Early Requirement of Rac1 in a Mouse Model of Pancreatic Cancer

BASIC AND TRANSLATIONAL—PANCREAS

Early Requirement of Rac1 in a Mouse Model of Pancreatic CancerIRINA HEID,* CLARA LUBESEDER–MARTELLATO,* BENCE SIPOS,‡ PAWEL K. MAZUR,* MARINA LESINA,*ROLAND M. SCHMID,* and JENS T. SIVEKE*

*II. Medizinische Klinik Klinikum rechts der Isar, Technische Universität München, Munich; and ‡Department of Pathology, University Hospital Tübingen, Tübingen,Germany

See editorial on page 427.

BACKGROUND & AIMS: Pancreatic ductal adenocarci-noma (PDAC) is a fatal disease without effective chemo-preventive or therapeutic approaches. Although the roleof oncogenic Kras in initiating development of PDAC iswell established, downstream targets of aberrant Ras sig-naling are poorly understood. Acinar-ductal metaplasia(ADM) appears to be an important prerequisite for devel-opment of pancreatic intraepithelial neoplasia (PanIN), acommon precursor to PDAC. RAS-related C3 botulinumsubstrate 1 (Rac1), which controls actin reorganization,can be activated by Ras, is up-regulated in several humancancers, and is required for cerulein-induced morphologicchanges in acini. We investigated effects of loss of Rac1 inKras-induced pancreatic carcinogenesis in mice. METH-ODS: Using a Cre/lox approach, we deleted Rac1 frompancreatic progenitor cells in different mouse models ofPDAC and in mice with cerulein-induced acute pancreati-tis. Acinar epithelial explants of mutant mice were used toinvestigate the role of Rac1 in vitro. RESULTS: Rac1expression increased in mouse and human pancreatic tu-mors, particularly in the stroma. Deletion of Rac1 inKrasG12D-induced PDAC in mice reduced formation ofADM, PanIN, and tumors and significantly prolongedsurvival. Pancreatic epithelial metaplasia was accompa-nied by apical-basolateral redistribution of F-actin, alongwith basal expression of Rac1. Acinar epithelial explantsthat lacked Rac1 or that were incubated with inhibitors ofactin polymerization had a reduced ability to undergoADM in 3-dimensional cultures. CONCLUSIONS: Inmice, Rac1 is required for early metaplastic changesand neoplasia-associated actin rearrangements in de-velopment of pancreatic cancer. Rac1 might be devel-oped as a diagnostic marker or therapeutic target forPDAC.

Keywords: Genetically Engineered Mice; Ductal Cell; Cyto-skeleton; Signaling.

Invasive pancreatic ductal adenocarcinoma (PDAC) isthe most frequent and highly lethal type of pancreatic

cancer. One of the reasons for the high lethality of PDACis the very late and difficult diagnosis due to absence of

early symptoms and of robust diagnostic markers. Todate, 3 main forms of precursor lesions have been de-scribed: pancreatic intraepithelial neoplasia (PanIN), in-traductal papillary mucinous neoplasm, and mucinouscystic neoplasm. The progression of PanIN lesions fromPanIN1 to PanIN3, the carcinoma in situ, is the mostcommon and best-characterized model for developmentof PDAC (for review, see Hezel et al1). Nevertheless, theinitiation process, the cell of origin, as well as the signal-ing events leading to precursor occurrence are still poorlyunderstood.

Morphologic similarities suggest PanINs derive fromductal cells.2 However, recent lineage tracing studies con-firmed that premalignant lesions can arise from differen-tiated acinar cells in part through a reprogrammingmechanism named acinar-ductal metaplasia (ADM).3Along this process, acinar cells react to various stimuli,including Notch signaling, growth factors, and inflamma-tion or cellular damage, thereby reducing expression ofexocrine markers and developing into tubular structureswith ductal properties. One of the best-known pathwaysinvolved in ADM is epidermal growth factor receptor(EGFR) signaling. Exposure of acinar cells to the EGFRmain ligand transforming growth factor (TGF)-! stronglyinduces ADM in vitro and in vivo.4 –9 In addition, aberrantactivation or overexpression of oncogenic KrasG12D resultsin occurrence of ADM.5,10 This process is dramaticallyaccelerated when oncogenic KrasG12D is combined withoverexpression of TGF-!11 or activation of Notch signal-ing.3 Finally, repeated injections of cerulein, a pancreati-tis-inducing cholecystokinin (CCK) analogue, have beenused in several studies to investigate the role of ADM ininitiation and progression of PDAC.12–14 Thus, EGFR-,Notch-, and/or Kras-activation driven as well as inflam-mation-induced transdifferentiation of acinar cells to a

Abbreviations used in this paper: ADM, acinar-ductal metaplasia;BrdU, bromodeoxyuridine; CK19, cytokeratin 19; CytD, cytochalasin D;EGFR, epidermal growth factor receptor; LatA, latrunculin A; PanIN,pancreatic intraepithelial neoplasia; PDAC, pancreatic ductal adenocar-cinoma; Rac1, RAS-related C3 botulinum substrate 1; TGF, tumorgrowth factor; WT, wild-type.

© 2011 by the AGA Institute0016-5085/$36.00

doi:10.1053/j.gastro.2011.04.043

BASI

CA

ND

TRA

NSL

ATI

ON

AL

PAN

CREA

S

GASTROENTEROLOGY 2011;141:719–730

ductal-like phenotype is believed to be a crucial step in theinitiation of pancreatic carcinogenesis.

However, little is known about the role of Kras andEGFR downstream targets in induction of ADM/PanINand progression of PDAC. One potential candidate isRAS-related C3 botulinum substrate 1 (Rac1), an EGFRand Kras effector molecule. Rac1 belongs to the Rhofamily of small guanosine triphosphatases.15 Guanosinetriphosphatases function as molecular switches betweeninactive guanosine diphosphate– and active guanosinetriphosphate– bound form, receiving extracellular signalsfrom growth factor receptors and integrins and transduc-ing them to the intracellular target.16 The best-describedfunction of Rac1 is the engagement in actin cytoskeletonrearrangements and cellular motility.17–19 Specifically, ac-tivated Rac1 causes the uncapping of the fast-growing endof actin filaments and induces rapid actin polymerization.In a mouse model of chemically induced pancreatitis,Rac1 was proven necessary for CCK-induced acinar mor-phologic changes, F-actin redistribution, and amylase se-cretion.20 Additionally, several studies reveal that Rac1 isimplicated in control of cell cycle, growth and sur-vival,21,22 Ras-induced transformation,23,24 and resistanceto chemotherapy.25 Overexpression of Rac1 has been de-tected in human patient samples of breast, gastric, testic-ular, oral squamous cell, lung, and pancreatic cancer(70%).26,27 In addition, the crucial role of Rac1 in thetumor initiation process has been described in mousemodels of colorectal,28 skin,29 and lung30 cancer, showingthat Rac1 overexpression promoted and ablation of Rac1decreased tumor progression. Here we describe the phe-notype of conditional Rac1 deletion in well-characterizedKrasG12D-induced murine models of PDAC and acute pan-creatitis. We provide evidence for an important role ofRac1 and actin cytoskeleton rearrangements during ADM,underlining the importance of Rac1 in the developmentof pancreatic precursor lesions.

Materials and MethodsMouse Strains and Experimental PancreatitisPtf1awt/Cre, Kraswt/LSL-G12D, Trp53wt/LSL-R172H, Rac1fl/fl, and

Rosa26wt/LSL-LacZ mouse strains have been described before.4,11,31–33

Mice were of mixed 129SV/C57BL/6 background. The genotypesare listed in Figure 2F. Acute pancreatitis was induced by ad-ministration of 8 hourly intraperitoneal injections of cerulein(10 "g/kg body wt) over 2 consecutive days as described previ-ously.34 Pancreata were analyzed 24 and 72 hours after the lastinjection. All experiments were performed according to theguidelines of the local animal use and care committees.

