-based therapies or detection techniques can be developed.

Neurotensin (NT) is a tridecapeptide discovered in bovine hypothalami ¹⁶ and later shown in humans to be localized in the GI tract to enteroendocrine cells of the distal gut. ¹⁷ NT acts through its high-affinity NT receptor (NTR), a G-protein-coupled receptor of the seven transmembrane domain family of receptors, ^18,19 to regulate a variety of functions in the GI tract, including stimulation of pancreatic and biliary secretion, colonic motility, and growth of normal intestinal mucosa and pancreas. ^20,21 In addition to its effects on normal GI tissues, NTR stimulation has been shown to promote carcinogenesis in the colon of rats and to exert a trophic effect on human colon cancer cell lines both in vitro and in vivo. ^22–24 These effects can be blocked by a selective NTR antagonist, SR48692. ^25,26 Although not universally expressed in human colon cancer cell lines, NTR is expressed in a more metastatic colon cancer cell but is not expressed in normal human colon. ^27,28 Similarly, NT and its antagonist have been shown to regulate growth of the human pancreatic cancer cell line MIA PaCa-2, which expresses NTR. ^29,30 However, few data exist regarding the prevalence of NTR expression in human pancreatic cancers. Recent work by Reubi et al ³¹ suggests that NTR may be selectively expressed in human pancreatic adenocarcinomas and not in normal or inflamed pancreatic tissue.

In addition to NTR, other peptide receptors may be present in pancreatic cancers. For example, substance P receptors (SPRs) have been detected in the rat pancreatic acinar cell tumor line AR4 to 2J. ³² Further, VIP receptors (VIPRs) have been detected in pancreatic carcinomas by scintigraphy. ^15,33 In addition, gastrin-releasing peptide (GRP), or its amphibian equivalent bombesin, has been shown to stimulate the growth of normal pancreas in vivo and to modulate growth (either increase or in some instances decrease) pancreatic cancer growth in experimental models, thus raising the possibility of GRP receptors (GRPRs) on certain pancreatic cancers. ³⁴ Therefore, the purpose of our present study was to determine the prevalence of NTR, SPR, VIPR, and GRPR expression in human pancreatic cancers, and to further assess NT signaling pathways in a metastatic human pancreatic cancer cell line.

METHODS

Materials

Radioactive compounds were obtained from DuPont—New England Nuclear (Boston, MA). Tissue culture media and reagents were obtained from Gibco BRL (Grand Island, NY). Nitrocellulose filters for Northern blots were from Sartorius Corp. (Göttingen, Germany). The Ultraspec RNA isolation system was purchased from Biotech Laboratories (Houston, TX). NT peptide was purchased from Peninsula Laboratories (Belmont, CA). The NTR antagonist, SR48692, was a generous gift from Dr. Danielle Gully (Sanofi Recherche, Toulouse, France). Fura-2/acetoxymethyl ester was purchased from Molecular Probes (Eugene, OR). Reverse transcription—polymerase chain reaction (RT-PCR) reagents, including superscript II reverse transcriptase and DNA polymerase enzyme, were obtained from Gibco BRL. Random primers were obtained from Stratagene (La Jolla, CA). Deoxynucleotide triphosphates were purchased from Promega (Madison, WI). Oligonucleotide primers, based on previously described sequences for human NTR, VIPR, SPR and GRPR, ^27,35 were purchased from Oligos (Wilsonville, OR). GAPDH primers were obtained from Clontech Laboratories, Inc. (Palo Alto, CA). All other reagents, including agarose and ethidium bromide, were molecular biology-grade and obtained from Sigma Chemical Company (St. Louis, MO).

Tissue Procurement and Cell Culture

Human pancreatic adenocarcinomas (n = 26) were obtained from patients undergoing pancreatic resection at the M.D. Anderson Cancer Center, Houston, Texas. In addition, five human pancreatic cancers, initially resected at The University of Texas Medical Branch at Galveston and established as xenografts in male athymic nude mice (BALB/c; Life Science, St. Petersburg, FL), were analyzed. Tissue acquisition and subsequent use were approved by the institutional review boards at each institution. All specimens were snap-frozen in liquid nitrogen and stored at −70°C until RNA extraction. The human pancreatic adenocarcinoma cell line L3.6 was obtained from of Dr. Isaiah J. Fidler of M.D. Anderson Cancer Center. ³⁶ Cells were cultured in modified Eagle’s medium supplemented with 10% fetal bovine serum (FBS), l-glutamine (200 mmol/L), sodium pyruvate (100 mmol/L), vitamin mix (2×), and NEAA (1×) and maintained in a humidified atmosphere of 95% air and 5% CO₂ at 37°C.

