role of the basic helix-loop-helix transcription factor...

Role of the Basic Helix-Loop-Helix Transcription Factor p48in the Differentiation Phenotype of Exocrine PancreasCancer Cells1

Teresa Adell, Alicia Gomez-Cuadrado,Anouchka Skoudy, Olive S. Pettengill,Daniel S. Longnecker, and Francisco X. Real2

Unitat de Biologia Cel.lular i Molecular, Institut Municipal d’InvestigacioMedica, Universitat Pompeu Fabra, 08003 Barcelona, Spain [T. A.,A. G-C., A. S., F. X. R.], and Department of Pathology, DartmouthMedical School, Lebanon, New Hampshire 03756 [O. S. P., D. S. L.]

AbstractThe majority of human pancreatic adenocarcinomasdisplay a ductal phenotype; experimental studiesindicate that tumors with this phenotype can arise fromboth acinar and ductal cells. In normal pancreas acinarcells, the pancreas transcription factor 1 transcriptionalcomplex is required for gene expression. Pancreastranscription factor 1 is a heterooligomer of pancreas-specific (p48) and ubiquitous (p75/E2A and p64/HEB)basic helix-loop-helix proteins. We have examined therole of p48 in the phenotype of azaserine-induced ratDSL6 tumors and cancers of the human exocrinepancreas. Serially transplanted acinar DSL6 tumorsexpress p48 whereas DSL6-derived cell lines, and thetumors induced by them, display a ductal phenotypeand lack p48. In human pancreas cancer cell lines andtissues, p48 is present in acinar tumors but not inductal tumors. Transfection of ductal pancreas cancerswith p48 cDNA did not activate the expression ofamylase nor a reporter gene under the control of therat elastase promoter. In some cell lines, p48 wasdetected in the nucleus whereas in others it wascytoplasmic, as in one human acinar tumor. Togetherwith prior work, our findings indicate that p48 isassociated with the acinar phenotype of exocrinepancreas cancers and it is necessary, but notsufficient, for the expression of the acinar phenotype.

IntroductionThe majority of tumors arising in human pancreas derive fromthe exocrine tissue and are classified as “ductal-type” on thebasis of histological appearance (1). Our laboratory and oth-ers have shown that these tumors express molecular mark-ers characteristic of normal ductal cells (2). Among them arecytokeratin polypeptides 7, 19, and 4 (3–5), MUC3 andMUC5B mucins (5–7), dipeptidylpeptidase IV (8), and thecystic fibrosis transmembrane regulator (9). In addition,these tumors lack expression of acinar markers, althoughoccasional reports have described such expression (10). Atthe ultrastructural level, cell lines derived from human pan-creas cancers resemble ductal cells and lack the main fea-tures of acinar cells, i.e., zymogen granules (2, 11).

The target cell for carcinogenesis in the human pancreas,leading to the common ductal-type adenocarcinoma, is notknown. Several candidates have been proposed (discussedin Refs. 12 and 13): (a) a multipotential stem cell presumablylocated in the pancreatic ducts; (b) acinar cells that lose theirdifferentiated properties and acquire a ductal-type pheno-type representing either dedifferentiation or transdifferentia-tion; (c) islet-derived cells that acquire a ductal phenotype;and (d) pancreatic ductal cells. The elucidation of this issueis hampered by the fact that it is not possible to followhistological changes in the human pancreas sequentially.Nevertheless, several observations suggest that pancreaticductal cells are the target for carcinogens in humans: (a)severe dysplasia, a general hallmark of high risk for neopla-sia, is rarely observed in acinar cells, even in tissue frompatients with pancreas cancer (14); (b) ductal cell hyperpla-sia, flat or papillary, can often be identified in tissue frompatients with pancreas cancer (15). However, similar histo-logical changes can also be identified in tissue from patientswithout cancer, in particular in old individuals and in patientswith chronic pancreatitis (15, 16); and (c) recent evidenceshows that some of the genetic lesions that are associatedwith pancreas cancer can also be detected in putative pre-neoplastic lesions, such as flat and papillary ductal cell hy-perplasia. For example, mutations in codon 12 of K-ras occurin a high proportion of pancreas cancers (17) and have alsobeen reported in ductal cell hyperplasia associated withchronic pancreatitis and pancreas cancer (18, 19). Suchevidence may not be considered strong because the muta-tion found in the tumor often, but not always, corresponds tothe mutation found in putative preneoplastic lesions (18–21)and because K-ras mutations are restricted to a few codons,thus increasing the likelihood that the presence of the samemutation in putative preneoplastic lesions and in tumor froma given patient may be attributable to chance. A strongerevidence comes from the study by Moskaluk et al. (21),showing that in two cases with pancreas cancer, putative

Received 9/14/99; revised 12/20/99; accepted 2/4/00.The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indi-cate this fact.1 Partially supported by Grant SAF94-0971 from the Comision Intermin-isterial de Ciencia y Tecnologıa, Grant PM97-0077 from the DireccionGeneral de Ensenanza Superior e Investigacion Cientıfica, Grant SGR-00433 Generalitat de Catalunya, and Grant BMH4-CT98.3085 from theBiomed program. T. A. and A. S. were supported by personal grants fromGeneralitat de Catalunya.2 To whom requests for reprints should be addressed, at Unitat de Bio-logia Cel.lular i Molecular, Institut Municipal d’Investigacio Medica, carrerdel Dr. Aiguader, 80, 08003 Barcelona, Spain. Phone 34-93-2257586; Fax:34-93-2213237; E-mail [email protected].

137Vol. 11, 137–147, March 2000 Cell Growth & Differentiation

preneoplastic lesions (in one case a flat pancreatic intraduc-tal lesion and in another case a papillary lesion with severeatypia) harbored the same mutation in p16INK4A as the tumor.Because mutations in p16 show a wider spectrum than mu-tations in K-ras, this finding is unlikely to result from chance.

Experimental models of pancreas cancer in rodents haveshown consistent genetic and phenotypic characteristicsand can be of help in the understanding of human pancreascancer (12, 13). In the Syrian hamster, nitrosamines induceductal-type tumors that morphologically resemble the com-mon human pancreas cancer and harbor K-ras and p53mutations as human tumors (22). In the rat, azaserine in-duces acinar cell hyperplasia and acinar tumors lacking K-ras mutations (23). From these tumors, several useful celllines have been derived: AR42J cells, displaying mixed aci-nar and neuroendocrine features, have been extensivelyused in studies of the physiology and regulation of geneexpression of acinar cells (24, 25); ARIP cells were derivedfrom the same tumor as AR42J but they lack these differen-tiation features (25). Pettengill et al. (26) have derived two celllines that were independently derived from a transplantableacinar cell carcinoma DSL6, established from a primary car-cinoma of the pancreas induced by azaserine in Lewis rats.After 1–2 weeks in culture, exocrine enzyme productionceased and the long-term cell lines obtained no longer dis-played acinar features. Upon s.c. injection in rats, DSL6Acells produced solid tumors containing duct-like structuresand a dense stroma; DSL6B produced cells with mixed glan-dular, squamous, and mucinous areas. In both cell lines,ultrastructural analysis revealed the loss of zymogen gran-ules (26). In addition, transgenic mice in which c-myc isexpressed in acinar cells under the control of the elastasepromoter develop tumors that can progress from an acinar toa ductal phenotype (27). These studies support the notionthat ductal-like tumors with a phenotype very similar to thatof the common pancreatic adenocarcinoma found in humanscan evolve from acinar tumors.

