Background
Vitamin A-containing cells were first reported in 1982 by Watari et al. in vitamin A loaded mice using fluorescence and electron microscopy [
1]. This cell type was subsequently identified by electron microscopy in normal rat and human pancreatic tissues [
2]. These cells were identified as pancreatic stellate cells (PSCs) by Apte et al and Bachem et al in 1998 [
3,
4]. In the normal pancreas, stellate cells are quiescent and can be identified by the presence of vitamin A-containing lipid droplets in the cytoplasm. In response to pancreatic injury or inflammation, PSCs are transformed ("activated") from quiescent phenotypes into highly proliferative myofibroblast-like cells which express the cytoskeletal protein α-smooth muscle actin (α-SMA), and produce type I collagen and other extracellular matrix components. Many of the morphological and metabolic changes associated with the activation of PSCs in animal models of fibrosis also occur when these cells are cultured on plastic in serum-containing medium.
Activated PSCs have also been implicated in the deposition of extracellular matrix components in pancreatic adenocarcinoma [
5]. In patients with pancreatic cancer, an intense, interstitial, fibrillar staining for PSCs is evident in the peritumoral fibrous regions. Procollagen I staining colocalized with α-SMA to these fibroblast-shaped cells suggests that they are responsible for the deposition of matrix components and the desmoplastic reaction that surrounds the pancreatic tumor [
5].
Cyclooxygenases (COXs) are key rate-limiting enzymes involved in the conversion of arachidonic acid to prostaglandin (PG) H2, the precursor of a variety of compounds including PGs, prostacyclin, and thromboxanes. Two isozymes are found in mammalian tissues, COX-1 and COX-2. COX-1 is expressed constitutively in a wide variety of tissues, where it is involved in the maintenance of tissue homeostasis. In contrast, COX-2, which is not expressed in resting cells, is the inducible form of the enzyme responsible for PG production at sites of inflammation. Growth factors, cytokines, tumor promoters, and other inflammatory mediators can induce COX-2 expression [
6,
7]. COX-2 expression and activity is up-regulated in pancreatic cancer, but absent in normal pancreatic acinar and duct cells [
8‐
10]. Some scattered cells in normal pancreatic tissues express COX-2 [
11,
12].
The current study revealed that COX-2 is expressed in primary cultured PSC. Furthermore, conditioned media from pancreatic cancer stimulates PSC proliferation and COX-2 expression. The increase in PSC proliferation in response to conditioned media is prevented by inhibition of COX-2.
Discussion
There is accumulating evidence that PSCs play a role in the development of pancreatic fibrosis [
13,
14]. Little is known regarding the relationship between PSCs and pancreatic cancer, or the role of COX-2.
The present study revealed that PSCs express COX-2 constitutively and when activated. The two isoforms of COX, COX-1 and COX-2, differ in many respects. COX-1 is a housekeeping gene that is expressed in most tissue, while COX-2 is not detected in most normal tissues. In the pancreas, islet cells display a strong expression of COX-2 [
9]; however, some scattered basal cells in normal pancreas express COX-2 as well, though less than seen in islet cells [
11,
12]. In hepatic stellate cells (HSCs) which are similar to PSCs, COX-2 expression is virtually undetectable by Western blot analysis in protein extracts obtained from freshly isolated HSC [
15]. However, serum-deprived unstimulated HSC express low levels of the COX-2 protein and expression is dramatically enhanced in response to IL-1α, TNF α or endothelin-1 [
15,
16]. The present study suggests that COX-2 expression is independent of the activation status in isolated PSCs. While there is no marked expression of COX-2 in desmoplastic areas of pancreatic cancers [
8‐
10], it is possible that the enzyme is up-regulated early in the activation of stellate cells
in vivo but increased expression may not be required for maintenance of stellate function once activated.