RNA Extraction, Primer Design, and RT-PCR

RNA was extracted from tissues by the method of Chomczynski and Sacchi ³⁷ using the Ultraspec RNA system according to the manufacturer’s instructions. ³⁸

For RT-PCR, the sequences of the specific primers used for the RT reaction were as follows: human NTR, (sense) 5′-CGAAGCCGCACCAAGTT-3′ and (antisense) 5′-AGGATGGGGTTGATGGTGGAG-3′; human VIPR, (sense) 5′-GGGCTCGGTGGGCTGTAAGG-3′ and (antisense) 5′-GACCAGGGAGACTTCGGCTTG-3′; human SPR (sense) 5′-TGACCGCTACCACGAGCAAGTCTC-3′ and (antisense) 5′-ATAGTCGCCGGCGCTGATGAA-3′; human GRPR (sense) 5′-ATCTATGTCATCCCTGCAG-3′ and (antisense) 5′-TACATGTCAAGAAGCAGATC-3′.

The reverse transcriptase reaction was performed using 5 μg total RNA and 100 ng random primers diluted to a volume of 12 μL. Each mixture was heated to 70°C for 10 minutes, chilled on ice, and added to 4 μL 5× First Strand Buffer, 2 μL 0.1 mol/L DTT, and 1 μL 10 mmol/L dNTP mix. Contents were mixed and incubated at 42°C for 2 minutes. After incubation, 1 μL superscript II enzyme was added to each reaction, and each was incubated at 42°C for 50 minutes, then heat-activated at 70°C for 15 minutes. Parallel reactions were performed without RNA to ensure lack of contamination by genomic DNA. After confirmation of cDNA product by gel electrophoresis, 1 μL cDNA template was used for PCR. Reaction mixtures contained 1 μL cDNA, 5 mL 10× PCR buffer, 3 μL 25 mmol/L MgCl₂, 1 μL 10 mmol/L dNTP mix, and 100 pmol sense and antisense primers. Autoclaved-distilled H₂O was added to a total volume of 49 μL. Reactions were mixed and covered with sterile silicone oil, and PCR was initiated with 95°C hot-start addition of the Taq DNA polymerase and carried out according to the following profile: an initial denaturation for 1 minute at 95°C followed by 40 cycles at 94°C for 1 minute, 63°C for 1 minute, and 72°C for 7 minutes. An aliquot of each PCR reaction was resolved by gel electrophoresis with 1% agarose gel containing ethidium bromide. Bands were visualized by ultraviolet illumination.

Northern Blot Analysis

For Northern blot analysis, total RNA (50 μg) was electrophoresed in 1.2% agarose-formaldehyde gels, transferred to nitrocellulose, and hybridized with a labeled cDNA probe to human NTR. ²⁸ Hybridization and washing conditions were described previously. ³⁹

Intracellular Calcium ([Ca²⁺]_i) Ratio Imaging

Real-time recording of [Ca²⁺] was performed in cultured cells using methods described previously. ^28,40 In brief, cells grown on glass coverslips for 48 hours were washed with Krebs-Ringer’s-Henseleit (KRH) buffer and loaded with 2 μmol/L fura-2 for 50 minutes at 25°C to minimize dye compartmentalization. Loaded cells were washed three times with KRH and incubated for 60 minutes at 25°C in the dark in KRH with 0.1% bovine serum albumin (BSA), mounted on a Leiden Cover Slip Dish, and placed in an Open Perfusion Micro-Incubator (Medical Systems Corp., NY) covered in 3 mL KRH with 0.1% BSA. Cells were imaged using a Nikon Diaphot inverted microscope (Garden City, NY), coupled to a dual monochrometer system, and NT-mediated [Ca²⁺]_i mobilization was calculated by the method of Grynkiewicz et al. ⁴⁰ Fluorescence was detected using an intensified CCD camera (Dage-MTI, Inc., Michigan City, IN), and images were processed with ImageMaster software (Photon Technologies International, Inc., S. Brunswick, NJ).

