Summary. Pancreatic cancer is a devastating diseasecharacterized by a dismal prognosis with most patientsdying within six months after diagnosis. Surgery is anoption in less than one in five of these patients, and evenwith tumor resection the majority of patients succumb tothe disease. Other effective treatment options are notavailable. Common features of pancreatic cancer aresevere cachexia, marked insulin resistance and diabetesmellitus. Several studies have demonstrated connectionsbetween pancreatic cancers and the endocrine pancreasand this has raised questions regarding the role of theislets of Langerhans in pancreatic adenocarcinoma. Thismanuscript reviews the recent literature in this field andaddresses several questions regarding the interactionbetween the islets of Langerhans and pancreatic cancer.This review considers the histological findings inpancreatic cancer, cell culture and animal experiments,the four islet cell types and the hormones they secrete, aswell as the influence of the arachidonic acid pathwayson islet cell function and pancreatic cancer. Whilepancreatic adenocarcinomas are ductal in nature, the cellof origin has not been identified and there is even someevidence that the islets may harbor the precursor cell.Considerable evidence suggests that the diabetes iscaused by the tumor, while other studies have identifieddiabetes as a risk factor. Clearly, the islets are importantin many aspects of this disease. However, even thoughprogress has been made, some questions regarding theinteraction of pancreatic cancer and the endocrinepancreas remain unanswered.

Key words: Pancreatic cancer, Islets, Diabetes, Stemcell, Transdifferentiation, Lipoxygenase

Introduction

Pancreatic cancer is extraordinary in its biology andaggressiveness. This devastating disease is a greatchallenge for oncologists and researchers throughout theworld in their efforts to help patients and improve themost abysmal prognosis in medicine. Pancreatic canceris the fourth leading cause of cancer death in men andwomen with a 5-year survival rate of 3% and a mediansurvival time of less than 6 months (Howard, 1966;Parker, 1996; Greenlee et al., 2000). Potential curativesurgery is only possible in about one in six patientsdiagnosed with pancreatic cancer and adjuvant orpalliative therapy with gemcitabine is the best availabletreatment option (Howard, 1996; Yeo and Cameron,1999). However, the incidence of pancreatic cancerequals its mortality, since 5-year survival is less than20% after tumor resection. While gemcitabine improvesthe quality of life of many patients, it prolongs survivalby only about 1 month (Greenlee et al., 2001;Heinemann, 2001; Abbruzzese, 2002; Ahrendt and Pitt,2002). In contrast to several other cancers, the incidenceof this disease is not decreasing. Indeed, incidence hasincreased in Japan and in African Americans in recentyears (Ohba et al., 1996; Lin et al., 1998; Oomi andAmano, 1998; Silverman et al., 1998; Greenlee et al.,2001; McCarty, 2001). The key for this trend appears tobe changes in lifestyle, with increased smoking and awestern world diet rich in ω-6 fats. Because of theextreme resistance of pancreatic cancer to availabletherapeutics, it is even more important to identify riskfactors and to prevent or diagnose this disease at an earlystage, before invasive spread. The high fat western dietis obviously a major risk factor for many diseasesincluding cancers of the colon, breast, prostate andpancreas as well as for type 2 diabetes mellitus and heartdisease (Rose, 1997; Woutersen et al., 1999). Severalstudies have shown a high incidence of diabetes inpancreatic cancer and it is generally believed thatdiabetes is a risk factor for this malignancy (Schwarts etal., 1978; Friedman and van den Eeden, 1993; Fisher et

Review

On the role of the islets of Langerhans in pancreatic cancerR. Hennig, X-Z. Ding and T.E. AdrianDepartment of Surgery, Feinberg School of Medicine, Northwestern University, Chicago, IL, USA

Histol Histopathol (2004) 19: 999-1011

Offprint requests to: Thomas E. Adrian Ph.D, FRCPath, Professor ofSurgery and Pathology, Department of Surgery, NorthwesternUniversity, Tarry Building, 4-711, 303 East Chicago Avenue, Chicago,Il l inois 60611, USA. Fax: (312) 503 3489. e-mail:tadrian@northwestern.edu

http://www.hh.um.es

Histology andHistopathology

Cellular and Molecular Biology

al., 1996). In contrast, however, other studies havesuggested that the diabetes is caused by the tumor ratherthan being a risk factor (Gullo et al., 1994; La Vecchia etal., 1994; Gullo, 1999). Remarkably, the firstobservations on the interaction between the pancreaticducts and the islets of Langerhans were made in 1911(Bensley, 1911). A lot of very interesting data has beenproduced since then and the connections between theendocrine pancreas and pancreatic cancer have become aresearch focus of several groups. This review willaddress the following questions: How is diabetesconnected to pancreatic cancer? What is the relationshipbetween the islets of Langerhans and pancreatic cancer?Do the islets harbor the cells of origin of pancreaticcancer? What roles do islet hormones play in thedevelopment and growth of pancreatic cancer?

Histological findings

The pancreas is a complex tissue and is perhaps theleast understood organ besides the brain in mammals.Three major components build the structure of thepancreas. Acinar cells represent 85% of the tissuevolume and are, therefore, considered as the leaders inorgan function and failure (Pour et al., 2002). Ductalcells contribute less than 10% and islets only 1-2% ofthe volume of the pancreas (Pour et al., 2002). It wouldbe misleading to consider the islets as the least importantfraction of the pancreas, however. The Islet cell-specifictranscription factor, PDX1 is absolutely essential for thedevelopment of the pancreas, while the α-cells whichproduce glucagon gene products and PYY in the fetusare one of the earliest cell types appearing duringpancreatic organogenesis (Herrera et al., 1991; Lukiniuset al., 1992; Jonsson et al., 1994; Slack, 1995; Offield etal., 1996). The pancreatic islets were first described byLangerhans in 1869 (Langerhans, 1869). Thedistribution of islets throughout the exocrineparenchyma, rather than forming a solid gland, isunusual for an endocrine tissue. This provides evidenceof a special role for the islets within the pancreas. Thepancreas originates from two outgrowths of the primitiveforegut, a ventral (head and uncinate process) and adorsal (pancreas body and tail) anlage. This has animpact on islet cell shape (ventral-diffuse and dorsal-compact spheric) and the cellular composition of theislets (Pour et al., 2002). There are four different isletcell types, the glucagon producing α-cells, the insulinand amylin (also called islet amyloid polypeptide, IAPP)producing ß-cells, the somatostatin producing δ-cellsand pancreatic polypeptide producing PP-cells. PP-cellsare mostly found in the uncinate process, while α-cellsare more common in the body and tail of the pancreas(Orci et al., 1978; Stefan et al., 1982). The arrangementof endocrine cells varies between species. In rats andhamsters, there is a clear structure with ß-cells in thecenter and α- and δ-cells in the periphery of the islets(Pour et al., 2002). In rodents, the microcirculation isorganized, with much of the blood flowing from the α-

and δ-cells to the ß-cells and only afterwards reachingthe acinar tissue (Pour et al., 2002). Thus, we speak froman islet-acinar portal system, which suggests that theislets are the important metabolizing station in thepancreas, perhaps somewhat akin to the role of the liverin first pass metabolism for the whole body. The cellulararrangement in human islets is not organized like that ofrodents, however, and this is reflected in the lack oforganized pattern in the microcirculation. PP cells arenot only associated with the islets and are foundscattered in normal ducts throughout the exocrine tissue(Pour et al., 2002). In contrast, ductal cells can belocalised within islets, an observation first made byBensley in 1911 (Bensley, 1911). Intrainsular ductuleswere recently described in tissues from patients withpancreatic cancer or chronic pancreatitis (Pour andSchmied, 1999). The islet cells in 70% of pancreaticcancer patients with altered glucose tolerance exhibitreduced insulin content (Kimura et al., 1998) but expressductal cell markers such as CA19-9, DU-PAN-2 andTag-72 (Kimura et al., 1998; Pour and Schmied, 1999).Furthermore, in the BOP-hamster model, the firstmorphological changes which occur during pancreaticcarcinogenesis are the appearance of ductular structureswithin or around the islets (Schmied et al., 1999). Theseductules proliferate and become microcystic adenomasor hyperplastic, dysplastic, atypical and finallymalignant glands, destroying the islets and invading thesurrounding tissue (Schmied et al., 1999). This is inaccordance with an incidental finding at autopsy, inwhich a microscopical pancreatic carcinoma wasidentified in an islet (Schmied et al., 1999). In Figure 1an islet from adjacent to a pancreatic cancer is shownwith the described intra-insular ductules.Immunohistochemical studies have shown the abnormalco-localization of islet hormones and the presence ofendocrine cells in invasive regions of pancreaticcarcinoma (Pour et al., 1993). The islets withinpancreatic cancer tissues contain less ß-cells but more α-cells and this is not compensated in the extrainsulartissue (Schmied et al., 2001). These data stronglysuggest a relationship between islets and pancreaticcancer. However, doublestaining for chromogranin Aand Ki-S5 revealed that the majority of scatteredendocrine cells within ductal adenocarcinomas are anon-neoplastic tumor-associated cell population, unlessthese cells are intimately integrated into the neoplasticglandular epithelium (Ohike et al., 2003). At present,pancreatic cancers are classified as ductaladenocarcinomas of the pancreas, based on thehistological finding that the tumor cells mimic ductalstructures. It is, therefore, widely believed that ductalcells are the cells of origin of pancreatic cancer and aprogression model from normal ductal epitheliumthrough intraepithelial neoplasia to invasive ductaladenocarcinoma has been developed (Hruban et al.,2000a,b). However, this assumption needs to bequestioned in context of above mentioned data and sinceSakaki et al. also described non-neoplastic endocrine