The expression of acinar cell-specific genes, such asamylase and elastase, is under the control of a pancreas-specific transcriptional complex designated PTF13, whichbinds to a bipartite site in the A element of the regulatoryregion of these genes (28) and whose activity closely paral-lels the expression of pancreas-specific gene products dur-ing development (29). PTF1 is constituted by three differentbHLH transcription factors: p64 and p75 (30), which areubiquitous, and p48 whose expression is restricted to thepancreas (31). p64 is the product of REB, and p75 is theproduct of E2A (32). Although p48 and p64 bind to DNA, p75is involved in the transport of the p48/p64 heterodimer to thenucleus (33). The bHLH region of p48 shares sequence ho-mology with myoD (31), a transcription factor that specifiesactivation of myocyte differentiation in nonmuscle cells (34).Such homology, together with the finding that p48 expres-sion is restricted to the exocrine pancreas, suggests that p48

may be involved in the activation of an acinar differentiationprogram. However, acinar gene expression also requires thecontribution of other tissue-restricted transcription factorssuch as HNF-3b (35) and Pdx-1 (36).

In this study, we have addressed two questions: (a)whether the phenotype of rat and human exocrine pancreatictumors is related to the expression of p48; and (b) whetherp48 is able to instruct an acinar differentiation program inpancreas cancer cells displaying a ductal phenotype. Ourfindings, together with previous work (31), provide evidencethat p48 is necessary, but not sufficient, for the activation ofthe acinar phenotype in ductal cells, even when they arederived from cells with acinar differentiation potential. Fur-thermore, we show that under certain circumstances, p48 isretained in the cytoplasm, thus losing its ability to form thePTF1 complex and contribute to the activation of acinargenes.

ResultsThe Azaserine-induced Rat Pancreas Cancer Model.Azaserine induces focal acinar cell hyperplasia and acinarcell tumors in rats. Serially transplanted acinar tumors in-duced by azaserine maintain acinar features (Fig. 1A). Pet-tengill et al. (26) have shown that when these tumors areplaced in culture, loss of acinar features and acquisition ofductal characteristics take place. In this way, two independ-ent cell lines have been established, DSL6A and DSL6B froma transplanted tumor. When these cells are injected s.c. ororthotopically into male Lewis rats, ductal-like tumors de-velop that are morphologically very similar to the commonductal adenocarcinoma occurring in humans (Fig. 1, C andE). At the ultrastructural level, the serially transplantedtumors have acinar features and contain cells with elec-tron-dense secretion granules (Fig. 1B), whereas the tu-mors induced by DSL6A and DSL6B cell lines displayepithelial features but lack exocrine secretion granules(Fig. 1, D and F).

Expression of the PTF1 Complex Components in Aza-serine-induced Rat Pancreatic Tumors and Derived CellLines. The PTF1 complex has been shown to be involved inthe activation of expression of acinar cell-specific genes(28–31). Therefore, we set out to examine the expression ofPTF1 components in the tumors and cell lines describedabove. To this end, we raised an antiserum against thepancreas-specific component p48.

Production of p48-specific, Affinity-purified Rabbit An-tibodies. Rabbit polyclonal serum against rat p48 producedin Escherichia coli was first preabsorbed with nuclear ex-tracts from a mixture of rat tissues (see “Materials and Meth-ods”) and subsequently purified by affinity chromatographywith His-tagged rat p48. Fig. 2A shows the reactivity of theaffinity-purified immunoglobulin fraction with nuclear ex-tracts from AR42J cells and nuclear extracts from a variety ofnormal rat tissues. A band of Mr ;48,000 was specificallyidentified in AR42J cells and in normal pancreas tissue.There was no reactivity with nuclear or cytoplasmic extractsfrom nonpancreatic tissues. To confirm the specificity of theantiserum, immunohistochemical assays on frozen sectionsof normal rat and human tissues were performed (Fig. 2B). A

3 The abbreviations used are: PTF1, pancreas transcription factor 1;bHLH, basic helix-loop-helix; HNF, hepatocyte nuclear factor; hGH, hu-man growth hormone; RT-PCR, reverse transcription-PCR.

138 Differentiation of Exocrine Pancreatic Tumors

nuclear pattern of staining was identified in the majority ofacinar cells though the intensity of staining showed somevariation from cell to cell. Acinar staining could be inhibitedby preincubation of the antibody with purified p48. All ductalcells were consistently unreactive with anti-p48 antibodies;double labeling with antibodies detecting p48 and antibodiesdetecting cytokeratins 7 and 19, which are restricted tocentroacinar and ductal cells in the pancreas (3–5), revealedno coexpression (data not shown). p48 was not detected inislet cells (Fig. 2B). Apart from the pancreas, there was noreactivity with a large panel of normal tissues: esophagus,stomach, small bowel, colon, gallbladder, liver, breast, ovary,cervix, larynx, trachea, lung, kidney, and thyroid, indicatingthat p48 is selectively expressed in normal acinar cells. Sim-ilar results were obtained using normal rat and human tis-sues.

Expression of PTF1 Complex Components in Azaser-ine-induced Rat Pancreatic Tumors. Expression of p48transcripts was analyzed by RT-PCR; p48 was detectedusing Western blotting and immunohistochemistry. AR42Jcultured cells, displaying acinar features, and normal ratpancreas tissue were used as controls.

p48 mRNA was detected in AR42J cells but not in ARIP,DSL6A, or DSL6B cells. p48 transcripts were detected in thetumors that were serially transplanted and in those inducedby DSL6B cells but not in tumors induced by DSL6A cells(Fig. 3A). p48 protein was exclusively detected in seriallytransplanted tumors (Fig. 3B). Amylase was detected onlywhenever p48 protein was also detectable; in tumors fromDSL6B cells, low levels of p48 mRNA were present, but nop48 protein was detected by western blotting (Fig. 3B). Using

immunohistochemistry, p48 was detected in the nucleus ofcancer cells in serially transplanted tumors, and amylase wasweakly detected, mainly in luminal areas (Fig. 4, A and B). Bycontrast, neither p48 nor amylase were detected in tumorsinduced by cultured DSL6A or DSL6B cells (Fig. 4, C–F).Altogether, these results indicate that loss of acinar featuresin cultured tumor cells and in tumor tissues is associated withthe loss of expression of p48.

To determine whether the other components of the PTF1complex were expressed in DSL6 cells and tumors, RT-PCRwas used. Transcripts of E2A and REB were detected in allcell lines and tumor tissues examined (Fig. 3A), indicatingthat lack of expression of these ubiquitous bHLH transcrip-tion factors does not occur during the loss of the acinarphenotype.

Expression of the PTF1 Components in Human Pan-creas Tumors. To examine p48 transcript expression inhuman pancreas tissue and cultured cells, a partial cDNAencoding the bHLH domain was isolated by RT-PCR usingdegenerate oligonucleotides. PCR products were cloned,and both strands of DNA were sequenced. The comparisonof the cDNA sequence coding for the bHLH region of rat andhuman p48 is shown in Fig. 5. The deduced amino acidsequence in this region is identical for rat and human p48,and there is 79% nucleotide identity; all differences betweenthe two sequences correspond to third nucleotides ofcodons.