MTT Cell Growth Determination Assay

Growth was assayed using the standard MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) assay, according to the supplier’s instructions, which colorimetrically measures the accumulation of a formazan dye generated by the conversion of MTT by active mitochondrial dehydrogenases in living cells. In brief, L3.6 cells (10⁵) were plated in 96-well plates for 24 hours. After this time, the medium was removed and replaced with fresh medium containing 1% FBS and various concentrations of NT (day 0). The medium was changed daily and NT treatments were repeated. On days 2, 4, or 6, 10 μL MTT reagent (5 mg/mL) in phosphate-buffered saline was added to each well. The plates were incubated for 2 hours at 37°C, 5% CO₂. The medium was gently aspirated, and the resulting formazan was dissolved in 100 μL of acidic (0.1 N HCl) isopropanol for 5 minutes at room temperature. Sample absorbance at 570 nm was measured using a spectrophotometer.

Statistical Analysis

The data on cell growth were analyzed using analysis of variance for a two-factor factorial experiment. The two factors were day (4 and 6 days) and dosage of NT. The Fisher least significant difference procedure was used for multiple comparisons. P < .05 was considered significant.

RESULTS

Expression of GI Peptide Receptors on Human Pancreatic Cancers

The presence of NTR has been previously demonstrated in pancreatic cancer cell lines. ^29,41 The purpose of the first part of the study was to assess the number of pancreatic adenocarcinomas that actually express NTR. In addition, we determined the expression of three other GI peptide receptors, VIPR, SPR, and GRPR. For all these studies, we used a sensitive RT-PCR procedure for detection of receptor expression.

In our initial series of 12 pancreatic adenocarcinomas, we found that NTR expression (523 bp) occurred in all 12 samples (Fig. 1). Expression of the expected PCR product for VIPR (754 bp), SPR (295 bp), and GRPR (710 bp) was noted in 8, 5, and 2 of the 12 specimens, respectively. Additional bands were noted in the VIPR and SPR reactions that may have represented splice variants of the receptors. To confirm that the RNA was not degraded, all samples were initially run on an agarose gel to confirm intact 28S and 18S ribosomal bands. In addition, a separate PCR reaction, using primers for the constitutively expressed GAPDH gene, was performed and demonstrated the expected PCR product for GAPDH (450 bp) in all samples (see Fig. 1, bottom panel).

Figure 1. Gut peptide receptor expression in human pancreatic cancers. RNA was extracted from pancreatic cancer specimens resected from patients (n = 12), and reverse transcriptase—polymerase chain reaction (RT-PCR) was performed to determine expression of neurotensin receptor (NTR), vasoactive intestinal peptide receptor (VIPR), substance P receptor (SPR), and gastrin-releasing peptide receptor (GRPR). To confirm integrity of the RT reaction, primers for the constitutively active gene GAPDH were used for amplification (lower panel).

Next, an additional 14 pancreatic cancers were analyzed. In this second group, NTR and SPR expression was noted in 11 and 2 of the 14 cancers, respectively; however, neither VIPR nor GRPR expression was detected in any of the 14 tumors (data not shown).

In addition to primary pancreatic cancer tissues, expression of NTR, VIPR, SPR, and GRPR was also analyzed in five human pancreatic cancer xenografts established in athymic nude mice in our laboratory. All five of the pancreatic cancer xenografts expressed NTR at varying levels; expression of VIPR was noted in four of the five cancer xenografts, SPR in three of the five, and GRPR in none (Fig. 2). To confirm the integrity of the RT reactions, each reaction was amplified with primers for GAPDH (see Fig. 2, bottom panel).

Figure 2. Gut peptide receptor expression in human pancreatic cancer xenografts. RNA was extracted from five pancreatic cancer xenografts established in athymic nude mice from resected pancreatic cancers. Reverse transcriptase—polymerase chain reaction (RT-PCR) was performed to detect expression of neurotensin receptor (NTR), vasoactive intestinal peptide receptor (VIPR), substance P receptor (SPR), and gastrin-releasing peptide receptor (GRPR). In addition, the integrity of the RT reaction was confirmed by amplification using primers to the constitutively expressed GAPDH gene.