1000

Islets and pancreatic cancer

cells between tumor cells which, in turn are in closecontact with surrounding islets (Sakaki et al., 2002). It isalso remarkable that islets tend to survive in invasivecancers while acinar cells do not (Kodama and Mori,1983). However, even islets far from the tumor mayexhibit changes, suggesting either primary damage by apossible carcinogen or that they are influenced by thecancer cells (Pour et al., 2002). It should be mentionedthat endocrine cells are only detected in well-differentiated tumors but not in poorly-differentiatedones, paralleling the loss of islet cell markers in culturedislets during the process of dedifferentiation (Pour et al.,2002). In addition, Regitnig et al. reported a case of auniform insulinoma in the pancreas where a livermetastasis developed with insular-ductulardifferentiation (Regitnig et al., 2001). An important

paper recently described nestin positive stem cells in theislets and ductal tissues of the mouse pancreas (Hunzikerand Stein, 2000). Furthermore nestin-positive progenitorcells have been obtained from adult human and ratpancreatic islets (Zulewski et al., 2001; Lechner et al.,2002). Clearly such stem cells could be the origin ofpancreatic cancer. However, once again this area ishighly controversial. Other groups have publishedpapers suggesting on the one hand that the nestinpositive cells in the pancreas are the endothelial cells innew developing blood vessels and on the other hand thatthe only nestin positive cells are of mesenchymal origin(Lardon et al., 2002). The observation that islet cells cantransform into duct cells suggests that at least somehuman pancreatic adenocarcinomas may originate withinislets via transdifferentiation. Considering all above

1001

Islets and pancreatic cancer

Fig. 1. 5-LOX and LTB4 receptor expression in human pancreatic tissues. Immunohistochemistry for 5-LOX is shown in panels A and B and for LTB4receptor in panels C and D. A and C. Normal human pancreas from a multi organ donor with an unstained islet. B and D. Pancreatic adenocarcinomawith islets surrounding the tumor showing strong positive staining in cytoplasm, nucleus and nuclear membrane. D. This islet also shows intrainsularductules. The immunohistochemistry for 5-LOX used mouse-monoclonal antibody, 1:250, microwave citrate buffer pretreatment, incubation over nightat 4 °C, and DAB and for LTB4 receptor rabbit-polyclonal antiserum, 1:200, incubation over night at 4 °C, and HistoMark Red. Hematoxylincounterstained all sections. x 400

mentioned histological findings a final statementwhether or not islet cells could be the cells of origin forpancreatic adenocarcinoma cannot be made, especiallysince most of the endocrine cells within tumors seem tobe non-neoplastic. Their presence and survival inpancreatic adenocarcinoma could be explained by theadvantage for tumor cells to get important growth factorssuch as insulin directly provided by them while, apossible role in originating these tumors cannot beexcluded.

Experimental studies to determine the cell of origin

In vitro studies

Numerous cell culture experiments have beenperformed to shed light on the cellular origin ofpancreatic cancer. However, the isolation, purificationand maintenance of human pancreatic cells are difficultand limited. While hamster ductal and islet cellsimmortalize spontaneously, their human counterpartsundergo senescence after approximately 10 months(Ulrich et al., 2002). Human acinar cells could becultured for only 21 days (Ulrich et al., 2002). Use of arecently described technique allowed the culture ofhuman ductal cells for more than 16 months, withoutgenetic manipulations or by introducing foreign genes asthe large T antigen of SV40 which alter regulatory genessuch as p53 and pRb (Ulrich et al., 2000, 2002). Onegroup were able to culture immortalized pancreaticductal cells using large T antigen of SV40 and thesecells acquired the ability to form tumors aftertransfection with mutated K-ras (Jesnowski et al., 1999;Lohr et al., 2001). Hamster islets could successfully bemaintained in culture indefinitely and human islets formore than 12 months, however, within 14 days theseislet cells lost the ability to produce hormones (Schmiedet al., 2000; Ulrich et al., 2002). When purified islets arecultured, acinar cells and ductular structures developprimarily within the center, and these structures expressductal cell markers such as cytokeratin 7, 19,pancytokeratin, carbonic anhydrase II, CA19-9 and DU-PAN-2; markers which are also expressed in mostpancreatic adenocarcinomas (Pour et al., 2002). Duringthe process of culture, the islet cells lose their endocrinegranules. Cells containing a mixture of insulin andglucagon then appear, imitating the embryonic pancreas.After 60 days, the islet cells are completely replaced byundifferentiated cells expressing cytokeratines, α1-antitrypsin, vimentin and nestin (Pour et al., 2002).Reports have suggested that nestin could be a marker forprecursor cells in the pancreas, however, results obtainedfrom western blotting and immunohistochemistry arecontroversial in localizing these nestin positive cells toeither ducts, islets or blood vessels and mesenchymalstructures (Lee et al., 2003). Very recently, immortalizedNestin positive cells have been developed by stableexpression of hTERT (Lee et al., 2003). This is a majoradvance that will hopefully help to define the biologicalrole of nestin positive cells and determine the cell of

origin for pancreatic cancer. Islet cells are also capableof transdifferentiation into acinar cells, hepatocytes ormucinous and oncocyte-like cells (Yuan et al., 1996;Pour and Schmied, 1999). Indeed, acinar, ductal and isletcells can all act as facultative stem cells, or are able totransdifferentiate in each other under reactivation ofPDX1 and finally redifferentiate into cells with a stemcell character (Bardeesy and DePinho, 2002). Whencultured hamster islet and ductal cells were treated withthe pancreatic carcinogen, BOP invasive ductaladenocarcinomas developed (Pour et al., 1975; Schmiedet al., 1999). However, only tumors derived fromhamster islet cell cultures showed mutations of K-rasand p16, which are the most frequent genetic mutationsin human pancreatic cancer (Ikematsu et al., 1997;Schmied et al., 1999; Muscarella et al., 2001).

Whether the appearance of exocrine cells withinislets is caused by transdifferentiation of endocrine cellsor activation and proliferation of pancreatic stem cells isstill not clear and needs to be determined. It is temptingto speculate that it is a specific genetic alteration ratherthan a specific target cell that is crucial for thedevelopment of pancreatic cancer. However, one celltype could still be more susceptible to genetic mutationsthan the others.

In vivo studies

Islets promote pancreatic cancer

Important information has also been gained fromdifferent animal models. Syrian golden hamsters treatedwith the carcinogen BOP develop pancreaticadenocarcinoma that is similar to the human disease bothgenetically and histologically (Fujii et al., 1990). This isa valuable model for studying tumor development andbiology. In this model, destruction of ß-cells withstreptozotocin inhibits pancreatic tumor development,while stimulation of islet cell proliferation augmentspancreatic carcinogenesis after BOP treatment (Pour etal., 1990; Pour and Kazakoff, 1996). This group reportedsimilar results in Chinese hamsters, where geneticallydiabetic hamsters with atrophic islets did not developpancreatic cancer following BOP treatment but non-diabetic controls did (Bell and Pour, 1987).

Islets as the origin of pancreatic cancer

Transplantation of hamster islets into thesubmandibular gland of Syrian golden hamstersfollowed by BOP treatment led to the development ofductal pancreatic adenocarcinoma in this site (Pour et al.,1997). Cancer was not observed after transplantingductal or acinar cells into this gland. Moreover, thesubmandibular gland itself is not a target for BOP.

Ductal cells as the origin of pancreatic cancer

Pancreatic adenocarcinoma can be induced in ratsusing the carcinogen DMBA (Jimenez et al., 1999). In

1002

Islets and pancreatic cancer

contrast to the systemic application of BOP in hamsters,crystals of DMBA are directly implanted into thepancreatic head of rats. Immunohistochemical studiesrevealed the expression of ductal markers such askeratin, cytokeratin 19 and 20 in very early stages oftumor development, suggesting a ductal cell of origin inthis model (Jimenez et al., 1999). Another interestingfinding is that after partial pancreatectomy in rats,regeneration of the exocrine and endocrine pancreasoccurs from differentiated cells as well as multipotentductal cells (Sharma et al., 1999). Proliferating duct cellsserve as source for both the endocrine and exocrinecompartment, recapitulating embryogenesis and showingthat differentiated duct cells can become multipotentduring this process (Sharma et al., 1999; Bardeesy andDePinho, 2002).

Acinar cells as the origin of pancreatic cancer

In the azaserine rat model, where acinar cells are thecellular target, duct-like cell lines have been derivedfrom an acinar cell carcinoma (Pettengill et al., 1993).Several groups have developed transgenic mouse modelsof pancreatic cancer. The elastase promoter of acinarcells is usually targeted for transgene expression. Sincepancreatic cancers develop in the progeny oftransforming growth factor alpha (TGF-α) transgenicmice crossed with p53-null mice, acinar cells can clearlyserve as the starting point for trans- or dedifferentiation(Greten et al., 2001; Wagner et al., 2001). In this model,duct-like tubular structures form, proliferate andeventually develop pancreatic ductal adenocarcinoma(Wagner et al., 2001).