Statistical AnalysesKaplan–Meier curves were calculated using the survival

time for each mouse from all littermate groups and log-rank testfor significance. For all other analyses, the unpaired 2-tailedStudent t test was performed with P ! .05 considered as signif-icant.

Supplementary DataDetailed description of additional procedures and sup-

plementary figures are provided in the Supplementary Materialsand Methods.

ResultsIncreased Rac1 Expression in Developmentand Progression of PDACTo determine whether Rac1 is involved in the ini-

tiation and/or progression of KrasG12D-induced PDAC, wecrossed Kraswt/LSL-G12D to Ptf1awt/Cre mice (referred to asKrasG12D) as previously described.11 Increased expression oftotal Rac1 in pancreatic whole tissue was detected approx-imately at the age of 4 weeks, when few ductal-like meta-plastic structures associated with some scattered PanINsstart to develop (Figures 1A and B and 2A). The levels ofactive, guanosine triphosphate– bound form of Rac1 inKrasG12D pancreata at this time point were similar to wild-type (WT) controls (Figure 1A). The amount of total Rac1was significantly increased during disease progression onboth messenger RNA and protein levels (Figure 1B and C).Immunofluorescence analysis of WT pancreata showedweak cytoplasmic acinar expression of Rac1 (Figure 1Dand E). Interestingly, Rac1 expression in ADM revealed anincreased basal localization suggesting Rac1 activation,while PanINs showed almost no positivity (Figure 1D andE). Strong Rac1 expression was notable in the PanIN andPDAC surrounding stroma in murine and human sam-ples (Figure 1D and Supplementary Figure 2C and D).

Loss of Rac1 Reveals No Major PancreaticAbnormalitiesRac1 is ubiquitously expressed during pancreas

development in ductal, acinar, endocrine, and mesenchy-mal cells. For pancreatic ablation of Rac1, we crossedpreviously described Rac1fl/fl33 with Ptf1awt/Cre lines to gen-erate mice with pancreatic Rac1 deficiency (referred to asRac1ko; Figure 2F). Rac1ko mice were viable and born atthe expected Mendelian ratio. Successful Cre-mediatedrecombination was assessed by polymerase chain reactionand X-Gal staining using a Rac1ko;Rosa26RLSL-LacZ reporterline (Supplementary Figure 1A and data not shown). Im-munofluorescence staining and Western blot analysis re-vealed loss of Rac1 in pancreata of 4-week-old animals(Supplementary Figure 1B).

Rac1ko pancreata revealed a grossly normal morphologyup to 18 months of age (Supplementary Figure 1C). Wenoted a minor impairment in the endocrine function in4-week-old Rac1ko mice (Supplementary Figure 1D), whichpersisted in older animals (data not shown). However, nei-ther weight nor size of Rac1ko mice were altered comparedwith WT littermates, and we observed similar staining pat-terns of exocrine and endocrine proteins (SupplementaryFigure 1D). Thus, we conclude that deletion of Rac1 resultsin no major pancreatic abnormalities. To test if this mildphenotype was due to reconstitution by similar factors,quantitative reverse-transcription polymerase chain reaction

BASIC

AN

DTRA

NSLA

TION

AL

PAN

CREAS

720 HEID ET AL GASTROENTEROLOGY Vol. 141, No. 2

analysis of Rac2/3 and Cdc42 activity assay was performed.Here, we found low and unchanged levels of Rac2 and Rac3as well as absent Cdc42-GTP in WT and Rac1ko pancreata(data not shown).

Ablation of Rac1 Reduces Early PanINBurden and Impairs Development of PDACFor analysis of Rac1 in pancreatic carcinogenesis,

we crossed Rac1fl/fl with Ptf1awt/Cre and Kras"/LSL-G12D linesto obtain Ptf1awt/Cre;Kraswt/LSL-G12D;Rac1fl/fl mice (KrasG12D;Rac1ko). First, we analyzed the presence of ADM in4-week-old KrasG12D and KrasG12D;Rac1ko animals. Asshown in Figure 2A and B, quantification of staining forcytokeratin 19 (CK19) illustrated a greatly reduced inci-

dence of ADM in KrasG12D;Rac1ko mice. At 6 months ofage, KrasG12D;Rac1wt/fl had significantly larger pancreas thanKrasG12D;Rac1ko mice (Figure 2C, P # .001). Comparison ofKrasG12D;Rac1wt/fl and KrasG12D cohorts (n # 5 each) revealedsimilar gross morphologic and histologic appearance (Fig-ure 2D and E), suggesting that one allele of Rac1 issufficient for the metaplastic transdifferentiation process.While large areas of KrasG12D;Rac1wt/fl pancreata were re-placed by ADM and PanIN1-2 lesions with surroundingstroma, Rac1-depleted animals showed an almost normalpancreas with rare focal ADM/PanIN lesions (Figure 2Dand E). Therefore, we conclude that deletion of Rac1 leadsto a substantial delay in the metaplastic process.

Figure 1. Rac1 expression increases during PDAC progression. (A) Affinity precipitation assay of pancreatic whole-tissue lysates indicates similarlevels of active guanosine triphosphate–bound Rac1 in 4-week-old WT and KrasG12D mice. (B) Immunoblot analyses of pancreatic whole-tissuelysates show enhanced protein levels of Rac1 at time points of 1 and 9 months. Hsp90 serves as loading control. (C) Rac1 messenger RNA issignificantly up-regulated in the pancreas of 6-month-old KrasG12D mice. Results of quantitative reverse-transcription polymerase chain reactionnormalized to GAPDH (n # 3 per group). (D) Immunofluorescence double stainings for Rac1 and amylase in pancreatic tissue cryosections from WTand KrasG12D mice. (Upper panels) Rac1 is weakly expressed in pancreatic acini. (Lower panels) Immunofluorescence (IF) analysis reveals slightlyincreased basal localization of Rac1 in the acinar compartment and strong Rac1 expression in the pancreatic stroma. Scale bar # 50 "m. (E)Immunofluorescence staining shows cytoplasmic expression of Rac1 in acinar cells and membranous expression in ADM, whereas PanINs arenegative. Scale bar # 50 "m.

BASI

CA

ND

TRA

NSL

ATI

ON

AL

PAN

CREA

S

August 2011 TARGETING RAC1 IN PANCREATIC CANCER 721

Figure 2. Loss of Rac1 impairs ADM and mPanIN development in KrasG12D;Rac1ko mice. (A) Deletion of Rac1 impairs the development ofmetaplastic acini in KrasG12D;Rac1ko compared with KrasG12D mice. CK19 staining of pancreatic tissue (left panel) from 4-week-old KrasG12D animalsshows focal positive ADM lesions (inset). In contrast, pancreas of KrasG12D;Rac1ko mice (right panel) reveals only very rare single ADM lesions. Scalebar # 50 "m. (B) Quantification of CK19-positive ADM lesions in pancreata of KrasG12D and KrasG12D;Rac1ko mice (n # 3 per group). (C) Macroscopicappearance of KrasG12D; Rac1wt/fl versus KrasG12D;Rac1ko pancreata at the age of 6 months. Note the smaller size and absence of visible cysts(arrowhead) in KrasG12D;Rac1ko pancreas. Ruler gap # 1 mm. (D) Deletion of Rac1 reduces PanIN burden in KrasG12D;Rac1ko compared withKrasG12D;Rac1wt/fl mice. H&E stain of 6-month-old KrasG12D;Rac1wt/fl pancreatic tissue (left) shows disseminated PanIN/ADM development spread inthe whole organ. In contrast, pancreata of KrasG12D;Rac1ko mice (right) reveal only rare focal PanIN lesions. Scale bar # 50 "m. (E) Quantificationof PanIN1 in 6-month-old KrasG12D, KrasG12D;Rac1wt/fl, and KrasG12D;Rac1ko mice. (F) Survival analysis of KrasG12D (n # 19; median, 411 days) versusKrasG12D;Rac1ko (n # 13; median, 529 days) mice. Pancreas-specific Rac1 deletion significantly prolongs the survival of KrasG12D;Rac1ko micecompared with the KrasG12D cohort. The table shows abbreviation of genotypes used throughout the text.