Expression of p48 in a panel of human pancreas cancercell lines and in normal pancreas tissue was analyzed byRT-PCR, Western blotting, and immunohistochemistry (Fig.6). p48 transcripts were detected in normal pancreas tissue

Fig. 1. The DSL6 rat pancreas cancer model. Light (A, C, and E) and electron microscopy (B, D, and F) analyses of a serially transplanted tumor showingglandular acinar differentiation (A and B) and the tumors induced by s.c. injection of DSL6A (C and D) and DSL6B (E and F) cells displaying solid and ductaldifferentiation. H&E stain (A, C, and E) was used. A, 3100; C and E, 3200; B and D, 33900; F, 34600.

139Cell Growth & Differentiation

but were undetectable, using a nested PCR, in all humanpancreas cancer cell lines displaying a ductal phenotypeexamined (n 5 8; Fig. 6A). In agreement with these findings,p48 was not detected in cell lysates of cultured pancreascancer cells (Fig. 6B). Using immunohistochemistry on frozentissue sections, p48 was absent from all ductal pancreasadenocarcinomas examined (n 5 18; Fig. 7B). By contrast,anti-p48 antibodies showed strong reactivity with the three

acinar pancreatic tumors examined. In two cases, classifiedas “predominantly acinar-glandular” and “predominantly sol-id,” all cells showed nuclear staining (Fig. 7C); in one addi-tional tumor, classified as “mixed acinar-solid,” p48 wasdetected exclusively in the cytoplasm of 30% of tumor cells,and it was detected in the nucleus of a small proportion oftumor cells (,5%; Fig. 7D). In these tumors, amylase expres-sion was analyzed using immunohistochemistry. In the aci-

Fig. 2. Specificity of polyclonal antibody raised againstrat p48. A, expression of p48 in normal rat tissues ana-lyzed by Western blotting using nuclear extracts. P, pan-creas; D, duodenum; S, spleen; Li, liver; Lu, lung; M,skeletal muscle. Reactivity with actin shown as a controlof protein loading. B, analysis of the reactivity of anti-p48 antibodies with normal human tissues using immu-noperoxidase and a light hematoxylin counterstaining.Selective reactivity with the nucleus of pancreatic acinarcells, but not with normal ducts or islet cells, is ob-served. a and b, pancreas; c, stomach; d, larynx; e, lung;f, brain. A, acini; D, duct; I, islet. 3200.

Fig. 3. Expression of PTF1 complex componentsin rat pancreas cancer cells. A, RT-PCR analysis oftranscripts of p48, E2A, and REB in cultured ratpancreas cancer cells and in DSL6-derived tumortissues. B, Western blotting analysis of p48 andamylase (Amyl) expression in cultured rat pancreascancer cells and in DSL6-derived tumor tissues. T,tumor; T from T, serially transplanted tumor; NPT,normal pancreas tissue. Normal pancreas tissueand AR42J cells are shown as controls in bothtypes of assays.


nar-glandular tumor, strong expression of amylase was ob-served; by contrast, in the mixed acinar-solid tumor, amylasewas detected in a very low proportion of cells (,5%; data notshown). Anti-p48 antibodies did not react with any of thenonpancreatic tumors analyzed (Table 1).

The expression of the other components of the PTF1 tran-scription complex in cell lines derived from ductal-type pan-creas cancers was examined by RT-PCR. The two isoforms

of E2A, E12 and E47, and HEB were detected in all cell linesexamined (Fig. 6C). Therefore, the only component whoseexpression is absent in ductal pancreatic tumors is p48.

Is Transfection of p48 cDNA into Ductal-like Cells Ableto Induce an Acinar Phenotype? The DSL6 cells and tu-mors provide an excellent model to analyze whether p48 canactivate an acinar differentiation program because the duc-tal-like DSL6A and DSL6B cells are derived from an acinar

Fig. 5. Alignment of the sequences of the bHLH region of human and rat p48 cDNA. A 79% nucleotide and 100% amino acid identity is observed. H,human; R, rat.

Fig. 4. Immunohistochemical analysis of theexpression of p48 and amylase in DSL6-de-rived tumor tissues; a light hematoxylin coun-terstaining was performed not to mask nuclearstaining with anti-p48 antibodies. A, C, and E,p48; B, D, and F, amylase; A and B, seriallytransplanted DSL6 tumor; C and D, DSL6A-induced tumor; E and F, DSL6B-induced tu-mor. A and B, 3100; C–F, 3200.


tumor. To this end, DSL6A and DSL6B cells were transientlytransfected with the rat p48 cDNA, and the activity of thePTF1 complex was analyzed by monitoring amylase expres-sion in cell lysates, amylase activity released to the culturemedium, and the activity of a reporter plasmid in which thehGH cDNA is under the control of a 6-mer of the A elementof the rat elastase gene (37). DSL6B cells could not betransfected using Lipofectamine, but p48 was detected intransfected DSL6A cells as well as in control AR42J andARIP cells (Fig. 8A). Amylase was not detected in the super-natant or lysates of DSL6A and ARIP cells transfected withp48 cDNA (Fig. 8A), and no activity of the hGH reporter couldbe demonstrated (Fig. 8B). By contrast, transfection ofAR42J cells led to an increase in nuclear p48 levels and to a2–4-fold increase in the activity of the hGH reporter (Fig. 8B).In addition, a small but reproducible increase in intracellularand secreted amylase was observed. The increase in se-creted amylase was not statistically significant, likely be-cause of the low efficiency of transient transfections in

AR42J cells. These results were confirmed in more than threeindependent assays.

The effect of transfection of rat p48 cDNA into humanRWP-1 pancreas cancer cells was also examined. p48 ex-pression was detected in transfectants, but amylase, as wellas the activity of the reporter construct, was undetectable(Fig. 8, A and B).

To examine whether lack of acinar gene activation andreporter activity could be attributable to mislocalization ofp48, as observed in one of the acinar tumors, the subcellulardistribution of p48 in transfectants was examined by immu-nohistochemistry. p48 was found in the nucleus of all trans-fected AR42J and ARIP cells (Fig. 8C, b and c) and in thecytoplasm of all transfected DSL6A and RWP-1 cells (Fig.8C, a and d).

DiscussionIn this work, we show that down-regulation of p48 wasassociated with the change of an azaserine-induced acinarcell carcinoma to a ductal phenotype and that human ductal-type pancreas cancers lack p48 expression. By contrast,expression of the other two members of the exocrine pan-creas-specific transcription complex PTF1 is maintainedboth in rat and human ductal tumors. These results indicatethat p48 is selectively absent from ductal-type tumors.