Stem cells as the origin of pancreatic cancer

Animal models indicate that there are severalpossible ways to develop pancreatic cancer, either from aparticular stem cell or through transdifferentiation. Itappears that, at a certain stage, all three cell types havethe plasticity to dedifferentiate and could be the cell oforigin of pancreatic adenocarcinoma. This suggests thatdifferentiation is not an irreversible process and thatislet, duct and acinar cells are facultative stem cells, ableto gain stem cell properties back during the process oftrans- and dedifferentiation. In this process of re-enteringthe cell cycle, PDX1, which is the transcription factor inpluripotent progenitor cells in the embryonic pancreas,becomes reactivated (Bardeesy and DePinho, 2002).Once again, we get the impression that a particular cellstate and increased proliferation rate, combined with theeffect of a carcinogen is much more important than thecell type for the development of pancreaticadenocarcinoma. However, it should be kept in mindthat only the BOP hamster model closely mimics allaspects of the human disease and only islet but notductal cells transplanted into the submandibular glandwere able to develop pancreatic cancer after BOP

treatment.

Islet hormones and pancreatic cancer

Diabetes and pancreatic cancer

The observation that diabetes occurs in two thirds ofpancreatic cancer patients is one of the most interestingaspects of this disease and has led to the speculation thatthe endocrine pancreas is in some way connected withthis cancer (Schwarts et al., 1978; Permert et al.,1993a,b). The diabetes in pancreatic cancer patientsoften remains unrecognised or is diagnosedconcomitantly with the pancreatic cancer or within a fewmonths before the cancer diagnosis, most probablybecause of an asymptomatic hyperglycemia in thesepatients (Chari et al., 2001). This is tragic, becauseepidemiological and in vivo data suggest that themetabolic defect occurs at a resectable stage of thisdisease (Liu et al., 1998; Chari et al., 2001; Permert etal., 2001). Logically, there are two possibilities eitherdiabetes is a risk factor for pancreatic cancer or thismetabolic disease arises from the tumor. Some datasuggest diabetes as a risk factor for pancreatic cancer(Gapstur et al., 2000; Ghadirian et al., 2003). However,several other studies, including epidemiological data,demonstrate that diabetes is a consequence of pancreaticcancer and moreover, that patients with diabetes type Iseem to be protected against this dreadful disease (Gulloet al., 1994; Fisher et al., 1996; Gullo, 1999; McCarty,2001). Based on these latter findings, suggest thatdiabetes mellitus is not really a risk factor for pancreaticcancer, but rather insulin resistance as a product ofobesity predisposes to the development of pancreaticadenocarcinoma as well as type 2 diabetes.

Obesity and insulin resistance in the context ofpancreatic cancer

Several epidemiological studies have shown thatobesity is a risk factor for pancreatic cancer and sodiabetes should be considered in the context of obesity(Silverman et al., 1998; Ghadirian et al., 2003).Pancreatic cancer patients are mostly lean or cachectic atthe time of diagnosis but 75% of these patients wereobese or had a significant higher BMI before the onset ofsymptoms (Chari et al., 2001). Chari et al. found in theirpatient population, that the prevalence of diabetesassociated with pancreatic cancer was higher in obesepatients compared to patients with a normal BMI and,therefore, concluded that obesity may predispose thesepancreatic cancer patients to diabetes (Chari et al.,2001). However, these data cannot answer the questionwhether the tumor or diabetes comes first. Therefore,whether obesity or the tumor itself is responsible forpancreatic cancer associated diabetes needs to be furtherinvestigated. Nonetheless, obesity is an important riskfactor for type 2 diabetes as well as pancreatic cancer

1003

Islets and pancreatic cancer

(Silverman et al., 1998, 1999). It is quite likely that bycausing insulin resistance, obesity has a promotionaleffect on early pancreatic cancer and then later the tumoritself causes insulin resistance and a further promotionaleffect. A high fat diet has a promotional effect in theBOP-hamster model of pancreatic cancer (Kazakoff etal., 1996). A high fat diet induces peripheral insulinresistance, leading to a compensatory islet cellhyperactivity and proliferation which is an importantpromoting factor for pancreatic cancer induction (Pourand Kazakoff, 1996). Metformin, a drug used in thetherapy of type 2 diabetes can prevent the induction ofpancreatic cancer promoted by a high fat diet (Schneideret al., 2001). This drug normalizes insulin levels and therate of islet cell turnover by improving insulin resistancethrough an increased glucose uptake, oxidation andglycogenesis in the skeletal muscle and by inhibitinghepatic gluconeogenesis (Schneider et al., 2001). Thesize of islets in the metformin-treated hamsters wassignificantly smaller than untreated controls, reflectingsuppression of islet cell proliferation and, therefore, adecreased response to the carcinogen BOP (Schneider etal., 2001). Although, this study leads to the impressionthat diabetes is a risk factor for pancreatic cancer, severalepidemiological studies show that diabetes is aconsequence of pancreatic cancer but also show thatobesity is a risk factor for the disease. Certainly, thiscould explain the increased incidence of pancreaticadenocarcinoma in Japan and in African Americans andeven more efforts needed to be put into this metabolicrelationship to pancreatic cancer.

ß-cells (insulin and islet amyloid polypeptide producingcells)

Islet ß-cells as possible stem cells, activators forcarcinogens and the source of growth factor that promotepancreatic cancer

Experiments using the ß-cell toxin streptozotocinshould be mentioned there, because there are at leastfour possible explanations for the preventive effect ofstreptozotocin in the BOP hamster model (Bell et al.,1988, 1989). The destruction of ß-cells: 1) abolishes asource of cells able to transdifferentiate anddedifferentiate; 2) prevents the metabolism of the BOPprocarcinogen to its active carcinogenic substance; 3)withdraws insulin as an important growth factor; and 4)may destroy putative pancreatic stem cells. Thesethoughts are consistent with the observations that insulinhas a significant impact on pancreatic cancer growth andthat patients with type I diabetes mellitus appear to beafforded some protection against this disease. Finally, weagain get the impression that islet ß-cells play animportant role in pancreatic cancer development andgrowth, because these cells appear to be in aproliferative or regenerative status susceptible topancreatic cancer causing carcinogens, provide insulinwhich can promote tumor growth by direct and indirect

mechanisms and they release amylin, a peptide whichmay play a role in peripheral insulin resistance andcachexia.

Insulin promotes pancreatic cancer

The impaired glucose metabolism in pancreaticcancer patients is characterized by hyperinsulinemia andperipheral insulin resistance, caused by a post-insulinreceptor defect which reduces skeletal muscle glycogensynthesis and glycogen storage (Liu et al., 2000). Insulinsecretion is not impaired, however, with different groupsreporting relatively normal to increased insulin secretion(Fogar et al., 1993; Valerio et al., 1999; Liu et al., 2000).Furthermore, the study that revealed relatively normalinsulin secretion demonstrated high basal plasma insulinlevels (Liu et al., 2000). The resulting hyperinsulinemiaappears to be important, because pancreatic cancerpatients with diabetes have an even worse prognosis anda shorter survival time than patients without diabetes(Fisher et al., 1996). The anatomical arrangement of theislets scattered throughout the exocrine parenchyma andthe estimated 20-fold higher concentrations of insulin inthe pancreas than in the systemic circulation, suggest akey role of insulin in pancreatic cancer growth anddevelopment (Chandrasekar and Korc, 1991; Ding et al.,2000). Insulin stimulates pancreatic cancer cellproliferation via MAP kinase activation and glucoseutilization by PI3 kinase activation and enhancement ofexpression of GLUT-1, a glucose transporter widelyexpressed in human tissues that regulates basal glucosetransport (Ding et al., 2000). Therefore, insulin promotespancreatic cancer growth directly as a growth factor andindirectly by increasing substrate uptake andmetabolism. Pancreatic cancer cells express insulin aswell as insulin-like growth factor 1 (IGF-1) receptorsand, regardless of which receptor is more involved, highintrapancreatic insulin concentrations give pancreaticcancer cells a growth advantage (Fisher et al., 1996;Ding et al., 2000). In contrast, one group showed thatexogenous insulin had an inhibitory effect on pancreaticcancer induction in the BOP-hamster model (Pour andLawson, 1984; Pour et al., 1990). However, insulinadministered systemically should block endogenousinsulin synthesis and secretion. In turn, this should leadto a substantial decrease in intrapancreatic insulinconcentrations. This could explain the inhibitory effectof exogenous insulin on pancreatic cancer, since highintrapancreatic insulin concentrations appear to be a verypotent stimulus. Interestingly, pancreatectomy inpancreatic cancer patients leads to an improvement inglucose metabolism and peripheral insulin resistance,suggesting that the tumor itself is responsible for thisdiabetic situation (Permert et al., 1993a,b). This is evenmore surprising, considering that approximately 85% ofthe pancreas is removed during surgery, which leads to amarked reduction of total islets (Permert et al., 1993).Therefore, pancreatic cancer associated diabetes couldperhaps be considered as an own entity, a type 3 diabetes

1004

Islets and pancreatic cancer

mellitus.