BASIC

AN

DTRA

NSLA

TION

AL

PAN

CREAS


Next, we analyzed the effect of Rac1 deficiency in asecond, more aggressive model with an additional domi-nant negative p53wt/R172H allele.42 At the age of 3 months,pancreatic tissue of KrasG12D;p53R172H mice was almostcompletely replaced by ADM/PanIN1-2 and accompaniedstromal reaction. In contrast, pancreata of KrasG12D;p53R172H;Rac1ko mice showed only very small areas withlow-grade ADM and almost no fibrosis or inflammatoryinfiltrates (Supplementary Figure 3A and B). Altogether,these results underline our hypothesis that Rac1 is crucialfor the development of PanIN lesions.

For analysis of PDAC development, a cohort of micewas followed up for signs of disease progression or death.The clinical course of KrasG12D mice has been describedbefore,11,35 with an approximately 50% to 60% rate ofPDAC incidence and a median survival of 411 days. No-tably, loss of Rac1 resulted in prevented PDAC develop-ment and a significantly prolonged survival rate (mediansurvival, 529 days; Figure 2F). Histologic end point anal-ysis of the KrasG12D;Rac1ko pancreata revealed mixed phe-notypes. The pancreas of most mice contained areas ofnormal-appearing pancreas with few chronic pancreatitis-like inflammatory infiltrates and low-grade PanINs. In 2animals, we noted extensive adipose replacement and de-velopment of extensive lymphoid infiltrates in at least 4 of8 analyzed mice (Supplementary Figure 3C and E). Fur-ther immunohistochemical analysis revealed these lym-phomas to be B-cell lymphoma. Polymerase chain reac-tion analysis of microdissected tumor tissue showed aweak band of recombined Rac1, most likely due to some

pancreatic Ptf1a-derived cells within the lymphoma tissue.Only one of the 8 mice showed higher-grade PanINs.Polymerase chain reaction of these microdissected lesionsshowed Cre-mediated recombination of the Rac1 alleles(Supplementary Figure 3D). Thus, Rac1 deletion nearlycompletely ablates development of PDAC.

Loss of Rac1 Reduces ADM in Cerulein-Induced Acute PancreatitisWe next addressed the question whether the im-

paired metaplasia following Rac1 deletion may also beobservable in a primarily nononcogenic but inflammatorysetting. Cerulein-induced acute pancreatitis is a well-es-tablished model for development of ADM.34 Two cohortsconsisting of WT and Rac1ko littermates were injectedwith cerulein over 2 days as described previously34 andanalyzed 24 and 72 hours thereafter. Deletion of Rac1significantly reduced development of CK19" ADM lesionsduring acute pancreatitis (Figure 3A and B). Although theamount of ADM lesions at 72 hours failed to show sig-nificant differences in Rac1ko compared with WT animals,a clear trend in favor of Rac1ko pancreata was visible(Figure 3C and D). Thus, we conclude that Rac1 is animportant determinant of ADM induction in both anoncogenic and inflammatory setting.

ADM Is Characterized by ActinReorganizationTo elucidate the underlying role of Rac1 during

pancreatic carcinogenesis, we assessed proliferation in

Figure 3. Loss of Rac1 reduces ADM in cerulein-induced acute pancreatitis. (A and C) Deletion of Rac1 reduces the development of ADM lesionsduring cerulein-induced acute pancreatitis in Rac1ko compared with WT mice. Arrowheads show CK19-positive ADM lesions (d, duct; I, islet). Scalebar # 50 "m. (B and D) Quantification of CK19-positive ADM lesions in WT versus Rac1ko mice (B) 24 hours after last injection (n # 6 per group) and(D) 72 hours after last injection (n # 4 per group).

BASI

CA

ND

TRA

NSL

ATI

ON

AL

PAN

CREA

S


KrasG12D versus KrasG12D;Rac1ko mice. Here we found nodifferences in BrdU-positive cells in 7-day-old pancreata,metaplastic duct-like structures, and PanIN lesions at 6months of age (Supplementary Figure 4A). Furthermore,evaluation of major Ras/Egfr downstream effectors andgene expression analysis of Notch and Hedgehog genesshowed no obvious differences during early carcinogene-sis (Supplementary Figure 4B–D). Hence, we hypothesizethat Rac1-dependent reduced cellular plasticity may bethe cause for the impaired ADM and PanIN development.Indeed, Rac1 is strongly involved in acinar morphologychanges, F-actin redistribution, and amylase secretionupon CCK treatment in vitro.20 Thus, we assessed F-actinlocalization and rearrangement in the cerulein-inducedpancreatitis model. Indeed, F-actin was clearly locatedapically and redistributed basolaterally following induc-tion of pancreatitis in WT acini (Figure 4A). This processwas reduced in Rac1ko mice, possibly explaining the im-pairment of ADM formation. In KrasG12D animals we ob-served a similar effect; both F-actin and CK19 were trans-located to the basal membrane during ADM in additionto the original apical location (Figure 4B and Supplemen-tary Figure 5). Thus, we propose a model for F-actinredistribution during ADM. The apical-located actin fila-ments protrude to the basolateral side and concentrate atboth sides of the metaplastic transdifferentiating cells, asign for loss of cell polarity (Figure 4C).

Rac1 Is Required for the Initiation of ADMin Pancreatic Acinar ExplantsTo analyze Rac1 involvement into the process de-

scribed previously, we used an in vitro system for ADM.Several studies show that acinar pancreatic epithelial ex-plants cultured in collagen matrix differentiate to duct-like structures after TGF-! stimulation mimicking in vivoADM.5,36 Therefore, we assessed the function of Rac1 inthe exocrine metaplastic conversion upon TGF-! stimu-lation and spontaneously in explanted cells with onco-genic KrasG12D expression.

Freshly harvested explants from pancreata of WT,Rac1ko, KrasG12D, and KrasG12D;Rac1ko littermates consistedof acinar cell clusters expressing amylase and were mor-phologically indistinguishable (Figure 5B). Immunofluo-rescence staining of acinar epithelial explants from WTand KrasG12D mice showed weak Rac1 expression afterisolation; however, upon 1 hour of TGF-! treatment,Rac1 expression strongly increased (Supplementary Fig-ure 2A). Treated with 50 ng/mL TGF-! for 5 days, 74.5%$ 22.4% (n # 3) of the WT acini acquired a duct-likemorphology, characterized by cystic structures lined withsimple epithelia and expressing the duct cell markerCK19. In contrast, only 26% $ 36.8% (n # 2) of acinarexplants from Rac1ko animals transdifferentiated to duct-like complexes on TGF-! treatment (Figure 5). Moreover,the duct-like structures of Rac1ko explants were smallerand lined with several cell layers. The less abundant CK19-expressing cells were mostly surrounded by cell clusters

with retained acinar phenotype, suggesting incompletetransdifferentiation.