A summary of the expression of these GI peptide receptors in the surgically resected samples and the tumor xenografts is shown in Table 1. NTR expression was noted in 88% of the resected surgical samples and 100% of the pancreatic cancer xenografts. Expression of the VIPR and SPR was noted in 31% and 27% of the surgically resected pancreatic carcinomas, respectively. GRPR expression was noted in only 8% of the resected cancers and in none of the xenografts. Therefore, our findings demonstrate that most pancreatic adenocarcinomas express NTR, with expression of the other peptide receptors occurring more infrequently.

Table 1. GASTROINTESTINAL PEPTIDE RECEPTOR EXPRESSION FROM SURGICAL SAMPLES AND PANCREATIC CANCER XENOGRAFTS

NTR mRNA Is Expressed in the L3.6 Pancreatic Cancer Cell Line

With these findings of NTR expression in most pancreatic cancers, we were interested in identifying a pancreatic cancer cell line that both expresses NTR and would represent a clinically relevant model that could be used for hormonal manipulation studies both in vitro and in vivo. Therefore, the L3.6 pancreatic cancer cell line, a highly metastatic variant derived from COLO 357 human pancreatic cancer cells, ³⁶ was analyzed for NTR expression by Northern blot using a human NTR cDNA; RNA from the HT29 colon cancer cell line was used as a positive control (Fig. 3). NTR mRNA was expressed in L3.6 cells, albeit at lower levels than in the HT29 positive control.

Figure 3. Neurotensin receptor (NTR) is expressed in L3.6 pancreatic cancer cells. Northern blot analysis of total RNA (50 μg) from L3.6 pancreatic cancer cells. RNA extracted from the HT29 cell line was used as positive control for NTR expression. Blots were probed with a ³²P-labeled human NTR cDNA probe.

NTR Is Functionally Coupled to the Ca²⁺ Pathway in L3.

6 Cells

To confirm that the endogenous NTR in L3.6 cells was functionally coupled to the Ca²⁺ signaling pathway, cells were loaded with fura-2 dye and fluorescence was measured after administration of NT (50 nM) (Fig. 4). A rapid increase in [Ca²⁺]_i was noted with NT treatment, as shown in the pseudocolor images in Figure 4A and in the graph in Figure 4B, demonstrating a sharp spike of [Ca²⁺]_i mobilization immediately after NT treatment. These cells were then treated with acetylcholine (100 μmol/L), a positive control, to demonstrate [Ca²⁺]_i mobilization using a different agent. In contrast to our previous findings with the MIA PaCa-2 pancreatic cancer cell line, in which [Ca²⁺]_i mobilization occurs in the entire cell population, ²⁹ we noted that approximately 50% of L3.6 cells responded to NT treatment, suggesting that a selective population of L3.6 cells expresses NTR. The experiment was repeated using the NTR antagonist SR48692 (see Fig. 4C). Pretreatment with SR48692 prevented the stimulation of [Ca²⁺]_i, indicating that the effect of NT was through its specific NTR.

Figure 4. Intracellular calcium ([Ca²⁺]_i) mobilization in L3.6 pancreatic cancer cells. (A) Pseudocolor images representing the relative [Ca²⁺]_i concentration as demonstrated by the color bar (left). Times listed are after treatment with neurotensin (NT; 50 nM). (B) Graph depicting [Ca²⁺]_i concentration over a time course after NT treatment (50 nM). Each line represents a single cell measurement. To ensure the ability of cells to reproduce a calcium response, NT treatment was followed with acetylcholine (Ach) treatment (100 mmol/L). (C) Graph depicting [Ca²⁺]_i concentration in L3.6 cells pretreated with SR48692 (1 mmol/L) before NT treatment (50 nM). To verify the specificity of inhibition of the NT receptor-mediated calcium response, cells were treated with Ach (100 mmol/L).

The dosage of NT required for a half-maximal response (EC₅₀) was next determined. As shown in Figure 5A, [Ca²⁺]_i was increased in a dose-dependent manner with a calculated EC₅₀ of 36.7 nM. Conversely, dose-response curves calculated for the antagonistic effects of SR48692 on NT agonist stimulation gave an IC₅₀ of 9.0 nM. Therefore, these studies confirm that NT is acting through its specific receptor, which is functionally coupled to the Ca²⁺ signaling pathway. In addition, the higher EC₅₀ value of 36 nM compared with MIA PaCa-2 cells ⁴¹ is consistent with the Ca²⁺ imaging studies demonstrating a response in approximately half of the cell population.