Stimulation of PSC by PANC-1 CM increased the expression of COX-2. Oncogenes, growth factors, cytokines, chemotherapy and tumor promoters stimulate COX-2 transcription via protein kinase C and RAS-mediated signaling. Stimulation of either protein kinase C or RAS-mediated signaling enhances mitogen-activated protein (MAP) kinase activity, which in turn, activates transcription of COX-2 [
17]. We have previously reported that PANC-1 CM enhances ERK 1/2 activation and growth of PSCs [
18]. We speculate that a growth factor is responsible for these effects, however, our attempts to identify the candidate using receptor antagonists and immunoneutralization have not been successful. Inhibition of ERK1/2 phosphorylation by U0126 prevented the PANC-1 CM-stimulated increase in PSC COX-2 protein production. In previous studies U0126 alone had no effect on ERK1/2 or COX-2 expression [
19,
20]. This suggests that the MAP kinase pathway plays a role in cancer-induced stimulation of COX-2 in PSCs. The reported biological consequences of COX-2 up-regulation include growth stimulation inhibition of apoptosis [
21], increased metastatic potential [
22] and promotion of angiogenesis [
23]. Increased expression of COX-2 in PSCs by PANC-1 CM may contribute to tumor progression.
Finally, the proliferation of PSCs was inhibited by treatment with NS398, a COX-2 inhibitor. In pancreatic carcinomas, COX-2 is overexpressed and NS398 inhibits tumor growth [
8,
9,
24]. This COX-2 inhibitor alone has no effect on expression of COX-2 or ERK1/2 and shows no toxicity at the concentration used in the present studies [
20,
24]. Recent studies have demonstrated a role for the COX-2 enzyme and PGE2 in the regulation of epithelial cell growth and angiogenesis [
23,
25,
26]. These properties will need to be studied further in pancreatic adenocarcinoma and stellate cells. NS-398 has been previously shown to inhibit cell proliferation of colorectal carcinoma by inducing apoptosis in a COX-2-independent fashion [
27]. More studies are needed to confirm the mechanism of inhibition by NS398.
Materials and methods
Materials
Iscove's modified Dulbecco's medium (IMDM), Dulbecco's modified Eagle's medium (DMEM), albumin, and pronase were purchased from Sigma Chemical (St Louis, MO). Fetal bovine serum (FBS), glutamine, and antibiotics were purchased from Mediatech, Inc. (Herndon, VA). Collagenase P was purchased from the Roche Diagnostics Corporation (Indianapolis, IN), and deoxyribonuclease from Amersham Biosciences (Piscataway, NJ). Nycodenz was obtained from Nycomed Pharma AS (Oslo, Norway). U0126, a specific inhibitor of extracellular signal-regulated kinase (ERK) activation, was obtained from Calbiochem (San Diego, CA). NS398, COX-2 inhibitor was obtained from Cayman Chemicals (Ann Arbor, MI). 3H-methyl thymidine was purchased from ICN Pharmaceuticals (Costa Mesa, CA).
Animals
Male Sprague-Dawley rats weighing 200–250 g were used in accordance with standard institutional animal welfare guidelines, and protocols were approved by the Institutional Animal Care and Use Committee, Northwestern University School of Medicine.
Isolation and culture of PSCs
Rat PSCs were isolated as previously described [
3]. Briefly, the pancreas was digested with a mixture of collagenase P and pronase and deoxyribonuclease in Gey's balanced salt solution. The resulting suspension of cells was centrifuged in a 28.7% Nycodenz gradient at 1400
g for 23 minutes. Stellate cells then separated into a hazy band just above the interface of the Nycodenz solution and the aqueous buffer. Cells were harvested, washed, and resuspended in IMDM containing 10% FBS, 4 mmol/l glutamine, and antibiotics. PSCs were all used within two passages following isolation.
Conditioned medium
The poorly differentiated pancreatic adenocarcinoma cell line PANC-1 (American Type Tissue Culture, Rockville, MD) was grown in DMEM in 75 cm2 flasks. When the cells reached confluence, the serum-containing medium was removed and the cells were cultured in 20 ml of serum-free medium. After 24 hours, the medium was collected and the peptide containing fraction obtained by semi-purifying on a Sep-Pak Plus C18 Cartridge (Waters, Milford, MA). After washing, Sep-Paks were eluted with 50% acetonitrile (EM Science, Gibbstown, NJ) with 0.1% trifluoroacetic acid (J. T. Baker, Phillipsburg, NJ). The eluates were lyophilized and reconstituted in fresh IMDM with 1% FBS, forming what we refer to as PANC-1 conditioned medium (CM). In total extracts of 100 ml pf PANC-1 conditioned media were purified and reconstituted in 20 ml of media for PSC culture. Control medium consisted of only serum-free medium without PANC-1 cells which underwent are same Sep-Paking procedure. The eluates were also lyophilized, and then reconstituted in fresh IMDM with 1% FBS.