Pancreatic acinar explants from 4-week-old KrasG12D an-imals transdifferentiated spontaneously (88.1% $ 22.7%;n # 4) due to the presence of KrasG12D activation. TGF-!treatment enhanced the metaplastic capacity of theseacini to 100% conversion (Figure 5C and E), consistentwith the accelerated carcinogenesis in the KrasG12D;Tgfamodel.11 Ablation of pancreatic Rac1 in KrasG12D acinarcells led to a substantial impairment of the metaplasticprocess; only 9.2% $ 12.9% (n # 3) of acini transdiffer-entiated spontaneously to duct-like structures after 3 daysin collagen cultures and exhibited even greater impair-ment of metaplastic capacity on TGF-! treatment. More-over, the KrasG12D-derived duct-like cysts displayed a dis-tinct enlargement, complete loss of exocrine markerexpression, and strong CK19 staining. The explants fromKrasG12D;Rac1ko mice revealed only rare duct-like struc-tures of smaller size (Figure 5C, data not shown). Theseobservations indicate that KrasG12D- or EGFR-activated sig-naling was sufficient to initiate the ADM process, shownby CK19 expression of Rac1-lacking explants. However,the morphologic changes, necessary for formation ofduct-like structures, were strongly inhibited in the ab-sence of Rac1. Taken together, loss of Rac1 is sufficient toprevent the full conversion of acinar cells to a metaplasticductal epithelium.

Rac1-Mediated ADM Is Actin DependentAs described previously, ADM is strongly associ-

ated with F-actin rearrangements. However, other majorcell integrity markers, like E-cadherin and ZO-1, which arelocated in adherent and tight junctions, respectively, wereunaffected (Supplementary Figure 1E). Therefore, we pre-sumed that this inability to properly reorganize actinfilaments in Rac1-deficient acinar cells was the main causefor the impaired ADM. Indeed, in freshly isolated acinifrom WT and KrasG12D pancreata, F-actin was mainly api-cally located, whereas in acini from Rac1ko and KrasG12D;Rac1ko mice, F-actin showed an additional basal distribu-tion (Figure 6A). To test whether impaired ADM was actindependent, we conducted the 3D ADM experiments using2 membrane-permeable toxins; latrunculin A (LatA) andcytochalasin D (CytD) bind to F-actin and prevent actinpolymerization.37 Acinar explants from WT mice treatedwith either agent showed strong inhibition of TGF-!–induced ADM (Figure 6B and C). Similarly to the Rac1-deficient acini, immunofluorescence staining of these ex-plants exhibited weak CK19 expression after TGF-!treatment, indicating that mainly the plastic transdiffer-entiation of acini was affected (Figure 6B). These resultsstrongly suggest that ADM is actin dependent and Rac1mediated.

DiscussionRas and Ras-dependent pathways play a central

role in pancreatic carcinogenesis. Given the high fre-

BASIC

AN

DTRA

NSLA

TION

AL

PAN

CREAS


quency of oncogenic Kras mutations in preneoplastic le-sions and the recapitulation of the entire carcinogenicprocess in mouse models with an endogenous oncogenicKras allele, activation of Ras is believed to be the crucial

step in the initiation of PDAC development.2,38 Ras-de-pendent effectors such as mitogen-activated protein ki-nase (MAPK) and phosphatidylinositol 3-kinase (PI3K),fundamental for regulation of proliferation and/or apo-

Figure 4. ADM is accompanied by actin reorganization. Localization and redistribution of F-actin and CK19 during ADM in cerulein-inducedpancreatitis and KrasG12D-induced pancreata. Immunofluorescence single/double stainings on pancreatic cryosections. Dashed line visualizes acinias example. Scale bar # 50 "m. (A) F-actin is clearly apically located in WT acini (left panel) and redistributes basolaterally following induction ofpancreatitis (white arrowhead). This process is considerably reduced in Rac1ko mice. (B) F-actin is clearly apically located (white arrowhead, upperpanel) in normal acini and shows basolateral localization (white arrows, lower panel) in ADM lesions (confocal analysis). (C) Model for F-actin (red line)redistribution during ADM. The apical localized actin filaments redistribute to the basolateral side and concentrate at both sides of the metaplastictransdifferentiating cells. This process is supported by Rac1 activation in pancreatic tissue.

BASI

CA

ND

TRA

NSL

ATI

ON

AL

PAN

CREA

S


Figure 5. ADM in primary acinar epithelial explants is Rac1 dependent. Pancreatic acinar epithelial explants undergo transdifferentiation to ductalstructures in collagen cultures under TGF-! treatment (A) or spontaneously due to the presence of oncogenic KrasG12D (C, middle panel). Phasecontrast images (A, C middle panel) and immunofluorescence staining (B and C) for amylase and CK19. After isolation, all acinar cells expressamylase independently of the genotype (B and C, left panels). (B) WT explants convert to duct-like structures positive for CK19 (white arrowhead)when cultured for 5 days with recombinant human TGF-!. Acinar epithelial explants of mice lacking Rac1 show no or aberrant duct-like morphologyand retain amylase expression (gray arrowhead). (C) Acinar explants of KrasG12D mice convert spontaneously to duct-like structures already at day3 and express only CK19 (top, arrowheads). In contrast, explants from KrasG12D mice lacking Rac1 are significantly impaired in their ability to form

BASIC

AN

DTRA

NSLA

TION

AL

PAN

CREAS


ptosis, have been intensively studied and subjected tovarious efforts of targeted drug development.39 However,the function of other Ras downstream effectors in PDAC,including the Rho family of guanosine triphosphatases, isless well defined. To date, many Rho proteins have beendescribed in virtually all stages of tumorigenesis.26 Therole of individual Rho guanosine triphosphatases, such asfor Rac1, was predominantly reported in epithelial trans-formation and rearrangement of the cytoskeleton.22,23,40

Although the cell of origin in PDAC is still largelyunknown, recent evidence from sophisticated transgenicmouse models supports the concept that cells from vari-ous pancreatic compartments are capable of transforminginto a PanIN precursor.38,41 Thus, preneoplastic lesionscan derive from progenitor, centroacinar, adult acinar,and/or islet cells depending on intracellular and extracel-lular signals. This tremendous plasticity involves onco-genic Kras; however, the requirement of Ras-dependentsignaling is still ill defined. Because Rac1 is an importantregulator of the transformation process and cytoskeletalrearrangements,20,23,40 we characterized the role of Rac1 inKrasG12D-induced carcinogenesis using a functional geneticin vivo approach.

Given the normal histology of Rac1ko mice and noreconstitution by Rac2/3, Rac1 does not seem to be es-sential for either pancreatic development and adult tissuehomeostasis. In accordance with previous in vitro find-ings,42 weak acinar Rac1 expression increased in vivo uponCCK treatment, supporting the importance of Rac1 inacinar plasticity following inflammatory stimuli. The de-tectable membranous staining in ADM but not in PanINlesions suggests a dynamic expression pattern within asmall window of Rac1 activation. The comparably higherexpression in lesion-surrounding stromal cells is consis-tent with a gene expression study of human pancreaticsamples where Rac1 was found to be highly up-regulatedin PDAC.27 Interestingly, the stromal Rac1 expression wasalready suggested in this study, because up-regulated Rac1was found mainly in bulk PDAC tissue and not in tumoraspirates.

Despite the seemingly low expression, deficiency ofRac1 had an impressive effect in KrasG12D-induced carci-nogenesis. We observed not only a significant reduction inPanIN and PDAC development but also a significantlyprolonged survival of KrasG12D;Rac1ko mice. This evidentrole was not explainable by differences in proliferation orchanges in RAS/EGFR-dependent pathways essential forsurvival. Although Kissil et al found that loss of Rac1affects the proliferative capacity of KrasG12D baby mousekidney cells,30 developing acinar cells may react differentlyto the oncogenic pressure of KrasG12D and rely on addi-tional signals absent in in vitro conditions.

