Alkaline cancer pancreatic vs

* Context.-Blood tests possessing higher diagnostic accuracy are needed for all the major pancreatic diseases. Glycoprotein 2 (GP2) is a protein that is specifically expressed by the pancreatic acinar cell and that has previously shown promise as a diagnostic marker in animal models of acute pancreatitis.

Objective.-This study describes the development of an assay for GP2, followed by the determination of plasma GP2 levels in patients with acute pancreatitis, chronic pancreatitis, and pancreatic cancer.

Design.-Rabbit polyclonal antisera and mouse monoclonal antibodies were generated against human GP2 and used to develop an enzyme-linked immunosorbent assay. The assay was tested in patients with an admitting diagnosis of pancreatic disease at 2 tertiary care facilities. The diagnosis of acute or chronic pancreatitis and pancreatic cancer was determined using previously established criteria that incorporated symptoms, radiology, pathology, and serology. Plasma GP2 levels were determined in 31 patients with acute pancreatitis, 16 patients with chronic pancreatitis, 36 patients with pancreatic cancer, and 143 control subjects without pancreatic disease. Amylase and lipase levels were also determined in patients with acute pancreatitis.

Results.-The GP2 assay's sensitivity values were 0.94 for acute pancreatitis, 0.81 for chronic pancreatitis, and 0.58 for pancreatic cancer, which were greater than the 0.71 for acute pancreatitis and 0.43 for chronic pancreatitis (P = .02) observed for amylase. The lipase assay sensitivity for acute pancreatitis was 0.66. The accuracy of the GP2 assay was greater than that of the amylase or lipase assays for acute pancreatitis (GP2 vs lipase, P = .004; GP2 vs amylase, P = .003) when analyzed using receiver operator characteristic curves. When daily serial blood samples were obtained for 13 patients with acute pancreatitis, GP2 levels remained abnormally elevated for at least 1 day longer than the amylase or lipase levels.

Conclusion.-The GP2 assay is a useful new marker for acute and chronic pancreatitis.

(Arch Pathol Lab Med. 2004;128:668-674)

Common pancreatic diseases for which diagnostic assays would be clinically useful include acute and chronic pancreatitis, and pancreatic cancer. Most serum diagnostic assays studied to date have used proteins normally expressed by the exocrine pancreas and include amylase, lipase, trypsin, and phospholipase A. The assays have established their utility mainly in the diagnosis of acute pancreatitis.12 A drawback of many assays, however, is the production of the same or similar proteins by non-pancreatic tissues. With respect to amylase, its short half-life in blood compromises its diagnostic utility.3 Previous studies have demonstrated that serum amylase levels are normal in 30% of patients with acute alcoholic pancreatitis who present 48 hours or later after the onset of symptoms.4 Thus, the development of assays that measure proteins with longer serum half-lives should enhance our ability to diagnose acute pancreatitis.

For other common pancreatic diseases, such as chronic pancreatitis and pancreatic cancer, diagnostic assays using blood or other body fluids are much less helpful. Serum assays that can accurately diagnose chronic pancreatitis do not exist.5 The most widely used serum assay for pancreatic cancer, CA 19-9, possess sensitivities as high as 80%. CA 19-9, however, is inadequately sensitive to be used in population screening.6

The present study examines plasma levels for the pancreatic protein glycoprotein 2 (GP2) in patients with pancreatitis and pancreatic cancer. GP2 is the major membrane protein present in the pancreatic exocrine zymogen granule.7 Although GP2's cellular function is unknown, there is considerable knowledge concerning its biochemical properties. GP2 is directly linked to the membrane via a glycosylphosphatidylinositol linkage.8,9 When pancreatic secretion is stimulated, Glycoprotein 2 (GP2) is transported to the apical plasma membrane of the acinar cell, where it is cleaved and secreted in a polarized manner into the pancreatic duct.

Many human diseases, including cancer or inflammatory conditions, perturb the polarity of epithelial cells. The majority of cells in the digestive system that participate in human disease are polarized epithelia. For inflammatory or neoplastic diseases involving the gastrointestinal tract, proteins that are normally secreted in a polarized manner into the intestinal lumen or bile duct may be inappropriately released into the vasculature via the serosal space. Several other glycosylphosphatidylinositol-linked proteins that exhibit cellular polarity have served as useful clinical markers for gastrointestinal disease and include carcinoembryonic antigen, alkaline phosphatase, and 5'-nucleotidase.10,11 Thus, studies focused on GP2 were begun because the protein is normally secreted in a polarized manner, and its expression is restricted to the pancreatic acinar cell.