Immunostaining
COX-2 and α-SMA expression in PSCs was evaluated by immunohistochemical staining. Cultured PSCs were grown directly on glass coverslips in six-well plates, and immunostained for COX-2 using peroxidase-labeled streptavidin for immunohistochemistry (KPL, Inc., Maryland) according to the manufacturer's instructions. Cells were fixed for 30 minutes in acetone at -20°C. Thereafter, glass coverslips were air-dried and stored at 4°C until the cells were stained. Endogenous peroxidase activity was blocked by incubation in methanol with 0.3% hydrogen peroxidase for 30 minutes. After immersion in normal goat serum for 30 minutes, the slides were incubated with COX-2 (murine) polyclonal antibody (Cell Signaling Technology, Inc., Beverly, MA) diluted 1:200 in tris-buffered saline (TBS) 1× bovine serum albumin and stored in a humid chamber overnight at 4°C. The slides were incubated with anti-mouse immunoglobulins for 10 minutes at 37°C, followed by peroxidase conjugated streptavidin for 30 minutes at room temperature. Finally, color was developed incubating the slides for 8 minutes, with diaminobenzine (DAB Reagent Set; KPL, Inc., Maryland). Expression of α-SMA was examined in a similar manner by using monoclonal anti-α-SMA antibody (Sigma, St Louis, MI).
Protein concentrations in the cell lysates were measured by the method of Lowry et al. [
28] using bovine serum albumin as the standard.
3H-methyl thymidine incorporation
Following the treatment of cells with PANC-1 CM for 48 hours, DNA synthesis was measured by adding to each well 0.5 μCi of 3H-methyl thymidine and incubating these plates for the final 24 hours. The cell protein was precipitated with 10% trichloroacetic acid overnight, washed twice with phosphate buffered saline (PBS) and then dissolved by adding 0.25 ml of 0.5 mol/l NAOH to each well. Incorporation of 3H-methyl thymidine into DNA was measured by adding 1 ml of scintillation cocktail (ScintiSafe Plus 50%, S × 25-5, Fischer Scientific, Pittsburgh, PA) followed by count measurements using the Wallac WinSpectral liquid scintillation counter (Wallac Turku, Finland).
Cell counts
Cells were washed with PBS, harvested by trypsinization using 0.5% trypsin-0.2% EDTA, resuspended in 100 μl culture medium, and counted using the Guava Personal Cytometer (Guava Technologies, Inc., Burlingame, CA).
Western blotting
Expression of COX-2 and α-SMA were detected by Western blotting. Protein extracts (5 μg) from each sample were separated by gel electrophoresis using a 10% sodium dodecyl sulfate-polyacrylamide gel (SDS-PAGE). Known molecular weight protein standards were run alongside the samples. Separated proteins were then transferred to a nitrocellulose membrane (Bio-Rad, Hercules, CA), which was incubated for one hour at room temperature in blocking buffer (TBS and 0.1% Tween 20 with 5% nonfat dry milk). A murine COX-2 polyclonal antibody was diluted 1:2000 buffer (TBS and 0.1% Tween 20 with 5% nonfat dry milk). After incubation with the primary antibody overnight at 4°C, the membrane was exposed to the secondary antibody with gentle agitation for one hour at room temperature. Western Blots were visualized using Chemiluminescence Luminol Reagent (Santa Cruz Biotechnology, Inc., CA). α-SMA expression was examined in a similar manner by using monoclonal anti-α-SMA antibodies. Bands from individual western blots were quantified densitometrically and the mean ± SEM for each time point or concentration calculated for presentation.
Statistical analysis
All experiments were repeated at least twice. Data are expressed as mean ± SEM. Statistical analysis was performed using ANOVA with the Prism software package (GraphPad, San Diego, CA).
Authors' contributions
SY, MU, CP and WD participated in the PSC isolation and western blot experiments, cell culture experiments and drafted the manuscript. TA, XD, RB, MT and WD participated in the design of the study and trouble-shooting of the experiments. All authors read and approved the final manuscript.