Initiation of pancreatic carcinogenesis goes along withwell-described changes in cellular morphology, startingwith ADM as a key event before development ofPanIN.2,3,38 Indeed, development of ADM was substan-tially reduced in KrasG12D;Rac1ko mice at the early age of 4weeks, suggesting necessity of Rac1 for the very first meta-plastic events in the acinar cells. Of note, this effect wasalso observable in the more aggressive triple-transgenicdominant negative KrasG12D; p53R172H;Rac1ko model, sup-porting the early carcinogenic requirement of Rac1. Fol-low-up of KrasG12D;Rac1ko mice showed a significantly pro-longed survival. Moreover, high-grade PanINs wereobserved only in one animal. This suggests that Rac1 hasa central function in development of PDAC, which is noteasily bypassed. Rac1-deficient PanINs showed no differ-ences in the expression of Ras-dependent effectors orgenes of the Notch and Hedgehog pathway. These lesionsshowed Cre-mediated recombination as proven by analy-sis of microdissected lesions. A likely explanation forlow-grade and, in the one case, high-grade PanIN devel-opment in our model would be through an ADM-inde-pendent route or from a yet unknown alternative preneo-plastic progenitor cell. The development of preneoplasticlesions and PDAC from different progenitor compart-ments is currently a widely discussed scenario and war-rants further investigation.

Besides activation of oncogenic Kras, additional signalsare believed to be required for preneoplastic precursordevelopment, such as inflammation-induced reprogram-ming of acinar cells.12,38 We noted a significantly reducedyet still existing ADM during cerulein-induced pancreati-tis. This observation suggests that Rac1 may play a pre-dominant role in Ras-dependent regulation of acinar plas-ticity, whereas cerulein-induced induction of ADM mightdepend on additional pathways such as EGFR and/orEGFR-downstream signaling events.5 Still, our results arein line with a recent study showing attenuation but notabolishment of cerulein-induced pancreatitis due to thechemical Rac1 inhibition.42 Furthermore, Binker et alshowed that blocking Rac1 activation by selective com-pound NSC23766 inhibited Rac1 translocation from cy-toplasm to the membrane of acinar cells, a process nec-essary for actin rearrangements. The actin cytoskeletonhas been previously described to play an important role inacinar cells, including mediation of exocytosis, secretion,and modulation of cell shape.43 Given the important roleof Rac1 in cytoskeletal reorganization, we focused onactin redistribution during ADM and PanIN develop-ment. Indeed, we found apically localized actin filamentsto localize basolaterally and concentrate at both sides ofthe metaplastic transdifferentiating cells. This loss of cel-lular polarity, along with the increasing metaplasia, was

duct-like structures. All nuclei are stained with Hoechst 33342. Scale bar # 50 "m. (D and E) Quantification of acinar explants from WT Rac1ko,KrasG12D, or KrasG12D;Rac1ko mice as described previously (one representative experiment). The difference in the amount of duct-like structures issignificant in D (WT/Rac1ko, P # .001; WT/Rac1ko " TGF-!, P # .0002) and in E (KrasG12D/KrasG12D;Rac1ko, P # .01; KrasG12D/KrasG12D;Rac1ko "TGF-!, P # .0006).

BASI

CA

ND

TRA

NSL

ATI

ON

AL

PAN

CREA

S


present in cerulein-induced pancreatitis as well as inKrasG12D-activated cells undergoing preneoplastic transdif-ferentiation. Furthermore, the similarity in phenotype ofRac1-deficient and chemically actin-disrupted cultured

acinar cells regarding ADM suggests that this axis plays afinely tuned role in acinar cell morphology. Thus, wepropose a model for F-actin redistribution during ADM(Figure 4C), where increasing ADM-dependent cytoskel-

Figure 6. Intact actin polymerization is crucial for ADM in primary acinar explant cultures. (A) In primary acinar explants from WT and KrasG12D mice,F-actin is apically localized (white arrowheads). Ablation of Rac1 leads to alterations in F-actin distribution visualized by an additional basal staining(gray arrowheads). Nuclei are stained with Hoechst33342. Confocal analysis, scale bar # 50 "m. (B) Phase contrast images and double immuno-fluorescence staining for amylase and CK19 of acinar epithelial explants treated with inhibitors of actin polymerization LatA and CytD. Note CK19expression in TGF-!–treated samples; however, there is a clear absence of ductal-like structures. Nuclei are stained with Hoechst 33342.Scale bar # 50 "m. (C) Quantification of duct-like or acinar structures shown in B at day 4. The absence of ductal-like structures indicates that ADMdepends on actin polymerization. Average of 2 ADM experiments performed in triplicate.

BASIC

AN

DTRA

NSLA

TION

AL

PAN

CREAS


etal changes require Rac1 overexpression and/or activa-tion. One caveat concerns actin function-independenteffects of the structurally different polymerization inhib-itors LatA and CytD. Both drugs alter actin polymeriza-tion and mechanical cell properties by different mecha-nisms. The concentrations of LatA and CytD used in thisstudy for inhibition of ADM were well within the rangeneeded for disruption of the actin cytoskeleton. However,we cannot rule out additional effects of both inhibitors,such as messenger RNA regulation and protein/glucosetransport, which may have an impact on the morphologicchanges during ADM.

Although Rac1-dependent regulation of the actin cyto-skeleton is well documented, most studies were per-formed in vitro using overexpression of dominant nega-tive or constitutively active mutants in cultured cells,leaving the functional role in vivo undefined. Our findingsof Rac1-dependent basolateral actin redistribution in pan-creatic acini in vivo as well as in 3D cultures representclear evidence for an important functional role of physi-ologically expressed Rac1 in this process. Moreover, wepropose that actin polymerization is a central and criticaldeterminant of acinar cell plasticity and metaplastic trans-differentiation during carcinogenesis. Thus, modulationof Rac1 activation and actin regulation is a potentiallypromising new target to interfere with the initiation ofpancreatic carcinogenesis.

Supplementary MaterialNote: To access the supplementary material

accompanying this article, visit the online version ofGastroenterology at www.gastrojournal.org, and at doi:10.1053/j.gastro.2011.04.043.

References1. Hezel AF, Kimmelman AC, Stanger BZ, et al. Genetics and biology

of pancreatic ductal adenocarcinoma. Genes Dev 2006;20:1218–1249.

2. Hruban RH, Adsay NV, Albores-Saavedra J, et al. Pancreatic intra-epithelial neoplasia: a new nomenclature and classification sys-tem for pancreatic duct lesions. Am J Surg Pathol 2001;25:579–586.

3. De La OJ, Emerson LL, Goodman JL, et al. Notch and Kras repro-gram pancreatic acinar cells to ductal intraepithelial neoplasia.Proc Natl Acad Sci U S A 2008;105:18907–18912.

4. Hingorani SR, Petricoin EF, Maitra A, et al. Preinvasive and inva-sive ductal pancreatic cancer and its early detection in the mouse.Cancer Cell 2003;4:437–450.

5. Means AL, Meszoely IM, Suzuki K, et al. Pancreatic epithelialplasticity mediated by acinar cell transdifferentiation and genera-tion of nestin-positive intermediates. Development 2005;132:3767–3776.

6. Strobel O, Dor Y, Alsina J, et al. In vivo lineage tracing defines therole of acinar-to-ductal transdifferentiation in inflammatory ductalmetaplasia. Gastroenterology 2007;133:1999–2009.