Until recently, p48 was thought to participate exclusively inexocrine pancreas differentiation. However, inactivation ofthe p48 gene in mice by homologous recombination leads toan apancreatic phenotype characterized by the lack of anexocrine pancreas and the presence of a reduced number ofindividual endocrine cells in the spleen (38). These findingsindicate that p48 plays a role both early and late in pancreasdevelopment and differentiation. However, little is knownabout the genetic mechanisms regulating the expression ofp48. The p48 promoter has been cloned and sequenced, butits activity is not restricted to pancreatic cells, and putativetissue-specific control elements have not been identified inas much as 10 kb of 59-flanking region. (39). Several expla-nations can account for the finding that, upon in vitro culture,DSL6 cells displaying an acinar phenotype lose p48 expres-sion: culture may select for populations lacking p48 expres-sion and having some growth advantage; alternatively, themaintenance of p48 expression may require cellular interac-tions with the mesenchyme or with neural cells that cannotbe maintained in vitro. In support of the latter hypothesis isthe observation that cultures of normal human exocrine pan-creas undergo an acinar-to-ductal phenotypic switch in vitrothat is also accompanied by the loss of expression of p48(40, 41). In these cells, a selective outgrowth of cells lackingp48 cannot be proposed because normal pancreatic cellsundergo limited replication in vitro (40). There is extensiveevidence that mesenchymal-epithelial interactions play amajor role in development (42) and differentiation (43). Forexample, the LIM-homeodomain protein Isl-1 is expressed inall postmitotic islet cells as well as in mesenchymal cellssurrounding the dorsal but not the ventral evagination of thegut endoderm during development. Inactivation of Isl-1 in themouse results in the lack of Isl-1-expressing dorsal mesen-chyme, lack of exocrine cell differentiation in the dorsal but

Fig. 6. p48 expression in human pancreas cancer cell lines displaying aductal phenotype. Panel A, RT-PCR analysis of p48 transcripts in pan-creas cancer cell lines. K-ras transcript amplification is shown as a controlof RNA quality and amount. Lane 1, AsPC-1; Lane 2, Capan-2; Lane 3,HPAF; Lane 4, SK-PC-1; Lane 5, IMIM-PC-1; Lane 6, IMIM-PC-2; Lane 7,MZPC-1; Lane 8, RWP-1. Panel B, Western blotting analysis of p48expression in pancreas cancer cells (IMIM-PC-2 and IMIM-PC-1) and incontrol colon cancer cells (HT-29). Normal human pancreas tissue (NPT)and AR42J cells are shown as positive controls. C, control for proteinloading. Panel C, RT-PCR analysis of the expression of HEB and E2Atranscripts in human pancreas cancer cells. Results are shown for the twovariants of E2A, E47 and E12.


not in the ventral pancreatic bud, and lack of endocrine cellsin all of the pancreas. In vitro reconstitution experiments haveshown that the wild-type pancreatic mesenchyme can sup-port the normal development of the dorsal exocrine pancreasfrom Isl-12/2 mice (43). A search for regulatory motifs in the10 kb upstream of the p48 gene has failed to provide cluesabout tissue-specific transcription factors/complexes thatmay be involved in its regulation, although a putative Pdx-1binding site has been identified (39). The molecular mecha-nisms underlying the down-regulation of p48 gene expres-sion in tumors remain unknown, although at least in thehuman tumors, the p48 gene is generally retained.4

Our findings support the contention that p48 is necessary,but not sufficient, for the activation of acinar genes in pan-creatic cells. This conclusion, which is also supported by invitro studies of promoter regulation (35, 36) and by transfec-tion of an antisense construct of p48 cDNA in AR42J cells(31), is drawn from our transfection studies using both rat

and human pancreas cancer cells. AR42J cells, which spon-taneously express amylase and other acinar genes (25), areequipped with the transcription factor machinery necessaryfor their expression, and transfection with p48 cDNA is ac-companied by enhanced expression both in transient andstable assays (this work and data not shown). By contrast,transfection of p48 cDNA into other pancreatic cells did notresult in the activation of the expression of amylase nor areporter construct containing the A element of the elastasepromoter, indicating lack of formation of an active PTF1complex. Two types of mechanisms seem to contribute tothe lack of effects of p48 overexpression. In DSL6A andRWP-1 cells, p48 was undetectable in the nucleus. p75,encoded by the ubiquitously expressed E2A gene, hasbeen shown to be responsible for the nuclear import of thep48/p64 complex (33); however, E2A is expressed in bothcells. Similarly, the REB/HEB gene is also expressed inboth lines. Thus, the molecular basis of the cytoplasmicaccumulation of p48 remains to be elucidated. Id proteins,initially identified as inhibitors of differentiation of muscleand lymphoid cells (44, 45), are candidates to play a rolein the subcellular distribution of p48. Id proteins, of whichfour members have been described thus far, contain anHLH domain that enables their dimerization with bHLHfactors but lack the basic domain that allows DNA binding,therefore sequestering bHLH-type transcription factorsand acting in a dominant-negative fashion (44 – 46). Id1has been directly implicated in the suppression of mam-mary cell differentiation (47). There is currently no directevidence that Id proteins interact with p48 nor that theycan block its transcriptional activity. In addition to them,other yet unidentified proteins might interact with p48 andretain it in the cytoplasm.

The pathophysiological relevance of the subcellular distri-bution of p48 is underlined by our findings in human acinar

4 T. Adell, X. Molero, A. Skoudy, M. A. Padilla and F. X. Real, unpublishedobservations.

Table 1 Immunohistochemical analysis of p48 expression in humanadenocarcinomas

Tumor type Positive/Total tested

PancreasDuctal 0/17Cystadenocarcinoma 0/1Acinar 3/3

Colon 0/5Breast 0/2Lung 0/2Biliary tract 0/2Ovary 0/1

Fig. 7. p48 expression in humanpancreatic tissues using immu-noperoxidase; tissues are lightlycounterstained with hematoxylin.A, normal pancreas; B, ductal-typeadenocarcinoma; C, acinar-typeadenocarcinoma showing homoge-neous nuclear staining; D, acinar-type adenocarcinoma showing het-erogeneous, predominantly cyto-plasmic staining. T, tumor cells. Ar-rowheads, areas of cytoplasmicdistribution of p48. 3200.


tumors. Unlike normal acinar cells, one acinar tumor showedp48 expression exclusively in the cytoplasm. Furthermore,preliminary data indicate that p48 preferentially accumulatesin the cytoplasm of acinar/ductular complexes in areas ofchronic pancreatitis associated with pancreas cancer.4 Theabnormal subcellular distribution of p48 might contribute tothe disregulation of acinar gene expression and pancreaticinsufficiency.

In ARIP cells, p48 did localize to the nucleus, but it did notactivate acinar gene expression nor a reporter construct forthe PTF1 complex. It has been shown that expression ofacinar enzymes in AR42J cells requires the expression of thewinged helix-loop-helix factor HNF-3b (35). Indeed, ARIPcells express low levels of HNF-3b. Because transfectionefficiency for these cells is very low, we have not been ableto conclusively show the precise role of HNF-3b by cotrans-fecting its cDNA with that of p48. Other mechanisms mayaccount for the lack of activation of acinar genes. For exam-ple, the histone acetyltransferase PCAF and p300/CBP pro-mote both MyoD-dependent transcription and myogenic dif-ferentiation in muscle cells (48).

It is currently not known whether p48 may play a role intumor progression. Acinar-to-ductal conversion has beendemonstrated in other tumors of the exocrine pancreas inrodents, most notably the Ela-myc transgenic mouse. Inthese animals, acinar tumors progress to develop a ductalphenotype similar to that of the majority of human exocrine

pancreas cancers (27). It is conceivable that loss of acinarfeatures and p48 expression might favor tumor progression,although this possibility has not been examined in detail.Studies on the role of the bHLH transcription factor MyoD inmuscle differentiation support such contention. MyoD in-duces the expression of the cyclin-dependent kinase inhib-itor p21 at the transcriptional level, leading to an arrest in G1

that allows the activation of a cell differentiation program innormal cells (49–51). Such effects are abrogated in a highproportion of tumor cells (52), despite that MyoD is ex-pressed in the majority of rhabdomyosarcomas (53, 54).Furthermore, pRb favors both cell cycle exit and the expres-sion of muscle-specific genes by cooperating with Myo-D,whereas in the absence of pRb, myogenic transcription fac-tors are inactive (55).