The role of amylin in pancreatic cancer

Several experimental studies support the hypothesisthat a tumor-associated factor causes peripheral insulinresistance and leads to diabetes. Transplanted MiaPaCa2cells induce impaired glucose tolerance in SCID mice, ametabolic effect that appeared to be caused by a peptidethat was detected in pancreatic cancer cell-conditionedmedium (Valerio et al., 1999). A factor that causeschanges in ß-cell secretion was also demonstrated inconditioned medium from PANC-1 and HPAF pancreaticcancer cells (Ding et al., 1998). This factor does notaffect insulin secretion but selectively stimulates amylin(IAPP) secretion (Ding et al., 1998). The peptidesdescribed by both groups could be the same factor,however, it has not as yet been fully characterized.Hypersecretion of amylin stimulated by this factor leadsto a significant reduction in intracellular amylin contentin the cells and this is consistent withimmunohistochemical findings, that islets adjacent totumors show substantially less staining for amylin buthave normal insulin content and mRNA for bothpeptides is readily detected by in situ hybridization(Ding et al., 1998). Amylin is a 37-amino-acidpolypeptide co-localized and co-secreted with insulinfrom ß-cells (Westermark et al., 1987). However, thesecretion seems to be independently controlled undercertain circumstances such as pancreatic cancer (Madsenet al., 1991). Amylin is a physiological inhibitor ofinsulin secretion and several studies have shown thatamylin can have diabetogenic effects in vitro and invivo, causing peripheral insulin resistance by inhibitingglucose uptake and glycogen synthesis in skeletalmuscle as well as opposing insulin effects on the liver(Leighton and Cooper, 1988; Young et al., 1990;Frontoni et al., 1991; Koopmans et al., 1991). However,the physiological significance of this has beenquestioned in many studies (Wilding et al., 1994; Arneloet al., 1997). Amylin is also the major constituent of thepancreatic amyloid observed in 90% of patients withtype 2 diabetes (Westermark et al., 1987). Plasma amylinlevels are significantly increased in pancreatic cancerpatients and normalize after surgical tumor extirpation(Permert et al., 1994). Furthermore, cell cultureexperiments confirmed these data, demonstrating adissociation of insulin and amylin secretion from isletsin pancreatic cancer cell conditioned medium resultingin an increased amylin/insulin molar ratio caused byhypersecretion of amylin (Ding et al., 1998). Thissupports the hypothesis that pancreatic cancer cellsproduce an amylin-releasing factor. In addition, amylinalso inhibits food intake at physiological concentrationsand may contribute to the anorexia associated with thesevere cachexia, a major problem in patients withpancreatic adenocarcinoma (Chance et al., 1991; Permertet al., 1994; Arnelo et al., 2000). It was suggested thatmeasurement of plasma amylin levels in patients with

recent onset of type 2 diabetes might be valuable fordetecting pancreatic cancer at an early stage, when theremay be a better chance for cure. However, twosubsequent clinical studies have shown that amylin issensitive enough to detect pancreatic cancer and isinferior to CA 19-9 (Chari et al., 2001; Brand et al.,2002). One problem is the lack of specificity of amylin,because elevated levels are also observed in acute andchronic pancreatitis or periampullary tumors (Chari etal., 2001). The amylin releasing factor could be a moresensitive marker, so efforts to identify this factor wouldbe worthwhile.

Glucagon-producing α-cells

Glucagon has no proliferative effect on pancreaticcancer cells, which is not really surprising becauseglucagon receptors are not expressed in cells from theexocrine pancreas or in pancreatic cancer cells (Ding etal., 2000). However, glucagon levels are elevated inpancreatic cancer, suggesting that α-cells are stimulatedin this disease (Permert et al., 1997). Like those ofamylin, glucagon levels normalize after subtotalpancreatectomy, suggesting that the hypersecretion is aresponse to stimulation by the tumor (Permert et al.,1997). Pancreatic cancer cell-conditioned media hasbeen shown to increase glucagon secretion from an isletcell line. While elevated glucagon levels seem to beunimportant for tumor growth, this catabolic hormonemay contribute to the cachectic state.

Somatostatin producing δ-cells

Effect of somatostatin on tumor cell growth

Pancreatic cancer cell growth may be regulated bysomatostatin and elevated plasma levels of this hormonehave been reported in patients with pancreatic cancerassociated diabetes (Takeda and Escribano, 1991;Permert et al., 1997; Wang et al., 1998; Ding et al.,2000). In contrast to amylin and glucagon, plasma levelsof somatostatin do not normalize after subtotalpancreatectomy (Permert et al., 1997). It is tempting tospeculate that hypersecretion of somatostatin may be adefence mechanism and the sustained elevation of thisislet hormone could indicate failure of cure of the cancer.It could be interesting to investigate plasma somatostatinconcentrations in patients with small tumors (<1cm)before and after surgery. While some recent studies havedemonstrated that somatostatin can induce significantinhibition of tumor cell growth and increase apoptosis,the literatrure is highly controversial on this point. Forexample, the growth response of MiaPaCa2 cells tosomatostatin and its analogues have been variouslyreported as inhibition, stimulation and lack of any effectbecause of absence of receptors (Gillespie et al., 1992;Fisher et al., 1996; Douziech et al., 1999).Concentration-dependent growth inhibition of PANC-1cells was seen after somatostatin treatment together with

1005

Islets and pancreatic cancer

stimulation of activity of membrane phosphotyrosinephosphatase SHP-1 (Douziech et al., 1999). In contrast,MiaPaCa2 cell growth was stimulated, regardless of theexpression level of somatostatin receptors (Douziech etal., 1999). SHP-1 was absent in MiaPaCa2 cells,suggesting that the growth inhibitory effect in PANC-1cells is mediated through SHP-1 activation (Douziech etal., 1999).

Expression of somatostatin receptors in pancreaticadenocarcinomas

Fisher et al. reported that out of four pancreaticcancer cell lines studied, somatostatin receptors wereonly expressed in MiaPaCa2 cells and in this cell linesomatostatin suppresses growth (Fisher et al., 1996). Asnoted above, Gillespie et al could not detect somatostatinreceptors in MiaPaCa2 cells and, therefore, found also afailure of response to somatostatin (Gillespie et al.,1992). These different results could be caused bydifferent technical approaches or just differences due topassage of this cell line. The expression of somatostatinreceptors in pancreatic cancer tissue is alsocontroversial, with one group reporting up-regulationcompared to normal pancreas and another reporting lossof receptors (Tang et al., 1998; Rochaix et al., 1999). Inone study, in which somatostatin receptors were detectedin pancreatic cancer cells, chemotherapeutics, such asgemcitabine caused down-regulation of the high affinityreceptors (Fueger et al., 2001). A novel therapeuticapproach could be the reintroduction of somatostatinreceptors via gene transfer, however, it is difficult to seehow this could be accomplished effectively in cancers(Rochaix et al., 1999; Vernejoul et al., 2002).

Clinical studies

A phase III clinical trial showed that thesomatostatin analog, octreotide was inferior to 5-fluorouracil in terms of delay of tumor progression andsurvival and as a consequence the octreotide arm of thestudy was abandoned (Burch et al., 2000). Clearly,somatostatin analogues cannot generally berecommended for treating pancreatic cancer. Screeningthe tumor for somatostatin receptors as well as SHP-1and testing the response of each individual tumor tosomatostatin e.g. in a xenograft model would have to beseriously considered before embarking on such anapproach.

Pancreatic polypeptide (PP) producing cells

PP cells are very interesting because of their relativelocalization in the head of the pancreas where mostpancreatic adenocarcinomas arise (Pour et al., 2002).Cytochrome P450 isoenzymes (CYP) are expressedprimarily or exclusively in islet cells (Pour et al., 2002).These enzymes are involved in the metabolism ofcarcinogens such as nitrosamines which are present in

cigarette smoke, a risk factor for pancreatic cancer (Pouret al., 2002). Four of the CYP isoenzymes areexclusively expressed in PP cells, which are frequentlyscattered within the ductal epithelium (Pour et al., 2002).The islet cell hormone, PP secreted by these cellsinhibits pancreatic adenocarcinoma growth (Fisher et al.,1998). Certainly, PP-cells warrant further attention andfuture investigation as the data to date are very limited.