Figure 5. Dose-response for neurotensin receptor (NTR) stimulation and inhibition in L3.6 pancreatic cancer cells. (A) Graph depicting maximal increase in [Ca²⁺]_i in spectrum response to increasing doses of neurotensin (NT). Each data point represents the mean of multiple cell measurements ± standard error. (B) Graph depicting maximal increase in [Ca²⁺]_i in cells pretreated with increasing doses of SR48692 followed by NT (50 nM). Data are expressed as percentage of increase compared with control cells treated with NT (50 nM) alone. Each data point represents the mean of multiple cell measurements ± standard error.

NT Treatment Stimulates Proliferation of L3.6 Cells In Vitro

To determine whether NT exerts a proliferative effect in L3.6 cells, we used an MTT cell growth determination assay. L3.6 cells were treated with various concentrations of NT beginning on day 0; cells were then harvested on days 2, 4, and 6 and MTT assays performed. A significant increase in MTT absorbance was noted on days 4 and 6 after initiating treatment with NT (10 μmol/L) (Fig. 6).

Figure 6. Neurotensin (NT) treatment increases growth of L3.6 pancreatic cancer cells. Graph depicting MTT absorbance at 570 nm in L3.6 cells treated with increasing doses of NT. MTT assays were performed on days 4 and 6. Data are expressed as mean ± standard error. (*P < .05 vs. control [CON]).

Taken together, the demonstration of NT-mediated [Ca²⁺]_i mobilization and growth in this NTR-positive pancreatic cancer cell line suggests that L3.6 cells may be a useful model to assess the effects of NT hormone manipulation both in vitro an in vivo. The propensity of hepatic metastases with in vivo injection of L3.6 cells further enhances the clinical applicability of this cell line for the assessment of potential therapeutic agents.

DISCUSSION

It is firmly established that hormones can affect the growth of certain human malignancies. ⁶ Similar to breast and prostate cancers, which are the best described of the hormonally responsive cancers, ⁴ certain GI cancers possess peptide receptors and are hormone-responsive and their growth is inhibited by peptide deprivation, the antagonism of peptide action, or the administration of inhibitory hormones. ^5,42 We have previously shown that the human pancreatic cancer cell line MIA PaCa-2 possesses high-affinity NTR coupled to multiple intracellular signaling pathways, including Ca²⁺²⁹; treatment of MIA PaCa-2 with NT results in increased cell growth both in vitro and in vivo. 29,30,41,43 In the present study, we examined the expression of NTR in a large series of resected pancreatic adenocarcinomas as well as pancreatic cancer xenografts established in athymic nude mice. In addition to NTR, we assessed other putative peptide receptors (VIPR, SPR, GRPR).

Using a sensitive RT-PCR procedure, we detected NTR expression in approximately 88% of the resected human pancreatic adenocarcinomas and in all five of the pancreatic cancer xenografts established from patient tumors. Consistent with our results, Reubi et al, ³¹ using an ex vivo scintigraphic technique, demonstrated the presence of NTR in 75% of 24 pancreatic adenocarcinomas. Therefore, our study, as well as that of Reubi et al, ³¹ conclusively demonstrates the presence of NTR expression in most pancreatic adenocarcinomas. These results are also consistent with other reports of NTR expression in pancreatic cancer cell lines (e.g., MIA PaCa-2 and PANC-1). ⁴¹ Although we did not specifically assess noncancerous pancreatic tissues, Reubi et al ³¹ analyzed samples of chronic pancreatitis as well as normal pancreatic glands and did not detect NTR in these tissues. Further, endocrine pancreatic tumors also lacked NTR expression despite the presence of high-density somatostatin receptors. ³¹

These findings are potentially exciting because by all accounts the expression of NTR appears to be specific for pancreatic adenocarcinomas. This could have significant clinical applications for cancer detection and treatment. However, the absence of NTR in normal pancreas must be confirmed, because NT is known to affect various pancreatic functions, and therefore it would seem likely that normal pancreatic tissues may express NTR. Further, our preliminary studies have demonstrated NTR expression in the pancreas of rats of different ages, as assessed by PCR techniques (B.M. Evers, unpublished observations). Future studies using sensitive PCR analyses are required to assess NTR expression in the normal human pancreas.