Finally, p48 is an excellent marker of the acinar cell differ-entiation in the pancreas. Although its selectivity is unlikely tobe of clinical interest in the setting of human exocrine pan-creatic cancers because acinar tumors are rare, the study ofthe expression, subcellular distribution, and transcriptionalactivity of p48 should shed light on the molecular mecha-nisms involved in the diseases of the exocrine pancreas.

Materials and MethodsCells and Tissues. The following human pancreas cancer cell lines wereused: AsPC-1, Capan-2, and HPAF cells, obtained from the AmericanType Culture Collection (Manassas, VA); SK-PC-1, IMIM-PC-1, and IMIM-

Fig. 8. Does p48 induce acinar cell-specific gene activation inDSL-derived ductal-like tumor cells? Panel A, Western blottinganalysis of p48 and amylase (Amyl) expression in pancreas cancercells transfected with pcDNA3 vector without insert (C) or with p48cDNA (p48). Panel B, hGH levels in the medium of pancreascancer cells transfected with empty pcDNA3 vector (M) or withp48 cDNA (f), normalized for transfection efficiency. Panel C,representative findings of the subcellular distribution of p48 intransfected cells. In AR42J cells (b), two subpopulations of cellsexpressing nuclear p48 are observed; endogenous p48 is presentin low levels, whereas transfected cells display high levels ofexpression. In DSL6A (a) and RWP-1 (d) cells, cytoplasmic accu-mulation of p48 is observed. In ARIP cells (c), p48 is exclusivelylocalized in the nucleus. 3400.


PC-2 cells, established in our laboratory (11); MZPC-1 cells, obtained fromA. Knuth (Nordwestern Krankenhaus, Frankfurt, Germany); and RWP-1cells, obtained from N. Vaysse (INSERM U151, Toulouse, France). AR42Jand ARIP rat pancreas cancer cells (24) were obtained from N. Vaysse andO. Hagenbuchle (ISREC, Lausanne, Switzerland), respectively. HT-29 co-lon cancer cells were from A. Zweibaum (INSERM U505, Paris, France).Conditions for culture of human pancreas cancer cells have been reportedelsewhere (11). AR42J cells were seeded at approximately 1.5 3 103

cells/cm2, cultured using DMEM supplemented with heat-inactivated fetalbovine serum (10%), L-glutamine, nonessential amino acids, penicillin,and streptomycin, and passed at weekly intervals. ARIP cells were cul-tured similarly except that they were seeded at a 10-fold lower density.

A serially transplanted tumor from the DSL6 transplantable tumor andDSL6A and DSL6B cell lines were obtained as described by Pettengill etal. (26). DSL6A and DSL6B cells were cultured in Waymouth’s MB 752/1medium supplemented with heat-inactivated fetal bovine serum (10%),L-glutamine, nonessential amino acids, penicillin, and streptomycin, andpassed at weekly intervals.

To obtain grafted tumors, cryopreserved pieces of the DSL6 tumor orin vitro cultured DSL6A and DSL6B cells (2 3 106) were injected s.c. tomale Lewis rats. Tumor development was monitored weekly, and tumorsof 1–2 cm diameter were surgically excised for analysis.

Normal human tissues were obtained from surgical samples except fornormal pancreas, which came from organ donors; tumor tissues wereobtained from surgery performed at Hospital del Mar, Barcelona, Spain,except for the three acinar tumors which were kindly provided by Dr. A.Scarpa (Istituto di Anatomia e Istologia Patologia, Facolta di Medicina eChirurgia, Verona, Italy).

Electron Microscopy. Tissue fragments of approximately 1–2 mm3

were fixed with 2% glutaraldehyde in PBS, postfixed with osmium tetrox-ide, dehydrated in ethanol, and embedded in Epon 812 resin. Thin sec-tions were obtained using a ultramicrotome, stained with uranyl acetateand lead citrate, and examined using a Philips 301 electron microscope.

Antibodies. To generate p48-specific polyclonal antibodies, rat p48cDNA was cloned in pBUC and used to transform BL21 strain bacteria.Recombinant colonies were individually analyzed. A positive colony wasselected and induced with isopropyl-1-thio-b-D-galactopyranoside for2.5 h at 37°C; cells were collected, snap frozen, thawed, and sonicated.The fraction that was soluble in 8 M urea contained recombinant p48. Aftercentrifugation at 30,000 3 g for 20 min at 4°C, the supernatant wascollected, dialyzed against 1 M urea, and used to immunize rabbits and toprepare an affinity chromatography matrix as indicated below.

Rabbits were immunized with 50 mg of p48 [diluted in 20 mM HEPES(pH 7.9), 1 M urea, 0.1 M KCl, 2 mM DTT, 10% glycerol, 0.1 mM EDTA, and0.1% NP40 in the presence of protease inhibitors] mixed with completeFreund’s adjuvant. Two, 4, and 6 weeks later, rabbits received boosts withthe same amount of protein mixed with incomplete Freund’s adjuvant.Preimmune and postimmune blood samples were collected, and theirreactivity with recombinant p48 was analyzed by Western blotting. Im-mune serum was preabsorbed three times with a mixture of nuclearextracts from rat kidney, heart, skeletal muscle, and liver coupled to Affigel

10 (Bio-Rad, Richmond, CA). Anti-p48 antibodies present in the fractionthat did not bind to this matrix were subsequently affinity-purified on aNi-NTA matrix to which His-tagged p48 had been previously bound fol-lowing the manufacturer’s recommendations (Qiagen, Valencia, CA).Bound antibodies were eluted with 3.5 M MgCl2; their reactivity withrecombinant p48 and specificity were analyzed by Western blotting asdescribed below. This antibody preparation recognizes exclusively a pro-tein of Mr ;48,000 in nuclear extracts from rat or human pancreas tissue.Rabbit polyclonal antibodies detecting human and rodent amylase werepurchased from Sigma Chemical Co. (St. Louis, MO).

Western Blotting. Normal rat tissues were immediately frozen at280°C. To prepare nuclear extracts, tissues were rapidly thawed andhomogenized with a Dounce homogenizer in 50 mM Tris (pH 7.5), 2 mM

EDTA, 150 mM NaCl, 0.5 mM DTT, and 0.3 M sucrose containing acocktail of protease inhibitors, filtered through a gauze, and separatedthrough a 0.9 M sucrose cushion by centrifugation at 7000 rpm for 15min at 4°C. The supernatant was recentrifuged under the same con-ditions to obtain cytoplasmic extracts. The pellet was resuspended inthe homogenization buffer, brought to 0.3 sucrose and 0.2% NP40,rehomogenized, and recentrifuged through a sucrose cushion as de-scribed above. The pellet, considered the nuclear extract, was used forWestern blotting after lysing and clearing by centrifugation. An aliquotof 40 mg was fractionated by 10% SDS-PAGE, transferred to nitrocel-lulose filters, and incubated with Western blotting blocking buffer(Tris-buffered saline containing 5% skim milk and 0.05% Tween 20).After incubating with specific antibodies (anti-p48 or anti-amylase, 0.5mg/ml in Western blotting blocking buffer), filters were washed andincubated with peroxidase-labeled goat antirabbit immunoglobulin(Dakopatts, Glostrup, Denmark), and reactions were developed withenhanced chemiluminescence (ECL; Amersham, Arlington Heights, IL).