Inflammatory pathway

This final section highlights the involvement of theinflammatory arachidonic acid pathways in islet cellphysiology and recent findings demonstratingoverexpression of these enzymes in pancreaticadenocarcinomas. Up-regulation of cyclooxygenase-2(COX-2) expression has been described in islets,pancreatic cancer cells and pancreatic intraepithelialneoplastic lesions (PanINs) (Koshiba et al., 1999;Molina et al., 1999; Okami et al., 1999; Tucker et al.,1999; Kokawa et al., 2001; Maitra et al., 2002).However, the COX inhibitors, aspirin and indomethacindo not influence insulin secretion, so the role of COX-2expression in islets is uncertain (Yamamoto et al., 1983;Pek and Nathan, 1994). In contrast, lipoxygenaseinhibitors like NDGA inhibit insulin and glucagonsecretion (Yamamoto et al., 1983; Pek and Nathan,1994). A negative feedback mechanism has beensuggested, where formed eicosanoids inhibit glucose-induced arachidonic acid and insulin release (Nathan andPek, 1990). Whether lipoxygenase metabolites such as5-, 12- and 15-HETE or leukotriene B4 (LTB4) inhibitor stimulate insulin secretion is controversial, perhapsreflecting the use of different study models (Yamamotoet al., 1983; Metz et al., 1984; Pek and Walsh, 1984,1985; Walsh and Pek, 1984a,b; Nathan and Pek, 1990;Pek and Nathan, 1994). Using immunocytochemistry,we recently found marked overexpression of 5-lipoxygenase (5-LOX) as well as the receptor for itsdownstream ligand, leukotriene B4 (LTB4) in humanpancreatic cancer cells as well as islets surrounding thetumors, while few cells in islets from normal pancreatictissues stained positive (Hennig et al., 2002) (Fig. 1).Interestingly, a clinical study revealed that patients witheither type 1 or type 2 diabetes mellitus have higherLTB4 levels than healthy controls and this was alsopositively correlated with glycated haemoglobin levels;(Parlapiano et al., 1999). Troglitazone inhibits LTB4production by direct inhibition of 5-LOX activity(Yamashita et al., 2000). This drug is a well-knownperoxisome proliferator activated receptor γ (PPARγ)agonist, improving insulin resistance and glucosehomeostasis in type 2 diabetics and also inhibitspancreatic cancer cell proliferation (Motomura et al.,2000). It should be kept in mind, that insulin is animportant growth factor and LTB4 a possible regulatoryelement. Therefore, the 5-LOX pathway may play animportant role in pancreatic adenocarcinoma and itsassociated diabetes mellitus. However, further studies

1006

Islets and pancreatic cancer

will be necessary to identify the underlying mechanisms.Nuclear localization of 5-LOX has been described.Finding this enzyme in the region of active genetranscription (euchromatin), raises the intriguingpossibility that 5-LOX might regulate transcriptionalevents in the nucleus (Peters-Golden and Brock, 2001).Since the nuclear envelope is the place of leukotrienebiosynthesis, it is reasonable to speculate that thesemetabolites might act as transcriptional regulators aswell as through G-protein coupled receptors (Funk,2001). The leukocyte type of 12-LOX is expressed inislet α- and ß-cells and this pathway participates incytokine-mediated ß-cell dysfunction and cytotoxicity(Shannon et al., 1992; Bleich et al., 1995, 1998; Kawajiriet al., 2000). There is sufficient evidence for aninteraction between lipoxygenase products and PPARγ,which is involved in cell differentiation and glucosehomeostasis (Shi et al., 2002). Activation of PPARγ withligands such as trogliatzone induces differentiation incolon cancer cells or liposarcoma cells, reversing themalignant phenotype (Hsi et al., 2001). PPARγ ligandsconsistently induce marked growth inhibition andapoptosis in pancreatic cancer cells (Motomura et al.,2000; Eibl et al., 2001). Furthermore, PPARγ has beenreported to be overexpressed in human pancreaticadenocarcinomas. While we could not confirm theseresults, we saw up-regulation of PPARγ in the isletsadjacent to tumors in our own immunohistochemidalstudy (unpublished data) (Motomura et al., 2000). Thecomplex interrelationship between lipoxygenasemetabolism, PPARs, islets, pancreatic cancer and theassociated diabetes is far from clear, but considering theinfluence of this inflammatory pathway on glucosemetabolism as well as pancreatic cancer growth anddevelopment, it is likely to be important. Future studiesin this field are anticipated with interest.

Conclusion

We are still not able to say whether or not islet cellsmay be the origin of pancreatic adenocarcinoma.However, some questions can be answered. The diabetesassociated with pancreatic cancer appears to be aconsequence rather than a risk factor. Islet cell hormonesplay very different roles in pancreatic cancer growth.Insulin is growth promoting, while somatostatin andpancreatic polypeptide generally appear to be growthinhibiting. The current data suggest that at a certainstage, all pancreatic cell types are capable of trans-differentiation or dedifferentiation and can ultimatelydevelop into pancreatic adenocarcinomas. However, thepresence of a multipotent pancreatic stem cell remains asthe possible cell of origin for pancreatic cancer, butpancreatic stem cell research is still in its infancy.Certainly, the topic of “islets and pancreatic cancer”remains an exciting one. The remaining questions willhopefully be answered in the near future. Consideringthe abysmal prognosis of this disease, the need forsignificant progress for these patients is even more

acute.

References

Abbruzzese J.L. (2002). New applications of gemcitabine and futuredirections in the management of pancreatic cancer. Cancer 95, 941-945.

Ahrendt S.A. and Pitt H.A. (2002). Surgical management of pancreaticcancer. Oncology (Huntingt) 16, 725-734.

Arnelo U., Herrington M.K., Theodorsson E., Adrian T.E., ReidelbergerR., Larsson J., Marcusson J., Strommer L., Ding X. and Permert J.(2000). Effects of long-term infusion of anorexic concentrations ofislet amyloid polypeptide on neurotransmitters and neuropeptides inrat brain. Brain Res. 887, 391-398.

Arnelo U., Permert J., Larsson J., Reidelberger R.D., Arnelo C. andAdrian T.E. (1997). Chronic low dose islet amyloid polypeptideinfusion reduces food intake, but does not influence glucosemetabolism, in unrestrained conscious rats: studies using a novelaortic catheterization technique. Endocrinology 138, 4081-4085.

Bardeesy N. and DePinho R.A. (2002). Pancreatic cancer biology andgenetics. Nat. Rev. Cancer 2, 897-909.

Bell R.H., McCullough P.J. and Pour P.M. (1988). Influence of diabeteson susceptibility to experimental pancreatic cancer. Am. J. Surg.155, 159-164.

Bell R.H. and Pour P.M. (1987). Pancreatic carcinogenicity of N-nitrosobis(2-oxopropyl)-amine in diabetic and non-diabetic Chinesehamsters. Cancer Lett. 34, 221-230.

Bell R.H., Sayers H.J., Pour P.M., Ray M.B. and McCullough P.J.(1989). Importance of diabetes in inhibition of pancreatic cancer bystreptozotocin. J. Surg. Res. 46, 515-519.

Bensley R.R. (1911). Studies on the pancreas of the guinea pig. Am. J.Anat. 12, 297.

Bleich D., Chen S., Gu J.L. and Nadler J.L. (1998). The role of 12-lipoxygenase in pancreatic cells. Int. J. Mol. Med. 1, 265-272.

Bleich D., Chen S., Gu J.L., Thomas L., Scott S., Gonzales N.,Natarajan R. and Nadler J.L. (1995). Interleukin-1 beta regulates theexpression of a leukocyte type of 12- lipoxygenase in rat islets andRIN m5F cells. Endocrinology 136, 5736-5744.

Brand R.E., Ding X.Z., Young C.M. and Adrian T.E. (2002). Thespecificity of amylin for the diagnosis of pancreatic adenocarcinoma.Int J Gastrointest. Cancer 31, 123-128.

Burch P.A., Block M., Schroeder G., Kugler J.W., Sargent D.J., BraichT.A., Mailliard J.A., Michalak J.C., Hatfield A.K., Wright K. andKuross S.A. (2000). Phase III evaluation of octreotide versuschemotherapy with 5- fluorouracil or 5-fluorouracil plus leucovorin inadvanced exocrine pancreatic cancer: a North Central CancerTreatment Group study. Clin. Cancer Res. 6, 3486-3492.

Chance W.T., Balasubramaniam A., Zhang F.S., Wimalawansa S.J. andFischer J.E. (1991). Anorexia following the intrahypothalamicadministration of amylin. Brain Res. 539, 352-354.

Chandrasekar B. and Korc M. (1991) Alteration of cholecystokinin-mediated phosphatidylinositol hydrolysis in pancreatic acini frominsulin-deficient rats. Evidence for defective G protein activation.Diabetes 40, 1282-1291.

Chari S.T., Klee G.G., Miller L.J., Raimondo M. and DiMagno E.P.(2001). Islet amyloid polypeptide is not a satisfactory marker fordetecting pancreatic cancer. Gastroenterology 121, 640-645.

Ding X., Flatt P.R., Permert J. and Adrian T.E. (1998). Pancreaticcancer cells selectively stimulate islet beta cells to secrete amylin.

1007

Islets and pancreatic cancer

Gastroenterology 114, 130-138.Ding X.Z., Fehsenfeld D.M., Murphy L.O., Permert J. and Adrian T.E.

(2000). Physiological concentrations of insulin augment pancreaticcancer cell proliferation and glucose utilization by activating MAPkinase, PI3 kinase and enhancing GLUT-1 expression. Pancreas 21,310-320.

Douziech N., Calvo E., Coulombe Z., Muradia G., Bastien J., AubinR.A., Lajas A. and Morisset J. (1999). Inhibitory and stimulatoryeffects of somatostatin on two human pancreatic cancer cell lines: aprimary role for tyrosine phosphatase SHP-1. Endocrinology 140,765-777.

Eibl G., Wente M.N., Reber H.A. and Hines O.J. (2001). Peroxisomeproliferator-activated receptor gamma induces pancreatic cancer cellapoptosis. Biochem. Biophys. Res Commun. 287, 522-529.

Fisher W.E., Boros L.G. and Schirmer W.J. (1996). Insulin promotespancreatic cancer: evidence for endocrine influence on exocrinepancreatic tumors. J. Surg. Res. 63, 310-313.