In addition to NTR expression, we also assessed expression of VIPR, SPR, and GRPR. The expression of both VIPR and SPR occurred in approximately 25% of the surgically resected cancers. Only 8% of the pancreatic cancers demonstrated expression of GRPR. This was surprising because both GRP and its amphibian equivalent, bombesin, have been shown to stimulate the growth of normal pancreas as well as to modulate proliferation of various pancreatic cancers. ³⁴ One possible explanation for this discrepancy is the fact that GRP can stimulate the release of all GI hormones, except for secretin, as well as stimulating pancreatic and intestinal secretions ⁴⁴; therefore, the well-characterized in vivo effects of GRP on pancreatic growth may be secondary to indirect effects as a result of stimulation of other hormones, which then act directly on the pancreas.

Our study evaluating the expression pattern of various GI peptide receptors in pancreatic cancers has potentially important clinical applications. For example, as suggested by Reubi et al, ³¹ if NTR expression is confirmed to occur only in pancreatic adenocarcinomas, then scintigraphic techniques could provide a useful means for early detection and possible immunomodulation of pancreatic cancer. For example, nuclear medicine scintigraphic scans are currently in use to detect endocrine tumors based on the fact that many of these tumors possess somatostatin receptors. ^45,46 In addition, using nuclear medicine scintigraphic techniques, radiolabeled VIP has been used in patients with colon and pancreatic cancers to detect the presence of VIPR. ^14,15 Although our data show only a 31% prevalence of VIPR expression in resected pancreatic adenocarcinoma, others have shown this technique to provide more than 90% sensitivity in detecting primary and recurrent metastatic disease. ¹⁵ Therefore, receptor scintigraphy using radiolabeled NT or VIP may be particularly useful in the detection of isolated primary or metastatic pancreatic cancers.

Another area of active interest is the potential use of receptor antagonists as adjuvant chemotherapeutic agents. Analogous to breast and prostate cancer, patients with GI malignancies expressing peptide receptors may be candidates for therapeutic agents that either block peptide binding or inhibit peptide release. For example, we have shown that NT stimulates the growth of MIA PaCa-2 cells both in vivo and in vitro. ³⁰ Further, we found that the NTR antagonist SR48692 blocked the NT-mediated trophic effect on MIA PaCa-2 pancreatic cancer xenografts in vivo, but we could not demonstrate an inhibitory effect on MIA PaCa-2 growth using SR48692 alone. In contrast, other investigators have reported an inhibitory effect of SR48692 alone on both NTR-positive colon and pancreatic cancers. ^26,41 If confirmed in subsequent studies using other GI cancers bearing NTR, then antagonists such as SR48692 may prove useful as adjuvant therapy in certain NTR-positive cancers. Therefore, clinically relevant models of pancreatic cancer are required to address these questions.

In an effort to identify a useful cell model, we have determined NTR expression in the human metastatic pancreatic cancer cell line L3.6. This cell line represents a potentially useful and clinically relevant model of human pancreatic cancer because the injection of cancer cells into the pancreatic capsule or spleen of athymic nude mice produces hepatic metastases in most of the animals. ³⁶ We found that similar to MIA PaCa-2 cells, L3.6 cells express NTR. NTR was functionally coupled to the Ca²⁺ pathway, as noted by a rapid increase in [Ca²⁺]_i with NT treatment; this effect was blocked by pretreatment with the NTR antagonist SR48692. However, in contrast to MIA PaCa-2, in which the great majority of the cells possess NTR and respond to NT stimulation, only approximately 50% of the L3.6 cells in our study responded to NT stimulation, as detected by fura-2 spectrofluorometry. This finding is also reflected in the higher EC₅₀ value for L3.6 compared with reported EC₅₀ values of MIA PaCa-2 and PANC-1. ^29,41

Because L3.6 cells possess functional NTR, we next determined whether NT treatment stimulates proliferation of these cells using an MTT assay. NT treatment (10 μmol/L) resulted in a 15% to 20% increase in MTT absorbance on days 4 and 6 after initiating treatment. The timing and percentage increase in cell proliferation are consistent with previous findings from our laboratory ²⁹ as well as those of others ⁴¹ using NT or other peptides to assess cell growth. In addition, the fact that a higher dose of NT was required to affect cell growth compared with MIA PaCa-2 is also consistent with the higher EC₅₀ values noted for L3.6 cells. The presence of NTR in these unique pancreatic cancer cells suggests that this cell line may serve as a useful model for future in vitro and in vivo studies to assess NT-mediated signaling pathways and effects on proliferation. Further, this cell line would be a useful model to determine whether pancreatic tumor growth or metastasis can be inhibited by treatment with NTR antagonists.