Cultured cells were lysed in 25 mM Tris (pH 7.5), 1 mM EGTA, 1 mM

EDTA, and 1% SDS containing a protease inhibitor cocktail; lysates wereboiled for 15 min and cleared, and protein concentration was determined.Proteins (50 mg) were fractionated by SDS-PAGE, and Western blottingwas carried out as described above.

Immunohistochemistry and Immunocytochemistry. Fresh tissueswere immediately frozen in isopentane cooled at 280°C. Five micronsections were fixed for 10 min with 4% paraformaldehyde, incubated for15 min with H2O2 to block endogenous peroxidase, and washed withPBS. After blocking with 1% BSA, 0.1% saponin, and 0.1% Triton X-100in PBS (blocking buffer) for 30 min, sections were incubated with primaryantibody diluted in the same buffer. After washing with PBS, sections wereincubated with biotin-conjugated goat antirabbit immunoglobulin (Dako-patts), washed with PBS, incubated with streptavidin-peroxidase (2 mg/ml; Pierce, Rockford, IL), and washed with PBS. Reactions were devel-oped using diaminobenzidine as a chromogen. Affinity-purified rabbitanti-p48 antibodies were used at 1–5 mg/ml. Anti-amylase antibodieswere used at 10 mg/ml, and normal rabbit serum diluted to contain acomparable concentration of IgG was used as control.

Immunocytochemical assays were performed on cells fixed with 4%paraformaldehyde, permeabilized with blocking buffer, and incubated

Table 2 PCR primers and conditions used for the detection of transcripts encoding for transcription factors

Sense primer Antisense primer Annealing (°C) Cycles (n) Fragment size (bp)

Ratp48 TGCAGTCTATCAACGACGC GGACAGAGATCTTCCAGTTC 55 37 714E2A CATCCACGTCCTCCGCAGTCA TCTGCCACTGATGTGACCTCC 55 37 339REB CTCTAGATCCTTTACAAGCAAAG GAGGAGTATGTGAGGCGGCAA 55 37 427b-Actin CGTAAAGACCTCTATGCCAA AGCCATGCCAAATGTCTCAT 58 27 473

Humanp48

(1) AAATGTACGGGAGCGGCGCA GGAAGTTGATGTATCCAATGG 60(2) GATGCAGTCCATTAACGATGC CCAATGGCCAGGCGCAGGGTATCG 58 31 106

E2A (E47) GCTGGCCTCAGGTTTCACC ATTGGCCATGCGCCTCTCC 55 37 364E2A (E12) GCTGGCCTCAGGTTTCACC CTCGTTGATGTCACGGACC 55 37 432HEB ACCACTCCATGACTCTGCAGC AGGAGTGTGTGAGGCAGCAAC 55 37 443b-Actin GACTTAGTTGCGTTACACCC CCTCCCCTGTGTGGACTTGG 55 27 349K-ras GCCTGCTGAAAATGACTGAA CTTTGTGTATTTGCCATAAA 54 31 265


with antibodies as described above. For these experiments, an irrelevantmouse monoclonal antibody (B12, to detect dextran) was used as control.

RNA Expression Analysis. RNA from tissues and cultured cells wasisolated using guanidinium isothiocyanate as described by Chomczynskiand Sacchi (56). For RT-PCR, 4 mg of DNase-treated total RNA were usedto synthesize cDNA using a mixture of oligo-d(T) and random hexamers.An aliquot of the cDNA products (1:16) was used for PCR using amplifi-cation conditions optimized for each primer pair. The PCR products wereseparated by agarose gel electrophoresis and visualized by ethidiumbromide staining. The primers and PCR conditions used are shown inTable 2. The identity of the amplified products was confirmed by digestionwith restriction enzymes or by direct sequencing.

To analyze human p48 mRNA expression, a partial cDNA encompass-ing the region coding for the conserved bHLH domain was obtained.Degenerate oligonucleotides flanking the conserved bHLH region wereused to amplify oligo-d(T)-primed cDNA from normal human pancreastissue: forward and reverse primers corresponded to the amino acidsequences ANVRER and YINFL, respectively. After amplification, a PCRproduct of the expected size was eluted from an agarose gel after elec-trophoresis and cloned in pGEMT using manufacturer’s recommenda-tions, and positive colonies were isolated. Three independent clones werestudied, and both strands of the insert DNA were sequenced using anApplied Biosystems A310 equipment. The sequence of the three cloneswas identical. This sequence has been submitted to GenBank with ac-cession number AF181999.

Transient Transfection and Reporter Gene Expression Assays.Rat and human pancreatic cells (6 3 105) were seeded on six-wellplastic plates (Costar, Cambridge, MA) or on sterile coverslips. Twenty-four h later, cells were transfected with 1 mg of linearized plasmid,either with empty vector (pcDNA3) or with the same vector containingthe full-length rat p48 cDNA (pcDNA3.p48) using Lipofectamine (Lipo-fectamine Plus Reagent; Life Technologies, Inc., Gaithersburg, MD),following the manufacturer’s instructions. Cells were harvested at dif-ferent time points to determine the optimal time for analysis; in general,assays were performed 48 h after transfection. For Western blottingand immunocytochemical procedures, cell lysates were prepared asdescribed above. To determine the transcriptional activity of the PTF1complex, cells were cotransfected with one of the two plasmids de-scribed above (0.2 mg) plus 0.2 mg of a plasmid, kindly provided by G.Swift and R. MacDonald (University of Texas Southwestern MedicalCenter, Dallas, TX), containing the hGH cDNA downstream from a6-mer of the A element from the rat elastase promoter (6A26Elp.hGH;Ref. 37) and 0.1 mg of plasmid pGL2-control vector (Promega Corp.,Madison, WI), containing the luciferase cDNA, to normalize for trans-fection efficiency. For these experiments, 105 cells were seeded inwells of 24-well plates (Costar) in duplicates, and transfection wasperformed 24 h later using the Lipofectamine Plus reagent. Forty-eighth later, hGH activity was measured in the culture medium using a RIA(Nichols Institute Diagnostics, San Juan Capistrano, CA), and amylaseactivity was determined using the AMYL commercial kit (BoehringerMannheim, Mannheim, Germany). Cells were lysed and processed forluciferase activity assays according to the manufacturer’s recommen-dations (Promega).

AcknowledgmentsWe thank the investigators mentioned in the text for providing cells andreagents, O. Hagenbuchle for the rat p48 cDNA and for many valuablediscussions, and C. Balague, E. Batlle, M. Garrido, J. Lloreta, P. Navarro,T. Palomero, and the members of the Unitat de Biologia Cel.lular i Mo-lecular, Institut Municipal d’Investigacio Medica, for valuable contribu-tions.