7. Wagner M, Greten FR, Weber CK, et al. A murine tumor progres-sion model for pancreatic cancer recapitulating the genetic alter-ations of the human disease. Genes Dev 2001;15:286–293.

8. Wagner M, Luhrs H, Kloppel G, et al. Malignant transformation ofduct-like cells originating from acini in transforming growth factortransgenic mice. Gastroenterology 1998;115:1254–1262.

9. Blaine SA, Ray KC, Anunobi R, et al. Adult pancreatic acinar cellsgive rise to ducts but not endocrine cells in response to growthfactor signaling. Development 2010;137:2289–2296.

10. Ji B, Tsou L, Wang H, et al. Ras activity levels control the devel-opment of pancreatic diseases. Gastroenterology 2009;137:1072–1082, 1082 e1–6.

11. Siveke JT, Einwachter H, Sipos B, et al. Concomitant pancreaticactivation of Kras(G12D) and Tgfa results in cystic papillary neo-plasms reminiscent of human IPMN. Cancer Cell 2007;12:266–279.

12. Guerra C, Schuhmacher AJ, Canamero M, et al. Chronic pancre-atitis is essential for induction of pancreatic ductal adenocarci-noma by K-Ras oncogenes in adult mice. Cancer Cell 2007;11:291–302.

13. Carriere C, Young AL, Gunn JR, et al. Acute pancreatitis markedlyaccelerates pancreatic cancer progression in mice expressing onco-genic Kras. Biochem Biophys Res Commun 2009;382:561–565.

14. Morris JP IV, Cano DA, Sekine S, et al. Beta-catenin blocks Kras-dependent reprogramming of acini into pancreatic cancer precur-sor lesions in mice. J Clin Invest 2010;120:508–520.

15. Didsbury J, Weber RF, Bokoch GM, et al. rac, a novel ras-relatedfamily of proteins that are botulinum toxin substrates. J Biol Chem1989;264:16378–16382.

16. Jaffe AB, Hall A. Rho GTPases: biochemistry and biology. Annu RevCell Dev Biol 2005;21:247–269.

17. Etienne-Manneville S, Hall A. Rho GTPases in cell biology. Nature2002;420:629–635.

18. Guo F, Gao Y, Wang L, et al. p19Arf-p53 tumor suppressor path-way regulates cell motility by suppression of phosphoinositide3-kinase and Rac1 GTPase activities. J Biol Chem 2003;278:14414–14419.

19. Ridley AJ, Schwartz MA, Burridge K, et al. Cell migration: integrat-ing signals from front to back. Science 2003;302:1704–1709.

20. Bi Y, Page SL, Williams JA. Rho and Rac promote acinar morpho-logical changes, actin reorganization, and amylase secretion.Am J Physiol Gastrointest Liver Physiol 2005;289:G561–G570.

21. Chiariello M, Marinissen MJ, Gutkind JS. Regulation of c-mycexpression by PDGF through Rho GTPases. Nat Cell Biol 2001;3:580–586.

22. Olson MF, Ashworth A, Hall A. An essential role for Rho, Rac, andCdc42 GTPases in cell cycle progression through G1. Science1995;269:1270–1272.

23. Quinlan MP. Suppression of epithelial cell transformation and induc-tion of actin dependent differentiation by dominant negative Rac1,but not Ras, Rho or Cdc42. Cancer Biol Ther 2004;3:65–70.

24. Qiu RG, Chen J, Kirn D, et al. An essential role for Rac in Rastransformation. Nature 1995;374:457–459.

25. Dokmanovic M, Hirsch DS, Shen Y, et al. Rac1 contributes totrastuzumab resistance of breast cancer cells: Rac1 as a poten-tial therapeutic target for the treatment of trastuzumab-resistantbreast cancer. Mol Cancer Ther 2009;8:1557–1569.

26. Karlsson R, Pedersen ED, Wang Z, et al. Rho GTPase function intumorigenesis. Biochim Biophys Acta 2009;1796:91–98.

27. Crnogorac-Jurcevic T, Efthimiou E, Capelli P, et al. Gene expres-sion profiles of pancreatic cancer and stromal desmoplasia. On-cogene 2001;20:7437–7446.

28. Espina C, Cespedes MV, Garcia-Cabezas MA, et al. A critical rolefor Rac1 in tumor progression of human colorectal adenocarci-noma cells. Am J Pathol 2008;172:156–166.

29. Wang Z, Pedersen E, Basse A, et al. Rac1 is crucial for Ras-dependent skin tumor formation by controlling Pak1-Mek-Erk hy-peractivation and hyperproliferation in vivo. Oncogene 2010;29:3362–3373.

30. Kissil JL, Walmsley MJ, Hanlon L, et al. Requirement for Rac1 ina K-ras induced lung cancer in the mouse. Cancer Res 2007;67:8089–8094.

31. Soriano P. Generalized lacZ expression with the ROSA26 Crereporter strain. Nat Genet 1999;21:70–71.

BASI

CA

ND

TRA

NSL

ATI

ON

AL

PAN

CREA

S


http://www.gastrojournal.org

http://dx.doi.org/10.1053/j.gastro.2011.04.043

http://dx.doi.org/10.1053/j.gastro.2011.04.043

32. Hingorani SR, Wang L, Multani AS, et al. Trp53R172H andKrasG12D cooperate to promote chromosomal instability andwidely metastatic pancreatic ductal adenocarcinoma in mice. Can-cer Cell 2005;7:469–483.

33. Glogauer M, Marchal CC, Zhu F, et al. Rac1 deletion in mouseneutrophils has selective effects on neutrophil functions. J Immu-nol 2003;170:5652–5657.

34. Siveke JT, Lubeseder-Martellato C, Lee M, et al. Notch signaling isrequired for exocrine regeneration after acute pancreatitis. Gas-troenterology 2008;134:544–555.

35. Mazur PK, Einwachter H, Lee M, et al. Notch2 is required forprogression of pancreatic intraepithelial neoplasia and develop-ment of pancreatic ductal adenocarcinoma. Proc Natl Acad Sci US A 2010;107:13438–13443.

36. Sphyris N, Logsdon CD, Harrison DJ. Improved retention of zymo-gen granules in cultured murine pancreatic acinar cells and induc-tion of acinar-ductal transdifferentiation in vitro. Pancreas 2005;30:148–157.

37. Wakatsuki T, Schwab B, Thompson NC, et al. Effects of cytocha-lasin D and latrunculin B on mechanical properties of cells. J CellSci 2001;114:1025–1036.

38. Morris JP IV, Wang SC, Hebrok M. KRAS, Hedgehog, Wnt and thetwisted developmental biology of pancreatic ductal adenocarci-noma. Nat Rev Cancer 2010;10:683–695.

39. Oskouian B, Saba JD. Cancer treatment strategies targeting sph-ingolipid metabolism. Adv Exp Med Biol 2010;688:185–205.

40. Parri M, Chiarugi P. Rac and Rho GTPases in cancer cell motilitycontrol. Cell Commun Signal 2010;8:23.

41. Gidekel Friedlander SY, Chu GC, Snyder EL, et al. Context-depen-dent transformation of adult pancreatic cells by oncogenic K-Ras.Cancer Cell 2009;16:379–389.

42. Binker MG, Binker-Cosen AA, Gaisano HY, et al. Inhibition of Rac1decreases the severity of pancreatitis and pancreatitis-associatedlung injury in mice. Exp Physiol 2008;93:1091–1103.

43. Torgerson RR, McNiven MA. Agonist-induced changes in cellshape during regulated secretion in rat pancreatic acini. J CellPhysiol 2000;182:438–447.

Received November 19, 2010. Accepted April 15, 2011.