Preliminary studies using a rodent model of acute pancreatitis established that serum GP2 levels correlate with disease severity.12 Additional experiments confirmed that GP2 is specifically expressed by the pancreas. The results encouraged the development of an assay for human GP2 (HGP2) and studies to determine plasma levels in patients with pancreatic diseases.

MATERIALS AND METHODS

Production of Anti-HCP2 Antibodies

Rabbit polyclonal anti-HGP2 sera was produced as previously described.13 The mouse anti-HGP2 monoclonal antibodies were produced using recombinant HGP2 protein expressed from Chinese hamster ovary cells. The recombinant GP2 protein was produced using polymerase chain reaction to mutate the HGP2 complementary DNA (cDNA), such that the amino acids necessary for the formation of the glycosylphosphatidylinositol linkage was deleted and replaced with 6 histidine residues.14 The resultant cDNA encodes for amino acids 1 through 505 of HGP2, in addition to the 6 histidine residues at the carboxyl-terminal end. The cDNA was subcloned into a plasmid expression vector, pDNA3.1 (Invitrogen, Carlsbad, Calif), which was used to transfect Chinese hamster ovary cells using Lipofectamine (Gibco-Life Technologies, Grand Island, NY). Gentamycin (G418, Gibco-BRL)-resistant, stable CHO-HGP2 clones were screened for high expression and used to produce soluble secreted HGP2. HGP2 that was secreted into the media was affinity-purified using a nickel-based column according to the manufacturer's instructions (Qiagen, Valencia, Calif). BALB/c mice and New Zealand white rabbits were immunized with the purified His-HGP2 for antibody production. Mouse monoclonal antibodies were produced as previously described.12

Enzyme-Linked lmmunosorbent Assay

An enzyme-linked immunosorbent assay (ELISA) was developed using the mouse monoclonal (clone 2G2/E10) and rabbit polyclonal anti-HGP2 antibodies. Polyvinyl microtiter plates (Dynex Technologies Inc, Chantilly, Va) were coated with mouse monoclonal antibody 2G2/E10 and incubated at 4?¡ãC for 16 hours and then blocked with 1% (wt/vol) bovine serum albumin (Fraction V, Sigma Co, St Louis, Mo) in 50mM Tris-buffered saline for 2 hours at 37?¡ãC. The plates were washed with phosphate-buffered saline (PBS) with 0.05% (vol/vol) Tween 20 (Sigma) (PBS-T). Each patient sample was serially diluted with buffer (Tris-buffered saline with 0.2% [wt/vol] bovine serum albumin, 0.05% [vol/vol] Tween 20 and 1OmM EDTA) before 100 ??L was added to each well. The plates were incubated for 1 hour at 37?¡ãC followed by a wash with PBS-T. The plates were then incubated for 1 hour at 37?¡ãC with the rabbit anti-HGP2 antisera (1:1000 dilution). The plates were washed again with PBS-T before an alkaline phosphatase-conjugated goat anti-rabbit IgG antibody (ICN Biochemical, Aurora, Ohio) was added to each well for 1 hour at 37?¡ãC. After a final wash, p-nitrophenyl phosphate (Sigma) was used for color development and detected with a microplate reader (Dynatech Laboratories, Chantilly, Va) at 405 nm. Purified recombinant HGP2 was used as a standard. Calculation of the number of moles of HGP2 was based on a molecular mass of 97 kd, of which 45% is composed of asparagine-linked glycosylation.13

Patient Recruitment and Sample Collection

Patients were recruited from the inpatient wards and outpatient clinics of Stanford Hospital and the Veterans Administration Palo Alto Health Care System (PaIo Alto, Calif). Thirty normal controls included healthy individuals with no history of pancreatic or gastrointestinal disease. A second set of negative control subjects consisted of 113 patients with no history of pancreatic disease, but with other gastrointestinal or nongastrointestinal problems.