In conclusion, our study demonstrates expression of NTR in a majority of human pancreatic adenocarcinomas. We also showed that VIPR and SPR were expressed in slightly more than 25% of these samples and GRPR was expressed in only 8% of the cancers. Finally, we identified expression of a functional NTR in a novel pancreatic cancer cell line, L3.6, which may serve as a useful and clinically relevant model to delineate the role of NT on pancreatic cancer cell growth. By assaying tumors for the presence of peptide receptors and further analyzing the effects of these agents on tumor growth, we will be able to determine more effectively the role that these hormones play in regulating pancreatic cancer cell growth. This understanding may lead to the development of more effective therapeutic strategies for patients based on receptor blockade or the alteration of endogenous hormone levels. Certainly, no one agent alone will be effective in treating patients with pancreatic cancers, but it is hoped that by using multiple agents that block different receptors or signal transduction pathways, we can increase the therapeutic efficacy and decrease the overall toxicity.

Acknowledgments

The authors thank Kirk Ives, Jing Li, and Jell Hsieh for technical assistance, Tatsuo Uchida for statistical analysis, and Eileen Figueroa and Karen Martin for manuscript preparation.

Discussion

Dr. John B. Hanks (Charlottesville, Virginia): Dr. Evers and his group have just reported their latest in a series of beautifully conducted efforts which have evaluated and identified GI hormone receptor resistance and activity in oncogenic processes at various GI sites. This study demonstrates neurotensin receptor activity in pancreatic adenocarcinoma. Additionally, VIP, substance P, and gastrin-releasing peptide receptor activity is demonstrated, albeit in the minority of those cases.

These and other studies by Evers and his group at Galveston raise the important question of applicability and feasibility of further studies in this area which might identify, localize, and possibly retard tumor growth and metastatic activity.

I have several brief questions about this fascinating work. My first may be out of my own ignorance in cell line analysis, but it allows the authors to expand on the clinical applicability of this important concept. Does the presence of the mRNA in your studies that you found from resected human pancreas correspond to the expression of the functional receptor in these patients? In other words, using your derived L3.6 cells, you confirm that the administration of neurotensin stimulated the calcium signaling pathways, but could you expand your comments on this in the human resected pancreas samples you obtained? To what extent does your derived data from L3.6 predict neurotensin functional studies in the human pancreas tissue?

Second, often pancreatic adenocarcinoma is comprised of a large number of nontumorous cells, including stroma and mesenchymal cells. Do such surrounding cells or cells involved in the desmoplastic reaction in these GI malignancies show receptors and their functional activity?

Third, you mention in your manuscript that your study failed to show neurotensin receptor activity in normal pancreatic tissue. Why is that, in the face of several proposed effects of neurotensin on the various pancreatic functions?

Finally, does the absolute level or the finding of any level of neurotensin receptor expression correlate with tumor stage, size, or clinical outcomes?

Dr. James V. Sitzmann (Rochester, New York): Dr. Evers and his group, including Dr. Townsend, are to be congratulated on this outstanding work, which is contributing to our knowledge of the progression of cancer. As we all know, the accepted theory of cancer development is based on the Knudson hypothesis, that a series of genetic alterations, whether congenital or environmentally induced, will lead to a cancer phenotype.

A distinct question raised by this hypothesis is how a cancer cell relates to its environment once mutated. This is why Dr. Evers’ and Dr. Townsend’s work is so important, as it focuses us on how the cancer cell can be regulated or controlled through external stimuli.

This paper reports a novel signal transduction system to the neurotensin receptor which is present on pancreatic cancer cells, but not on normal pancreatic cells. This observation leads to several questions.

If the neurotensin receptor is present on 100% of xenografted cells and 88% of harvested pancreatic cancers and not on normal pancreas, does this imply that the neurotensin receptor is a marker for the degree of differentiation of the tumor or, further, for the tumor phenotype?

Also, do we know that the neurotensin receptor binds only neurotensin, or can other ligands bind to this receptor?