Reprint requestsAddress requests for reprints to: Jens T. Siveke, PD Dr Med, 2

Medizinische Klinik und Poliklinik, Klinikum rechts der Isar,Technische Universität München, Ismaninger Str 22, 81675München, Germany. e-mail: [email protected]; fax: (49) 89-41406794.

AcknowledgmentsThe authors thank M. Neuhofer and S. Ruberg for excellent

technical assistance.I.H. and C.L.-M. contributed equally to this work.

Conflicts of interestThe authors disclose no conflicts.

FundingSupported by grants from the Wilhelm Sander Stiftung

(2010.021.1 to J.T.S.) and the German Federal Ministry of Educationand Research (National Genomic Research Network [NGFN-Plus];01GS08115 to J.T.S. and R.M.S.).

BASIC

AN

DTRA

NSLA

TION

AL

PAN

CREAS


mailto:[email protected]

Supplementary Methods

Preparation of Pancreatic Epithelial ExplantsCulturePancreatic epithelial explants of adult mouse pan-

creas (4 – 6 wk) were established by modification of pre-viously published protocols.1 In brief, whole pancreas washarvested and digested twice in 1.2 mg/ml collagenase-VIII (Boehringer Mannheim, Mannheim, Germany). Fol-lowing multiple wash steps with McCoy=s medium con-taining soybean trypsin inhibitor (SBTI, 0.2 mg/ml),digested samples were filtered through a 100 "m filter,resuspended in culture medium (Waymouht MB 752/1supplemented with 0.1% BSA, 0.2 mg/ml SBTI; 50 ug/mlbovine pituary extract, 10 ug/ml Insulin, 5 ug/ml trans-ferrin, 6.7 ng/ml selenium in 30% FCS) and allowed torecover for 1h at 37°C. Thereafter, cells were pellet andresuspended in culture medium media supplementedwith penicillin G (1000 U/ml), streptomycin (100 "g/ml),amphotericin B, 0.1% FCS, and an equal volume of rattail collagen type I (RTC) (BD Bioscience). The cellular/RTC suspension, was immediately plated on plates pre-coated with 2.5 mg/ml of rat tail collagen type I. In someexperiments recombinant human TGF-! (rh TGF!, R&DSystems) was added (final concentration 50 ng/ml). Forinhibition of actin, polymerisation cytochalasin D orlatrunculin A (Sigma) was added at day 2 both at finalconcentration of 2 and 0.2 "M.

For each genotype, experiments with acinar epithelialexplants were performed using different numbers ofmice: wt n#13; wt " actin inhibitors n#2; Rac1ko n#4;KrasG12D n#10; KrasG12D,Rac1ko n#7. Results were quali-tatively reproducible for each mouse. For quantificationacinar explants were seeded in triplicates; cells clusterswere counted from at least 3 optical fields/well. Thisquantification was performed for the number of micereported in the main text.

Quantification of Proliferation andADM/PanINFor proliferation analysis, mice were injected i.p.

with 100 mg/kg BrdU (Sigma) 2 hours prior to sacrifice.Three paraffin sections per mice from 3 animals pergenotype were stained using anti-BrdU antibody. To as-sess the proliferative capacity, 3 optical fields per section(in total 18/genotype) were analyzed using AxioVision 4.8software and displayed as the percentage of BrdU positivecells per region.

For PanIN/ADM quantification, H&E or CK19 stain-ings were quantified using 1 representative slide permouse. At least 10 pictures were taken from each slide,and the amount of lesions was calculated manually.

Glucose Tolerance TestGlucose tolerance test (GTT) was performed on

mice after 12 hours overnight fasting. Glucose was in-jected i.p. (2 g/kg, using 20% glucose solution). Blood

glucose levels were assessed by collecting tail blood, andglucose level data were recorded at indicated time points.

Immunofluorescence of Acinar EpithelialExplantsFor immunofluorescent labeling of explanted

pancreatic tissue, collagen gels containing explanted pan-creas were fixed in the chamber slides in 4:1 methanol:DMSO overnight at 4°C, washed and stored at –20°C in100% methanol. For phalloidin staining, collagen diskswere fixed in 1:1:2 PFA 8%:PBS:DMSO. Cultures in col-lagen disks were rehydrated and washed in PBS. Forimmunofluorescent staining, collagen gels were perme-abilized with TritonX-100 0.1% (Sigma) for 5 minutes atRT and washed in PBT (PBS " TritonX-100 0.5%). Gelswere blocked with 5% normal goat serum in PBBT (PBT" BSA 2%) for 2 hours at room temperature, then incu-bated sequentially with the primary (see SupplementaryTable 1 for a list of primary antibodies) and secondaryantibodies diluted in PBSBT, overnight at 4°C. Followingeach antibody, gels were washed in PBT. After the finalovernight incubation, the cultures were washed in PBTand in PBS, then counterstained with Hoechst 33342 1"g/ml (Invitrogen) and washed in PBS. Images were cap-tured on a Zeiss inverse fluorescente microscope or witha Leica TCS SP5 confocal microscope.

Immunohistochemistry andImmunofluorescenceImmunoperoxidase staining was performed on 4%

PFA/PBS fixed paraffin-embedded tissue slides using theVectastain ABC Elite kit (Vector Labs) following manu-facturer’s instructions. See Supplementary Table 1 for alist of primary antibodies used.

For immunofluorescence staining, cryosections (6 "M)were fixed with PFA 2% for 20 min and permeabilizedwith Triton X-100 0.1 % for 5 minutes. The first antibod-ies (Supplementary Table 1) were incubated overnight at4°C, and secondary antibodies were incubated for 1 h(Alexa dyes, Invitrogen) at room temperature. Sectionswere mounted with Vectashield hard mounting mediumcontaining DAPI (Vector Laboratories) and examined us-ing a Zeiss inverse fluorescent microscope or with a LeicaTCS SP5 confocal microscope.

Western Blotting and Rac1 Activity AssayA piece of pancreas was frozen in liquid nitrogen

immediately after sacrificing the mice. GTP-bound Rac1was pulled down from whole pancreas tissue lysate inMLB-buffer according to the manufacture instruction(Upstate, Millipore). The visualization of total and GTP-bound Rac was performed using standard Western-blot.For whole-tissue pancreatic protein lysates were preparedby homogenizing pancreatic tissue in RIPA buffer con-taining complete protease and phosphatase inhibitors(Roche) and separated on a 7.5-15% polyacrylamide-SDS

August 2011 TARGETING RAC1 IN PANCREATIC CANCER 730.e1

gel, transferred to PVDF membranes and blocked in 5%BSA or skim milk followed by antibody incubation over-night at 4°C. Antibodies used are shown in Supplemen-tary Table 1. Antibody binding was visualized usinghorseradish peroxidase-labeled secondary antibodies andECL reagent (Amersham).

RT-PCR AnalysisTotal RNA was extracted from shock-frozen pancre-

atic tissues of at least 3 different regions with the RNeasyKit (Qiagen). cDNA was synthesized with SuperScript IIreverse transcriptase (Invitrogen). Real-time PCR analysiswas performed on a Gene Amp 7700 sequence detectionsystem from Applied Biosystems using SYBR Green PCRmix. PCR reactions were performed as described previ-ously17 using the %CT method. Primers: Rac1-Forward5=-CGGAGCTGTTGGTAAAACCTG; Rac1-Reverse ATAG-GCCCAGATTCACTGGT; Rac2-Forward TGGGCCTCA-

GATGCAATGCAG Rac2-Reverse ATAGTCCTCCTGACC-TGCGG; Rac3-Forward ATCAAGTGCGTGGTGGTTGGCRac3-Reverse TCCCGCAGCCGTTCAATCGT; GAPDH-Forward TACTAGCGGTTTTACGGGCG; GAPDH-ReverseTCGAACAGGAGGAGCAGAGAGCGA. For other primerssee.2 Expression of Rac3 was tested on brain cDNA as aposotive control.