References1. Kloppel, G. Pathology of nonendocrine pancreatic tumors. In: V. L. W.Go, E. P. DiMagno, J. D. Gardner, E. Lebenthal, H. A. Reber, and G. A.Scheele (ed.), The Pancreas: Biology, Pathobiology, and Disease, pp.871–898. New York: Raven Press, 1993.

2. Real, F. X. The cell biology of pancreatic cancer. In: J. Neoptolemosand N. R. Lemoine (eds), Pancreatic Cancer, pp. 3–17. Oxford: BlackwellScience Press, 1995.

3. Schussler, M. H., Skoudy, A., Ramaekers, F., and Real, F. X. Interme-diate filaments as differentiation markers of normal pancreas and pan-creas cancer. Am. J. Pathol., 140: 559–568, 1992.

4. Real, F. X., Vila, M. R., Skoudy, A., Ramaekers, F. C. S., and Coromi-nas, J. M. Intermediate filaments as differentiation markers of exocrinepancreas. II. Expression of cytokeratins of complex and stratified epitheliain normal pancreas and in pancreas cancer. Int. J. Cancer, 54: 720–727,1993.

5. Vila, M. R., Balague, C., and Real, F. X. Cytokeratins and mucins asmolecular markers of cell differentiation and neoplastic transformation inthe exocrine pancreas. Zentralbl. Pathol, 140: 225–235, 1994.

6. Balague, C., Gambus, G., Carrato, C., Porchet, N., Aubert, J. P., Kim,Y. S., and Real, F. X. Altered expression of MUC2, MUC4, and MUC5mucin genes in pancreas tissues and cancer cell lines. Gastroenterology,106: 1054–1061, 1994.

7. Balague, C., Audie, J. P., Porchet, N., and Real, F. X. In situ hybrid-ization shows distinct patterns of mucin gene expression in normal, be-nign, and malignant pancreas tissues. Gastroenterology, 109: 953–964,1995.

8. Hartel, S., Gossrau, R., Hanski, C., and Reutter, W. Dipeptidyl pepti-dase (DPP) IV in rat organs. Histochemistry, 89: 151–161, 1988.

9. Marino, C. R., Matovcik, L. M., Gorelick, F. S., and Cohn, J. A. Local-ization of the cystic fibrosis transmembrane conductance regulator inpancreas. J. Clin. Investig., 88: 712–716, 1988.

10. Kim, J. H., Ho, S. B., Montgomery, C. K., and Kim, Y. S. Cell lineagemarkers in human pancreatic cancer. Cancer (Phila.), 66: 2134–2143,1990.

11. Vila, M. R., Lloreta, J., Schussler, M. H., Berrozpe, G., Welt, S., andReal, F. X. New pancreas cancer cell lines that represent distinct stagesof ductal differentiation. Lab. Investig., 72: 395–404, 1995.

12. Scarpelli, D. M., Rao, M. S., and Reddy, J. K. Are acinar cells involvedin the pathogenesis of ductal adenocarcinoma of the pancreas? CancerCells, 3: 275–277, 1991.

13. Pour, P. M. The role of Langerhans islets in pancreatic ductal ade-nocarcinomas. Front. Biosci., 2: D271–D282, 1997.

14. Longnecker, D. S., Shinozuka, H., and Dekker, A. Focal acinar celldysplasia in human pancreas. Cancer (Phila.), 45: 534–540, 1980.

15. Kloppel, G., and Longnecker, D. S. Hyperplastic and metaplasticchanges in pancreatic ducts: nomenclature and preneoplastic potential.Ann. NY Acad. Sci., 880: 66–73, 1999.

16. Yanagisawa, A., Ohtake, K., Ohashi, K., Hori, M., Kitagawa, T.,Sugano, H., and Kato, Y. Frequent c-Ki-ras oncogene activation in mu-cous cell hyperplasias of pancreas suffering from chronic inflammation.Cancer Res., 53: 953–956, 1993.

17. Caldas, C., and Kern, S. E. K-ras mutation and pancreatic adenocar-cinoma. Int. J. Pancreatol., 18: 1–6, 1995.

18. Luttges, J., Schlehe, B., Menke, M. A., Vogel, I., Henne-Bruns, D., andKloppel, G. The K-ras mutation pattern in pancreatic ductal adenocarci-noma usually is identical to that in associated normal, hyperplastic, andmetaplastic ductal epithelium. Cancer (Phila.), 85: 1703–1710, 1999.

19. Caldas, C., Hahn, S. A., Hruban, R. H., Redson, M. S., Yeo, C. J., andKern, S. E. Detection of K-ras mutations in the stool of patients withpancreatic adenocarcinoma and pancreatic ductal hyperplasia. CancerRes., 54: 3568–3573, 1994.

20. Malats, N. Prevalenca i factors associats a les mutacions enl’oncogen Ki-ras en el cancer de pancrees exocrı i els cancers del sistemabiliar extrahepatic (Estudi PANK-ras I). Doctoral Thesis, Facultat de Me-dicina, Universitat Autonoma de Barcelona, Barcelona (Spain), 1995.

21. Moskaluk, C. A., Hruban, R. H., and Kern, S. E. p16 and K-ras genemutations in the intraductal precursors of human pancreatic adenocarci-noma. Cancer Res., 57: 2140–2143, 1997.

22. Erill, N., Cuatrecasas, M., Sancho, F. J., Farre, A., Pour, P. M., Lluis,F., and Capella, G. K-ras and p53 mutations in hamster pancreatic ductaladenocarcinomas and cell lines. Am. J. Pathol., 149: 1333–1339, 1996.

23. Schaeffer, B. K., Zurlo, J., and Longnecker, D. S. Activation of c-Ki-rasnot detectable in adenomas or adenocarcinomas arising in rat pancreas.Mol. Carcinog., 3: 165–170, 1990.


24. Jessop, N. W., and Hay, R. J. Characteristics of two rat pancreaticexocrine cell lines derived from transplantable tumors. In Vitro, 16: 212,1980.

25. Christophe, J. Pancreatic tumoral cell line AR42J: an amphicrinemodel. Am. J. Physiol., 266: G963–G971, 1994.

26. Pettengill, O. S., Faris, R. A., Bell, R. H., Jr., Kuhlmann, E. T., andLongnecker, D. S. Derivation of ductlike cell lines from transplantableacinar cell carcinoma of the rat pancreas. Am. J. Pathol., 143: 292–303,1993.

27. Sandgren, E. P., Quaife, C. J., Paulovich, A. G., Palmiter, R. D., andBrinster, R. L. Pancreatic tumor pathogenesis reflects the causative ge-netic lesion. Proc. Natl. Acad. Sci. USA, 88: 93–97, 1991.

28. Cockell, M., Stevenson, B. J., Strubin, O., Hagenbuchle, O., andWellauer, P. K. Identification of a cell-specific DNA-binding activity thatinteracts with a transcriptional activator of genes expressed in the acinarpancreas. Mol. Cell. Biol., 9: 2464–2476, 1989.

29. Petrucco, S., Wellauer, P. K., and Hagenbuchle, O. The DNA-bindingactivity of transcription factor PTF1 parallels the synthesis of pancreas-specific mRNAs during mouse development. Mol. Cell. Biol., 10: 254–264,1990.