Fisher W.E., Muscarella P., Boros L.G. and Schirmer W.J. (1998).Gastrointestinal hormones as potential adjuvant treatment ofexocrine pancreatic adenocarcinoma. Int. J. Pancreatol. 24, 169-180.

Fogar P., Basso D., Panozzo M.P., Del Favero G., Briani G., Fabris C.,D'Angeli F., Meggiato T., Ferrara C. and Plebani M. (1993). Cpeptide pattern in patients with pancreatic cancer. Anticancer Res.13, 2577-2580.

Friedman G.D. and van den Eeden S.K. (1993). Risk factors forpancreatic cancer: an exploratory study. Int. J. Epidemiol. 22, 30-37.

Frontoni S., Choi S.B., Banduch D. and Rossetti L. (1991). In vivoinsulin resistance induced by amylin primarily through inhibition ofinsulin-stimulated glycogen synthesis in skeletal muscle. Diabetes40, 568-573.

Fueger B.J., Hamilton G., Raderer M., Pangerl T., Traub T.,Angelberger P., Baumgartner G., Dudczak R. and Virgolini I. (2001).Effects of chemotherapeutic agents on expression of somatostatinreceptors in pancreatic tumor cells. J. Nucl. Med. 42, 1856-1862.

Fujii H., Egami H., Chaney W., Pour P. and Pelling J. (1990). Pancreaticductal adenocarcinomas induced in Syrian hamsters by N-nitrosobis(2-oxopropyl)amine contain a c-Ki-ras oncogene with apoint-mutated codon 12. Mol. Carcinog. 3, 296-301.

Funk C.D. (2001). Prostaglandins and leukotrienes: advances ineicosanoid biology. Science 294, 1871-1875.

Gapstur S.M., Gann P.H., Lowe W., Liu K., Colangelo L. and Dyer A.(2000). Abnormal glucose metabolism and pancreatic cancermortality. JAMA 283, 2552-2558.

Ghadirian P., Lynch H.T. and Krewski D. (2003). Epidemiology ofpancreatic cancer: an overview. Cancer Detect. Prev. 27, 87-93.Gillespie J., Poston G.J., Schachter M. and Guillou P.J. (1992).Human pancreatic cancer cell lines do not express receptors forsomatostatin. Br. J. Cancer 66, 483-487.

Greenlee R.T., Hill-Harmon M.B., Murray T. and Thun M. (2001).Cancer statistics, 2001. CA Cancer J. Clin. 51, 15-36.Greenlee R.T., Murray T., Bolden S. and Wingo P.A. (2000). Cancerstatistics, 2000. CA Cancer J. Clin. 50, 7-33.

Greten F.R., Wagner M., Weber C.K., Zechner U., Adler G. and SchmidR.M. (2001). TGF alpha transgenic mice. A model of pancreaticcancer development. Pancreatology 1, 363-368.

Gullo L. (1999). Diabetes and the risk of pancreatic cancer. Ann. Oncol.10 Suppl 4, 79-81.

Gullo L., Pezzilli R. and Morselli-Labate A.M. (1994). Diabetes and the

risk of pancreatic cancer. Italian Pancreatic Cancer Study Group. N.Engl. J. Med. 331, 81-84.

Heinemann V. (2001). Gemcitabine: progress in the treatment ofpancreatic cancer. Oncology 60, 8-18.

Hennig R., Ding X.Z., Tong W.G., Schneider M.B., Standop J., FriessH., Buchler M.W., Pour P.M. and Adrian T.E. (2002). 5-Lipoxygenase and leukotriene B(4) receptor are expressed inhuman pancreatic cancers but not in pancreatic ducts in normaltissue. Am. J. Pathol. 161, 421-428.

Herrera P.L., Huarte J., Sanvito F., Meda P., Orci L. and Vassalli J.D.(1991). Embryogenesis of the murine endocrine pancreas, earlyexpression of pancreatic polypeptide gene. Development 113, 1257-1265.

Howard T.J. (1996). Pancreatic adenocarcinoma. Curr. Probl. Cancer20, 281-328.

Hruban R.H., Goggins M., Parsons J. and Kern S.E. (2000a).Progression model for pancreatic cancer. Clin. Cancer Res. 6, 2969-2972.

Hruban R.H., Wilentz R.E. and Kern S.E. (2000b). Genetic progressionin the pancreatic ducts. Am. J. Pathol. 156, 1821-1825.

Hsi L.C., Wilson L., Nixon J. and Eling T.E. (2001). 15-lipoxygenase-1metabolites down-regulate peroxisome proliferator- activatedreceptor gamma via the MAPK signaling pathway. J. Biol. Chem.276, 34545-34552.

Hunziker E. and Stein M. (2000). Nestin-expressing cells in thepancreatic islets of Langerhans. Biochem. Biophys. Res. Commun.271, 116-119.

Ikematsu Y., Liu G., Fienhold M.A., Cano M., Adrian T.E., HollingsworthM.A., Williamson J.E., Sanger W., Tomioka T. and Pour P.M. (1997).In vitro pancreatic ductal cell carcinogenesis. Int. J. Cancer 1095-1103.

Jesnowski R., Muller P., Schareck W., Liebe S. and Lohr M. (1999).Immortalized pancreatic duct cells in vitro and in vivo. Ann. NYAcad. Sci. 880, 50-65.

Jimenez R.E., Z'graggen K., Hartwig W., Graeme-Cook F., WarshawA.L. and Fernandez-del Castillo C. (1999). Immunohistochemicalcharacterization of pancreatic tumors induced by dimethyl-benzanthracene in rats. Am. J. Pathol. 154, 1223-1229.

Jonsson J., Carlsson L., Edlund T. and Edlund H. (1994). Insulin-promoter-factor 1 is required for pancreas development in mice.Nature 371, 606-609.

Kawajiri H., Zhuang D., Qiao N., Yoshimoto T., Yamamoto M., Iseki S.and Hamaguchi K. (2000). Expression of arachidonate 12-lipoxygenase in rat tissues: a possible role in glucagon secretion. J.Histochem. Cytochem. 48, 1411-1419.

Kazakoff K., Cardesa T., Liu J., Adrian T.E., Bagchi D., Bagchi M., BirtD.F. and Pour P.M. (1996). Effects of voluntary physical exercise onhigh-fat diet-promoted pancreatic carcinogenesis in the hamstermodel. Nutr. Cancer 26, 265-279.

Kimura W., Morikane K., Esaki Y., Chan W.C. and Pour P.M. (1998).Histologic and biologic patterns of microscopic pancreatic ductaladenocarcinomas detected incidentally at autopsy. Cancer 82,1839-1849.

Kodama T. and Mori W. (1983). Morphological behavior of carcinoma ofthe pancreas. 2. Argyrophil cells and Langerhans' islets in thecarcinomatous tissues. Acta Pathol. Jpn. 33, 483-493.

Kokawa A., Kondo H., Gotoda T., Ono H., Saito D., Nakadaira S.,Kosuge T. and Yoshida S. (2001). Increased expression ofcyclooxygenase-2 in human pancreatic neoplasms and potential for

1008

Islets and pancreatic cancer

chemoprevention by cyclooxygenase inhibitors. Cancer 91, 333-338.Koopmans S.J., van Mansfeld A.D., Jansz H.S., Krans H.M., Radder

J.K., Frolich M., de Boer S.F., Kreutter D.K., Andrews G.C. andMaassen J.A. (1991). Amylin-induced in vivo insulin resistance inconscious rats: the liver is more sensitive to amylin than peripheraltissues. Diabetologia 218-224.

Koshiba T., Hosotani R., Miyamoto Y., Wada M., Lee J.U., Fujimoto K.,Tsuji S., Nakajima S., Doi R. and Imamura M. (1999).Immunohistochemical analysis of cyclooxygenase-2 expression inpancreatic tumors. Int. J. Pancreatol. 26, 69-76.

La Vecchia C., Negri E., Franceschi S., D'Avanzo B. and Boyle P.(1994). A case-control study of diabetes mellitus and cancer risk. Br.J. Cancer 70, 950-953.

Langerhans P. (1869). Beiträge zur mikroskopischen Anatomie derBauchspeicheldrüse. Inaugural Dissertation.

Lardon J., Rooman I. and Bouwens L. (2002). Nestin expression inpancreatic stellate cells and angiogenic endothelial cells. Histochem.Cell Biol. 117, 535-540.

Lechner A., Leech C.A., Abraham E.J., Nolan A.L. and Habener J.F.(2002). Nestin-positive progenitor cells derived from adult humanpancreatic islets of Langerhans contain side population (SP) cellsdefined by expression of the ABCG2 (BCRP1) ATP-binding cassettetransporter. Biochem. Biophys. Res. Commun. 293, 670-674.

Lee K.M., Nguyen C., Ulrich A.B., Pour P.M. and Ouellette M.M. (2003).Immortalization with telomerase of the Nestin-positive cells of thehuman pancreas. Biochem. Biophys. Res. Commun. 301, 1038-1044.

Leighton B. and Cooper G.J. (1988). Pancreatic amylin and calcitoningene-related peptide cause resistance to insulin in skeletal musclein vitro. Nature 335, 632-635.

Lin Y., Tamakoshi A., Wakai K., Kawamura T., Aoki R., Kojima M. andOhno Y. (1998). Descriptive epidemiology of pancreatic cancer inJapan. J. Epidemiol. 8, 52-59.