References1. Means AL, Meszoely IM, Suzuki K, et al. Pancreatic epithelial

plasticity mediated by acinar cell transdifferentiation and genera-tion of nestin-positive intermediates. Development 2005;132:3767–3776.

2. Mazur PK, Einwachter H, Lee M, et al. Notch2 is required forprogression of pancreatic intraepithelial neoplasia and develop-ment of pancreatic ductal adenocarcinoma. Proc Natl Acad SciU S A 2010;107:13438–13443.

730.e2 HEID ET AL GASTROENTEROLOGY Vol. 141, No. 2

Supplementary Figure 1. Pancreas specific ablation of Rac1 does not lead to gross histological or functional changes. (A) Diagram of Cre-mediated recombination and PCR analyses of fresh isolated tissue samples from a 6-month-old Rac1ko mouse. Rac1 is specifically deleted inpancreas tissue, as shown by the selective presence of the recombined allele. (B) Immunfluorescence staining and immunoblot analysis of Rac1expression in 4-week-old pancreatic tissue from wt and Rac1ko animals. Rac1 expression is abrogated specifically in the pancreas. Dashed lineindicates an acinus. Scale bar # 50 "m. (C) Pancreas morphology of a 6-month-old Rac1ko mouse is normal, as shown by H&E staining. I # islet;d # duct. Scale bar # 50 "m. (D) Analysis of the endocrine function in 4-week-old Rac1ko mice. Glucose tolerance test reveal only slight delay inglucose uptake (n # 6/group, P # .2386) in animal lacking Rac1. Representative example of 2 independent experiments. IHC analysis shows normalexpression of endocrine (insulin and glucagon) markers. (E) Immunofluorescence analysis of E-cadherin and ZO-1 expression in pancreata of wt andRac1ko mice show normal acinar cellular integrity and intact adherent junctions. Bar # 50 "m.


Supplementary Figure 2. Rac1 expression in murine pancreatic acinar explants and murine and human pancreatic tissue. (A) Acinar epithelialexplants from wt or KrasG12D mice were treated with (") or without (&) 50 ng/ml rhTGF! for 1 hour. Immunofluorescence staining for Rac1 showsweak expression after the isolation and a strong induction of Rac1 expression in the acini upon TGF! treatment. Nuclei are counterstained withHoechst33342. Scale bar # 50 "m. (B) Immunofluorescence double stainings for Rac1 and amylase (top) or CK19 (bottom) in 6-month-oldpancreatic tissue of KrasG12D mice. Top: Rac1 is weakly expressed in normal acini and localizes basally while ADM progression (white arrow). At thesame time, Rac1 expression increases in surrounding stromal tissue. Bottom: In PanIN lesions Rac1 expression is nearly absent. Scale bar # 50 "m.(C) Immunohistological staining for Rac1 in 6-month-old KrasG12D pancreas shows absent or low expression in PanINs and strong staining in stromalcells. Scale bar # 50 "m. (D) Immunohistological staining for Rac1 in human pancreatic tissue reveals absent or only weak expression of Rac1 inPanIN lesions and PDAC cells, while stromal and inflammatory cells show strong positive staining.


Supplementary Figure 3. Loss of Rac1 impairs acino-ductal metaplasia and PanIN development in KrasG12D; p53R172H; Rac1ko mice. (A)Macroscopic photographs of KrasG12D;p53R172H;Rac1"/fl versus KrasG12D; p53R172H;Rac1ko pancreata at the age of 3 months. Note the smaller sizein KrasG12D; p53R172H;Rac1ko pancreas. Ruler gap # 1 mm. (B) Deletion of Rac1 reduces mPanIN burden in KrasG12D;p53R172H;Rac1ko micecompared to KrasG12D;p53R172H;Rac1wt/fl animals. H&E staining of 3-month-old KrasG12D;p53R172H;Rac1wt/fl pancreatic tissue (top) shows diffusedmPanIN/ADM lesions development spread in the whole organ. In contrast, pancreata of KrasG12D;p53R172H;Rac1ko mice (bottom) reveal only rarefocal mPanIN lesions. Scale bar # 50 "m. (C) Representative examples of end-point histology (H&E staining) of pancreatic tissue from KrasG12D;Rac1ko mice. Top panel: high grade PanIN lesions. Middle panel: normal pancreas with adipose replacement. Bottom panel: normal pancreas andlymphoma. (D) PCR for Rac1 wt, Rac1 LOX, and Rac1 recombined from microdissected samples as indicated in the figure. (E) Table summarizingthe phenotype of 8 KrasG12D;Rac1ko mice at the end-point stage.


Supplementary Figure 4. Deletion of Rac1 has no impact on central RAS/EGFR-dependent effectors and proliferation rate. (A) Quantification ofproliferation rate in 7-day-old pancreas of wt, KrasG12D and KrasG12D;Rac1ko animals (n # 3/group, P # .65); in ADM and PanINs of 6-month-oldpancreata of wt, KrasG12D (n#3) and KrasG12D;Rac1ko (n # 4) animals (ADM, P # .4; PanIN, P # .5) and in 3-month-old KrasG12D;p53R172H;Rac1wt/fl

(n # 3) and KrasG12D; p53R172H;Rac1ko (n#2) animals (ADM, P # .81; PanIN P # .88). (B) Immunoblot analyses of the main RAS/EGFR pathwayeffectors in 4-week-old whole-pancreatic lysates reveal no significant changes. Hsp90 serves as loading control. (C) Immunohistochemical analysisof exocrine pancreas and PanINs from KrasG12D and KrasG12D;Rac1ko animals shows no difference for the centroacinar marker Hes1 and in P-Erk1/2expression. (D) Relative mRNA expression of Pdx1, Notch and Hedgehog signaling genes in the pancreata of 1-month-old KrasG12D and KrasG12D;Rac1ko animals (left panel). No significant differences between the genotypes are notable.


Supplementary Figure 5. Acino-ductal metaplasia is accompanied by actin reorganisation. Localization and redistribution of F-actin and amylaseduring ADM in pancreas of 6-month-old KrasG12D mouse. Immunofluorescence double staining on pancreatic cryosections. F-actin is clearly apicallylocalized (white arrowhead, top) in the normal acini and shows basolateral localization (white arrowheads, bottom) in ADM lesions. Confocal pictures.Dashed line visualizes acini as example. Scale bar # 50 "m.

Supplementary Table 1. Antibodies Used for Indicated Applications in This Study

Ab Species Dilution Application Company

Hsp90 Rabbit 1:1000 WB Cell SignalingActin Mouse 1:5000 WB SigmaAmylase Goat 1:1000, 1:200 WB/IHC/IF Santa CruzE-cadherin Goat 1:200 IHC/IF R&DBrdU Rat 1:250 IHC SerotecMuc5 Mouse 1:100 IHC NeomarkersRac1 Mouse 1:1000 WB UpstateRac1 Mouse 1:100 IHC, IF BDInsulin Rabbit 1:200 IHC DAKOGlucagon Gp 1:200 IHC DAKOCK19 (TROMAIII) Rat 1:200 IHC DSHBE-cadherin Rabbit 1:100 IF Cell SignalingZO-1 Rabbit 1:200 IF Invitrogen


Early Requirement of Rac1 in a Mouse Model of Pancreatic Cancer

Documents

Transcript of Early Requirement of Rac1 in a Mouse Model of Pancreatic Cancer