30. Roux, E., Strubin, M., Hagenbuchle, O., and Wellauer, P. K. Thecell-specific transcription factor PTF1 contains two different subunits thatinteract with the DNA. Genes Dev., 3: 1613–1624, 1989.

31. Krapp, A., Knofler, M., Frutiger, S., Hughes, G. J., Hagenbuchle, O.,and Wellauer, P. K. The p48 DNA-binding subunit of transcription factorPTF1 is a new exocrine pancreas-specific basic helix-loop-helix protein.EMBO J., 15: 4317–4329, 1996.

32. Hagenbuchle, O., Krapp, A., Knofler, M., and Wellauer, P. K. ThebHLH protein p48, a DNA-binding subunit of PTF1, acts as a morphogenduring pancreas development. Exp. Clin. Endocrinol. Diabetes, 105: 4,1997.

33. Sommer, L., Hagenbuchle, O., Wellauer, P. K., and Strubin, M. Nu-clear targeting of the transcription factor PTF1 is mediated by a proteinsubunit that does not bind to the PTF1 cognate sequence. Cell, 67:987–994, 1991.

34. Walsh, K., and Perlman, H. Cell cycle exit upon myogenic differenti-ation. Curr. Opin. Genet. Dev., 7: 597–602, 1997.

35. Cockell, M., Stolarczyk, D., Frutiger, S., Hughes, G. J., Hagenbuchle,O., and Wellauer, P. K. Binding sites for hepatocyte nuclear factor 3b or3g and pancreas transcription factor 1 are required for efficient expressionof the gene encoding pancreatic a-amylase. Mol. Cell. Biol., 15: 1933–1941, 1995.

36. Swift, G. H., Liu, Y., Rose, S. D., Bischof, L. J., Steelman, S., Buch-berg, A. M., Wright, C. V., and MacDonald, R. J. An endocrine-exocrineswitch in the activity of the pancreatic homeodomain protein PDX1through formation of a trimeric complex with PBX1b and MRG1 (MEIS2).Mol. Cell. Biol., 18: 5109–5120, 1998.

37. Rose, S. D., Kruse, F., Swift, G. H., MacDonald, R. J., Hammer, R. E.A single element of the elastase I enhancer is sufficient to direct tran-scription selectively to the pancreas and gut. Mol. Cell. Biol., 14: 2048–2057, 1994.

38. Krapp, A., Knofler, M., Ledermann, B., Burki, K., Berney, C., Zoerkler,N., Hagenbuchle, O., and Wellauer, P. K. The bHLH protein PTF1–p48 isessential for the formation of the exocrine and the correct spatial organi-zation of the endocrine pancreas. Genes Dev., 12: 3752–3763, 1998.

39. Knofler, M., Krapp, A., Hagenbuchle, O., and Wellauer, P. K. Consti-tutive expression of the gene for the cell-specific p48 DNA-binding sub-unit of pancreas transcription factor 1 in cultured cells is under control of

binding sites for transcription factors Sp1 and aCbf. J. Biol. Chem., 271:21993–22002, 1996.

40. Vila, M. R., Lloreta, J., and Real, F. X. Normal human pancreascultures display functional ductal characteristics. Lab. Investig., 71: 423–431, 1994.

41. Adell, T., Skoudy, A., Gomez-Cuadrado, A., Palomero, T., and Real, F.X. The bHLH transcription factor p48 in exocrine pancreas differentiationin vitro. In: T. M. Gress (ed.), Molecular Pathogenesis of Pancreatic Can-cer, in press. Amsterdam: IOS Press, 2000.

42. Duluc, I., Freund, J. N., Leberquier, C., and Kedinger, M. Fetalendoderm primarily holds the temporal and positional information re-quired for mammalian intestinal development. J. Cell. Biol., 126: 211–221,1994.

43. Ahlgren, U., Pfaff, S. L., Jessell, T. M., Edlund, T., and Edlund, H.Independent requirement for ISL1 in formation of pancreatic mesenchymeand islet cells. Nature (Lond.), 385: 257–260, 1997.

44. Benezra, R., Davis, R. L., Lockson, D., Turner, D. L., and Weintraub,H. The protein Id: a negative regulator of helix-loop-helix DNA bindingproteins. Cell, 61: 49–59, 1990.

45. Sun, X-H. Constitutive expression of the Id1 gene impairs mouse Bcell development. Cell, 79: 893–900, 1994.

46. Norton, J. D., Deed, R. W., Craggs, G., and Sablitzky, F. Id helix-loop-helix proteins in cell growth and differentiation. Trends Cell Biol., 8: 58–65, 1998.

47. Desprez, P-Y., Hara, E., Bissell, M. J., and Campisi, J. Suppression ofmammary epithelial cell differentiation by the helix-loop-helix protein, Id-1.Mol. Cell. Biol., 15: 3398–3404, 1995.

48. Puri, P. L., Sartorelli, V., Yang, X. J., Hamamori, Y., Ogryzko, W.,Howard, B. H., Kedes, L., Wang, J. Y., Graessmann, A., Nakatani, Y., andLevrero, M. Differential roles of p300 and PCAF acetyltransferases inmuscle differentiation. Mol. Cell, 1: 35–45, 1997.

49. Halevy, O., Novitch, B. G., Spicer, D. B., Skapek, S. X., Rhee, J.,Hannon, G. J., Beach, D., and Lassar, A. B. Correlation of terminal cellcycle arrest of skeletal muscle with induction of p21 by MyoD. Science(Washington DC), 267: 1018–1021, 1995.

50. Skapek, S. X., Rhee, J., Spicer, D. B., and Lassar, A. B. Inhibition ofmyogenic differentiation in proliferating myoblasts by cyclin D1-depend-ent kinase. Science (Washington DC), 267: 1022–1024, 1995.

51. Parker, S. B., Eichele, G., Zhang, P., Rawls, A., Sands, A. T., Bradley,A., Olson, E. N., Harper, J. W., and Elledge, S. J. p53-independentexpression of p21Cip1 in muscle and other terminally differentiating cells.Science (Washington DC), 267: 1024–1027, 1995.

52. Otten, A. D., Firpo, E. J., Gerber, A. B., Brody, L. L., Roberts, J. M.,and Tapscott, S. J. Inactivation of MyoD-mediated expression of p21 intumor cell lines. Cell Growth Differ., 8: 1151–1160, 1997.

53. Frascella, E., and Rosolen, A. Detection of the MyoD1 transcript inrhabdomyosarcoma cell lines and tumor samples by reverse transcriptionpolymerase chain reaction. Am. J. Pathol., 152: 577–583, 1998.

54. Cui, S., Hano, H., Harada, T., Takai, S., Masui, F., and Ushigome, S.Evaluation of new monoclonal anti-MyoD1 and anti-myogenin antibodiesfor the diagnosis of rhabdomyosarcoma. Pathol. Int., 49: 62–68, 1999.

55. Novitch, B. G., Spicer, D. B., Kim, P. S., Cheung, W. L., and Lassar,A. B. pRb is required for MEF2-dependent gene expression as well ascell-cycle arrest during skeletal muscle differentiation. Curr. Biol., 9: 449–459, 1999.

56. Chomczynski, P., and Sacchi, N. Single-step method for RNA isola-tion using guanidinium thiocyanate-phenol-chloroform extraction. Anal.Biochem., 162: 156–159, 1987.


role of the basic helix-loop-helix transcription factor...

Documents