Liu J., Kazakoff K., Pour P.M. and Adrian T.E. (1998). The intracellularmechanism of insulin resistance in the hamster pancreatic ductaladenocarcinoma model. Pancreas 17, 359-366.

Liu J., Knezetic J.A., Strommer L., Permert J., Larsson J. and AdrianT.E. (2000). The intracellular mechanism of insulin resistance inpancreatic cancer patients. J. Clin. Endocrinol. Metab. 85, 1232-1238.

Lohr M., Muller P., Zauner I., Schmidt C., Trautmann B., Thevenod F.,Capella G., Farre A., Liebe S. and Jesnowski R. (2001).Immortalized bovine pancreatic duct cells become tumorigenic aftertransfection with mutant k-ras. Virchows Arch. 438, 581-590.

Lukinius A., Ericsson J.L., Grimelius L. and Korsgren O. (1992).Ultrastructural studies of the ontogeny of fetal human and porcineendocrine pancreas, with special reference to colocalization of thefour major islet hormones. Dev. Biol. 153, 376-385.

Madsen O.D., Nielsen J.H., Michelsen B., Westermark P., Betsholtz C.,Nishi M. and Steiner D.F. (1991). Islet amyloid polypeptide andinsulin expression are controlled differently in primary andtransformed islet cells. Mol. Endocrinol. 5, 143-148.

Maitra A., Ashfaq R., Gunn C.R., Rahman A., Yeo C.J., Sohn T.A.,Cameron J.L., Hruban R.H. and Wilentz R.E. (2002).Cyclooxygenase 2 expression in pancreatic adenocarcinoma andpancreatic intraepithelial neoplasia: an immunohistochemicalanalysis with automated cellular imaging. Am. J Clin. Pathol. 118,194-201.

McCarty M.F. (2001). Insulin secretion as a determinant of pancreatic

cancer risk. Med. Hypotheses 57, 146-150.Metz S.A., Murphy R.C. and Fujimoto W. (1984). Effects on glucose-

induced insulin secretion of lipoxygenase-derived metabolites ofarachidonic acid. Diabetes 33, 119-124.

Molina M.A., Sitja-Arnau M., Lemoine M.G., Frazier M.L. and SinicropeF.A. (1999). Increased cyclooxygenase-2 expression in humanpancreatic carcinomas and cell l ines: growth inhibit ion bynonsteroidal anti-inflammatory drugs. Cancer Res. 59, 4356-4362.

Motomura W., Okumura T., Takahashi N., Obara T. and Kohgo Y.(2000). Activation of peroxisome proliferator-activated receptorgamma by troglitazone inhibits cell growth through the increase ofp27KiP1 in human. Pancreatic carcinoma cells. Cancer Res 60,5558-5564.

Muscarella P., Knobloch T.J., Ulrich A.B., Casto B.C., Moniaux N.,Wittel U.A., Melvin W.S., Pour P.M., Song H., Gold B., Batra S.K.and Weghorst C.M. (2001). Identification and sequencing of theSyrian Golden hamster (Mesocricetus auratus) p16(INK4a) andp15(INK4b) cDNAs and their homozygous gene deletion in cheekpouch and pancreatic tumor cells. Gene 278, 235-243.

Nathan M.H. and Pek S.B. (1990). Lipoxygenase-generated icosanoidsinhibit glucose-induced insulin release from rat islets. ProstaglandinsLeukot. Essent. Fatty Acids 40, 21-25.

Offield M.F., Jetton T.L., Labosky P.A., Ray M., Stein R.W., MagnusonM.A., Hogan B.L. and Wright C.V. (1996). PDX-1 is required forpancreatic outgrowth and differentiation of the rostral duodenum.Development 122, 983-995.

Ohba S., Nishi M. and Miyake H. (1996). Eating habits and pancreascancer. Int. J. Pancreatol. 20, 37-42.

Ohike N., Jurgensen A., Pipeleers-Marichal M. and Kloppel G. (2003).Mixed ductal-endocrine carcinomas of the pancreas and ductaladenocarcinomas with scattered endocrine cells: characterization ofthe endocrine cells. Virchows Arch. 442, 258-265.

Okami J., Yamamoto H., Fujiwara Y., Tsujie M., Kondo M., Noura S.,Oshima S., Nagano H., Dono K., Umeshita K., Ishikawa O., SakonM., Matsuura N., Nakamori S. and Monden M. (1999).Overexpression of cyclooxygenase-2 in carcinoma of the pancreas.Clin. Cancer Res. 5, 2018-2024.

Oomi K. and Amano M. (1998). The epidemiology of pancreaticdiseases in Japan. Pancreas 16, 233-237.

Orci L., Malaisse-Lagae F., Baetens D. and Perrelet A. (1978).Pancreatic-polypeptide-rich regions in human pancreas. Lancet 2,1200-1201.

Parker S.L. (1996). Cancer statistics. 1996. CA Cancer J. Clin. 46, 5-27.Parlapiano C., Danese C., Marangi M., Campana E., Pantone P.,

Giovanniello T., Zavattaro E. and Sanguigni S. (1999). Therelationship between glycated hemoglobin and polymorphonuclearleukocyte leukotriene B4 release in people with diabetes mellitus.Diabetes Res. Clin. Pract. 46, 43-45.

Pek S.B. and Nathan M.H. (1994). Role of eicosanoids in biosynthesisand secretion of insulin. Diabete Metab. 20, 146-149.

Pek S.B. and Walsh M.F. (1984). Leukotrienes stimulate insulin releasefrom the rat pancreas. Proc. Natl. Acad. Sci. USA 81, 2199-2202.

Pek S.B. and Walsh M.F. (1985). Eicosanoids as regulators ofpancreatic islet hormone secretion. Adv. ProstaglandinThromboxane Leukot. Res. 13, 221-230.

Permert J., Ihse I., Jorfeldt L., von Schenck H., Arnquist H.J. andLarsson J. (1993a). Improved glucose metabolism after subtotalpancreatectomy for pancreatic cancer. Br. J. Surg. 80, 1047-1050.

Permert J., Ihse I., Jorfeldt L., von Schenck H., Arnqvist H.J. and

1009

Islets and pancreatic cancer

Larsson J. (1993b). Pancreatic cancer is associated with impairedglucose metabolism. Eur. J. Surg. 159, 101-107.

Permert J., Larsson J., Westermark G.T., Herrington M.K.,Christmanson L., Pour P.M., Westermark P. and Adrian T.E. (1994).Islet amyloid polypeptide in patients with pancreatic cancer anddiabetes. N. Engl. J. Med. 330, 313-318.

Permert J., Larsson J., Fruin A.B., Tatemoto K., Herrington M.K., vonSchenck H. and Adrian T.E. (1997). Islet hormone secretion inpancreatic cancer patients with diabetes. Pancreas 15, 60-68.

Permert J., Herrington M., Kazakoff K., Pour P.M. and Adrian T.E.(2001). Early changes in islet hormone secretion in the hamsterpancreatic cancer model. Teratog. Carcinog. Mutagen. 21, 59-67.

Peters-Golden M. and Brock T.G. (2001). Intracellularcompartmentalization of leukotriene synthesis: unexpected nuclearsecrets. FEBS Lett. 487, 323-326.

Pettengill O.S., Faris R.A., Bell R.H., Kuhlmann E.T. and LongneckerD.S. (1993). Derivation of ductlike cell lines from a transplantableacinar cell carcinoma of the rat pancreas. Am. J. Pathol. 143, 292-303.

Pour P.M. and Kazakoff K. (1996). Stimulation of islet cell proliferationenhances pancreatic ductal carcinogenesis in the hamster model.Am. J. Pathol. 149, 1017-1025.

Pour P.M. and Lawson T. (1984). Modification of pancreaticcarcinogenesis in the hamster model. XV. Preventive effect ofnicotinamide. J .Natl. Cancer Inst. 73, 767-770.

Pour P.M. and Schmied B.M. (1999). One thousand faces ofLangerhans islets. Int. J. Pancreatol. 25, 181-193.

Pour P., Mohr U., Cardesa A., Althoff J. and Kruger F.W. (1975).Pancreatic neoplasms in an animal model: morphological, biological,and comparative studies. Cancer 36, 379-389.

Pour P.M., Kazakoff K. and Carlson K. (1990). Inhibit ion ofstreptozotocin-induced islet cell tumors and N- nitrosobis(2-oxopropyl)amine-induced pancreatic exocrine tumors in Syrianhamsters by exogenous insulin. Cancer Res. 50, 1634-1639.

Pour P.M., Permert J., Mogaki M., Fujii H. and Kazakoff K. (1993).Endocrine aspects of exocrine cancer of the pancreas. Theirpatterns and suggested biologic significance. Am. J. Clin. Pathol.100, 223-230.

Pour P.M., Weide L., Liu G., Kazakoff K., Scheetz M., Toshkov I.,Ikematsu Y., Fienhold M.A. and Sanger W. (1997). Experimentalevidence for the origin of ductal-type adenocarcinoma from the isletsof Langerhans. Am. J. Pathol. 150, 2167-2180.

Pour P.M., Standop J. and Batra S.K. (2002). Are islet cells thegatekeepers of the pancreas? Pancreatology 2, 440-448.