The work also notes that the neurotensin receptor is a classic G-protein coupled receptor. Following binding of a ligand, a G-protein normally will disassociate to its alpha and beta gamma subunits. It has been shown in several cell systems, especially in liver regeneration and others, that the beta gamma subunit can associate with Ras proteins to stimulate the meg-kinase pathway. The meg-kinase pathway is a critical growth regulatory path for most cells. This pathway is of special significance for pancreatic cancers as mutations in the Ras protein, the so-called Kirsten Ras or K-Ras mutation, has been identified as being present in over 80% to 85% of pancreatic cancers derived from humans. The question begged by Dr. Evers’ work is whether the neurotensin receptors associated with Ras proteins were the meg-kinase activation. Have you had a chance to survey your cells for associated Ras oncogene expression?

Dr. Benjamin Li (Shreveport, Louisiana): Modelled after hormone receptor analysis in prostate and breast cancer, Dr. Evers’ and Dr. Townsend’s group examined the expression of gut peptide receptor for the potential of exploiting these receptors in GI malignancy for diagnostic and therapeutic manipulations. In this innovative study, neurotensin receptor expression as detected by RT-PCR for the NTR-mRNA was studied in human and animal carcinoma of the pancreas, and they reported that greater than 80% of this specimen had detectable NTR expression.

As their elegant electrogelphoresis photographs show, NTR expression was actually at various degrees of overexpression relative to the GAPDH constitutive protein.

I, like Dr. Hanks, am extremely interested in seeing whether this variable degree of overexpression translates to TNM staging, presence of metastases and, ultimately, clinical outcome.

The second question is that in breast cancer, various receptors have been studied for their role as prognostic and therapeutic markers, for example, estrogen receptor or HER-2/neu receptor. These studies suggest that the presence of the end product, i.e., the peptide receptor, rather than the mRNA may be more feasible to be measured. I would like very much for the authors to comment on why they chose to study mRNA level, a potentially more difficult entity to maintain the stability and to measure than the NTR receptor itself, rather than the NTR protein.

And finally, the potential implications for diagnostic imaging using an NTR receptor-specific radiolabeled molecule or an NTR antagonist as an adjuvant therapy are exciting. I would like the authors to comment on whether they have studied the degree of NT receptor expression in histologically normal tissue of patients with pancreatic cancer, to exploit the specificity of the NT receptor expression in cancer alone and to avoid background targeting of normal tissue.

Dr. B. Mark Evers (Closing Discussion): I thank all of the discussants for their questions, and I will try to group them in the interests of time. I think that all of the discussants have asked similar questions about clinical correlations, that is, does receptor expression correlate with tumor size, and how does this correlate with patient outcome? We are currently going back to hospital records to determine whether there is a clinical correlation related to the receptor expression.

Dr. Hanks asked about the L3.6 cell line that we have used as a human pancreatic cancer cell model in our studies. We have used this cell line in deference to some of the other human pancreatic cancer cell lines because the L3.6 cells serve as a useful in vitro and in vivo model. These cells, when injected into the pancreatic capsule of nude mice,metastasize to the liver routinely. Therefore, this cell line serves as a useful model to test potential therapeutic agents because studies can be carried out in vitro, and then, if effective, can be used in vivo in a clinically relevant model of metastatic pancreatic cancer. An excellent question regarding neurotensin receptor expression in stromal cells was raised. This is something that we need to assess because the surrounding cells may be quite important in terms of response and release of growth factors. Although other investigators have looked at the normal pancreas for neurotensin receptor expression, we have not specifically done this study. However, Reubi et al published a study in Gut in which they looked at neurotensin receptor expression by autoradiography and found, similar to our results, approximately 75% of cancers expressed the neurotensin receptor. In addition, they looked at 20 samples of chronic pancreatitis, as well as endocrine tumors and normal pancreas, and did not note receptor expression. We plan on evaluating normal pancreas by the more sensitive RT-PCR procedure to determine whether we detect similar results.

Dr. Sitzmann asked whether the neurotensin receptors were associated with Ras oncogene expression. We have not specifically looked at Ras expression in our pancreatic cancers. However, we do know that the Ras pathway is coupled to the MAPK and JNK pathways, which we have previously shown in both colonic and pancreatic cancers with neurotensin receptors. Neurotensin stimulates both of these pathways, which would be consistent with its effect as a growth factor.

Dr. Li asked why we chose mRNA levels for our assessment. This is what is usually done in our laboratory, and we feel that these levels provide a useful screen to rapidly assess whether these tumors express the receptor gene in question. We have not, as yet, evaluated overexpression of the neurotensin receptor in these cancers.