Regitnig P., Spuller E. and Denk H. (2001). Insulinoma of the pancreaswith insular-ductular differentiation in its liver metastasis--indicationof a common stem-cell origin of the exocrine and endocrinecomponents. Virchows Arch. 438, 624-628.

Rochaix P., Delesque N., Esteve J.P., Saint-Laurent N., Voight J.J.,Vaysse N., Susini C. and Buscail L. (1999). Gene therapy forpancreatic carcinoma: local and distant antitumor effects aftersomatostatin receptor sst2 gene transfer. Hum. Gene Ther. 10, 995-1008.

Rose D.P. (1997). Dietary fat, fatty acids and breast cancer. BreastCancer 4, 7-16.

Sakaki M., Sano T., Hirokawa M., Takahashi M. and Kiyoku H. (2002).Immunohistochemical study of endocrine cells in ductaladenocarcinoma of the pancreas. Virchows Arch. 441, 249-255.

Schmied B., Liu G., Moyer M.P., Hernberg I.S., Sanger W., Batra S. and

Pour P.M. (1999). Induction of adenocarcinoma from hamsterpancreatic islet cells treated with N-nitrosobis(2-oxopropyl)amine invitro. Carcinogenesis 20, 317-324.

Schmied B.M., Ulrich A. Matsuzaki H., Ding X., Ricordi C., Moyer M.P.,Batra S.K., Adrian T.E. and Pour P.M. (2000). Maintenance ofhuman islets in long-term culture. Differentiation 66, 173-180.

Schmied B.M., Ulrich A.B., Friess H., Buchler M.W. and Pour P.M.(2001). The patterns of extrainsular endocrine cells in pancreaticcancer. Teratog. Carcinog. Mutagen. 21, 69-81.

Schneider M.B., Matsuzaki H., Haorah J., Ulrich A., Standop J., DingX.Z., Adrian T.E. and Pour P.M. (2001). Prevention of pancreaticcancer induction in hamsters by metformin. Gastroenterology 120,1263-1270.

Schwarts S.S., Zeidler A., Moossa A.R., Kuku S.F. and Rubenstein A.H.(1978). A prospective study of glucose tolerance, insulin, C-peptide,and glucagon responses in patients with pancreatic carcinoma. Am.J. Dig. Dis. 23, 1107-1114.

Shannon V.R., Ramanadham S., Turk J. and Holtzman M.J. (1992).Selective expression of an arachidonate 12-lipoxygenase bypancreatic islet beta-cells. Am. J. Physiol 263, E828-E836.

Sharma A., Zangen D.H., Reitz P., Taneja M., Lissauer M.E., MillerC.P., Weir G.C., Habener J.F. and Bonner-Weir S. (1999). Thehomeodomain protein IDX-1 increases after an early burst ofproliferation during pancreatic regeneration. Diabetes 48, 507-513.

Shi Y., Hon M. and Evans R.M. (2002). The peroxisome proliferator-activated receptor delta, an integrator of transcriptional repressionand nuclear receptor signaling. Proc. Natl. Acad. Sci USA 99, 2613-2618.

Silverman D.T., Swanson C.A., Gridley G., Wacholder S., GreenbergR.S., Brown L.M., Hayes R.B., Swanson G.M., Schoenberg J.B.,Pottern L.M., Schwartz A.G., Fraumeni J.F. and Hoover R.N. (1998).Dietary and nutritional factors and pancreatic cancer: a case-controlstudy based on direct interviews. J. Natl. Cancer Inst. 90, 1710-1719.

Silverman D.T., Schiffman M., Everhart J., Goldstein A., Lillemoe K.D.,Swanson G.M., Schwartz A.G., Brown L.M., Greenberg R.S.,Schoenberg J.B., Pottern L.M., Hoover R.N. and Fraumeni J.F.(1999). Diabetes mellitus, other medical conditions and familialhistory of cancer as risk factors for pancreatic cancer. Br. J. Cancer80, 1830-1837.

Slack J.M. (1995). Developmental biology of the pancreas.Development 121, 1569-1580.

Stefan Y., Grasso S., Perrelet A. and Orci L. (1982). The pancreaticpolypeptide-rich lobe of the human pancreas: definitive identificationof its derivation from the ventral pancreatic primordium [letter].Diabetologia 23, 141-142.

Takeda Y. and Escribano M.J. (1991). Effects of insulin andsomatostatin on the growth and the colony formation of two humanpancreatic cancer cell lines. J. Cancer Res. Clin. Oncol. 117, 416-420.

Tang C., Biemond I., Verspaget H.W., Offerhaus G.J. and Lamers C.B.(1998). Expression of somatostatin receptors in human pancreatictumor. Pancreas 17, 80-84.

Tucker O.N., Dannenberg A.J., Yang E.K., Zhang F., Teng L., Daly J.M.,Soslow R.A., Masferrer J.L., Woerner B.M., Koki A.T. and FaheyT.J. (1999). Cyclooxygenase-2 expression is up-regulated in humanpancreatic cancer. Cancer Res. 59, 987-990.

Ulrich A.B., Schmied B.M., Matsuzaki H., El Metwally T., Moyer M.P.,Ricordi C., Adrian T.E., Batra S.K. and Pour P.M. (2000).

1010

Islets and pancreatic cancer

Establishment of human pancreatic ductal cells in a long-termculture. Pancreas 21, 358-368.

Ulrich A.B., Schmied B.M., Standop J., l Schneider M.B. and Pour P.M.(2002). Pancreatic cell lines: a review. Pancreas 24, 111-120.

Valerio A., Basso D., Brigato L., Ceolotto G., Baldo G., Tiengo A. andPlebani M. (1999). Glucose metabolic alterations in isolated andperfused rat hepatocytes induced by pancreatic cancer conditionedmedium: a low molecular weight factor possibly involved. Biochem.Biophys. Res. Commun. 257, 622-628.

Wagner M., Greten F.R., Weber C.K., Koschnick S., Mattfeldt T.,Deppert W., Kern H., Adler G. and Schmid R.M. (2001). A murinetumor progression model for pancreatic cancer recapitulating thegenetic alterations of the human disease. Genes Dev. 15, 286-293.

Walsh M.F. and Pek S.B. (1984a). Effects of lipoxygenase andcyclooxygenase inhibitors on glucose- stimulated insulin secretionfrom the isolated perfused rat pancreas. Life Sci. 34, 1699-1706.

Walsh M.F. and Pek S.B. (1984b). Possible role of endogenousarachidonic acid metabolites in stimulated release of insulin andglucagon from the isolated, perfused rat pancreas. Diabetes 33,929-936.

Wang F., Larsson J., Adrian T.E., Gasslander T. and Permert J. (1998).In vitro influences between pancreatic adenocarcinoma cells andpancreatic islets. J. Surg. Res. 79, 13-19.

Westermark P., Wernstedt C., Wilander E., Hayden D.W., O'Brien T.D.and Johnson K.H. (1987). Amyloid fibrils in human insulinoma andislets of Langerhans of the diabetic cat are derived from aneuropeptide-like protein also present in normal islet cells. Proc.Natl. Acad. Sci. USA 84, 3881-3885.

Wilding J.P., Khandan-Nia N., Bennet W.M., Gilbey S.G., Beacham J.,Ghatei M.A. and Bloom S.R. (1994). Lack of acute effect of amylin

(islet associated polypeptide) on insulin sensit ivity duringhyperinsulinaemic euglycaemic clamp in humans. Diabetologia 37,166-169.

Woutersen R.A., Appel M.J., Garderen-Hoetmer A. and Wijnands M.V.(1999). Dietary fat and carcinogenesis. Mutat. Res. 443, 111-127.

Yamamoto S., Ishii M., Nakadate T., Nakaki T. and Kato R. (1983).Modulation of insulin secretion by lipoxygenase products ofarachidonic acid. Relation to lipoxygenase activity of pancreaticislets. J. Biol. Chem. 258, 12149-12152.

Yamashita M., Kushihara M., Hirasawa N., Takasaki W., Takahagi H.,Takayanagi M. and Ohuchi K. (2000). Inhibition by troglitazone ofthe antigen-induced production of leukotrienes in immunoglobulin E-sensitized RBL-2H3 cells. Br. J. Pharmacol. 129, 367-373.

Yeo C.J. and Cameron J.L. (1999). Improving results ofpancreaticoduodenectomy for pancreatic cancer. World J. Surg. 23,907-912.

Young D.A., Deems R.O., Deacon R.W., McIntosh R.H. and Foley J.E.(1990). Effects of amylin on glucose metabolism and glycogenolysisin vivo and in vitro. Am. J. Physiol. 259, E457-E461.

Yuan S., Rosenberg L., Paraskevas S., Agapitos D. and Duguid W.P.(1996). Transdifferentiation of human islets to pancreatic ductal cellsin collagen matrix culture. Differentiation 61, 67-75.

Zulewski H., Abraham E.J., Gerlach M.J., Daniel P.B., Moritz W., MullerB., Vallejo M., Thomas M.K. and Habener J.F. (2001). Multipotentialnestin-positive stem cells isolated from adult pancreatic isletsdifferentiate ex vivo into pancreatic endocrine, exocrine, and hepaticphenotypes. Diabetes 50, 521-533.

Accepted April 6, 2004

1011

Islets and pancreatic cancer