neuroendocrine differentiation in the 12t-10 transgenic prostate mouse model mimics endocrine...

The Prostate 68:50 ^ 60 (2008)

NeuroendocrineDifferentiation inthe12T-10Transgenic ProstateMouseModelMimics Endocrine

Differentiationof Pancreatic BetaCells

Aparna Gupta,1 Yongqing Wang,1 Christopher Browne,1 Susan Kim,2

Tom Case,1 Manik Paul,1 Marcia L. Wills,3 and Robert J. Matusik1,4,5,6,7,8*1DepartmentofUrologic Surgery,Vanderbilt UniversityMedical Center,Nashville,Tennessee

2MeharryMedical College,Nashville,Tennessee3Departmentof Pathology,Vanderbilt UniversityMedical Center,Nashville,Tennessee

4Departmentof CellandDevelopmental Biology,Vanderbilt UniversityMedical Center,Nashville,Tennessee5Departmentof Cancer Biology,Vanderbilt UniversityMedical Center,Nashville,Tennessee

6TheVanderbilt Prostate Cancer Center,Vanderbilt UniversityMedical Center,Nashville,Tennessee7Center for Reproductive Biology Research,Vanderbilt UniversityMedical Center,Nashville,Tennessee

8TheVanderbilt-IngramCancer Center,Vanderbilt UniversityMedical Center,Nashville,Tennessee

BACKGROUND. Neuroendocrine (NE)prostate cancerdevelops as anaggressivedisease thatdoes not respond to androgen ablation therapy. It has been demonstrated that the paracrineaction of NE cells facilitates the progression of androgen dependent adenocarcinoma to anandrogen independent state, suggesting a significant role for NE cells during failure ofandrogen ablation therapy.METHODS. To investigate the pathways that are involved in NE differentiation of prostatecancer, we have looked at the expression of genes known to be involved in endocrinedifferentiation of b-cells in the pancreas. This study has been performed using the NE prostatecancer mouse model (12T-10) and the derivative allograft model (NE-10).RESULTS. Immunohistochemical studies have shown that the neuroendocrine prostatetumors express the transcription factors Foxa2, mouse achaete-scute homolog-1 (mash-1),neurogenin3 (Ngn3) andNkx2.2. These tumors show a loss of hairy/enhancer of split (Hes-1), agene that inhibits NE differentiation. Human NE prostate cancers also express Foxa2 andhumanachaete-scute homolog-1 (HASH-1). These genes are expressed inNEprostate tumors inthe similar sequential manner as they appear in a pancreatic b-cell endocrine differentiation.Foxa2 expression is detected in early prostatic intraepithelial neoplasia (PIN). Mash-1expression is detected in a few clusters within low grade PIN lesions and Nkx2.2 expressionis rarely detected in individual scattered cells within the PIN lesion. Ngn3 andNkx2.2 frequently appear in the invasive NE cancer. Subsequent NE metastasis to lung andliver show a distinct gene expression pattern. The lungmetastasis expresses Ngn3 but does notexpress Nkx2.2 whereas liver metastases do not express Ngn3 but express Nkx2.2.CONCLUSIONS. These results suggest thatNgn3 andNkx2.2 expression aremarkers for site-specificmetastasis and/or transcriptionally regulated genes that are required for organ-specific

Abbreviations: NE, neuroendocrine; NED, neuroendocrine dif-ferentiation; PIN, prostatic intraepithelial neoplasia; AR, androgenreceptor; RT-PCR, reverse transcription-polymerase chain reaction;LGPIN, low grade prostatic intraepithelial neoplasia, HGPIN, highgrade prostatic intraepithelial neoplasia.

Grant sponsor: National Institutes of Health; Grant numbers: R01-CA76142, R01-AG023409.

*Correspondence to: Robert J. Matusik, PhD, Department of UrologicSurgery, A-1302 Medical Center North, Vanderbilt UniversityMedical Center, Nashville, TN 37232-2765.E-mail: [email protected] 30 May 2007; Accepted 17 July 2007DOI 10.1002/pros.20650Published online 14 November 2007 in Wiley InterScience(www.interscience.wiley.com).

� 2007Wiley-Liss, Inc.

metastasis. This study indicates that a pathway similar to pancreatic b-cell differentiation isinvolved inNE differentiation of prostate cancer. Prostate 68: 50–60, 2008. # 2007Wiley-Liss, Inc.

KEY WORDS: prostate; neuroendocrine differentiation; 12T-10; beta cells; metastasis;Foxa2

INTRODUCTION

In recent years, neuroendocrine differentiation(small cell carcinoma) in prostate cancer has beenreceiving increasing attention. Neuroendocrine tumoris an androgen insensitive tumor and neuroendocrinedifferentiation in prostatic adenocarcinoma couldbe one of the many possible mechanisms by whichprostate cancers evolve to androgen insensitive state.Neuroendocrine differentiation is present focally invirtually all cases of prostate cancer. The number ofcells detected in each case varies based on the tissuefixation, sectioning of the tissue, antibodymethod usedand the number of tissue sections examined [1]. Somestudies have reported as many as 92% of prostatecancers contain some NE cells [1,2]. On the contrary,others have reported much lower numbers such as25% of the prostate cancers have some NE cells. It isimportant to better understand the role of neuro-endocrine differentiation of prostate cancer since thegreater the proportion of neuroendocrine phenotypeaccompanying prostate adenocarcinoma, the poorerthe prognosis for the patient and the greater thelikelihood that the tumor will progress to androgenindependence [3–5]. Also, some metastases fromconventional ductal prostate cancers exhibit a neuro-endocrine phenotype [6]. Recent research shows that, aNE-10 allograft model (established from the 12T-10transgenic model of NE prostate cancer), will promotegrowth of the LNCaP xenograft in castrated mice [7].This study demonstrated that the neuroendocrinetumor secretions have the potential to maintainandrogen dependent LNCaP cells in the absence ofandrogens [7]. Further, the NE-10 tumor promotespulmonarymetastasis of thehumanprostate cancer cellline LNCaP xenograft when transplanted together in aathymic nude mice [16]. Understanding the molecularmechanisms that lead to NE differentiation (NED)would be an important step towards early diagnosisand effective treatment for this aggressive and andro-gen insensitive form of prostate cancer.

In this study, we have used the 12T-10 neuroendo-crine mouse model of prostate cancer. This transgenicline was created by targeting expression of the SV-40large T-antigen by the probasin promoter to prostaticluminal epithelial cells (8). With advancing age, the12T-10 mouse prostate tumor progresses throughstages starting with low-grade prostatic intraepithelialneoplasia, to high-grade prostatic intraepithelial

neoplasia, and advancing to microinvasion, invasivecarcinoma, poorly differentiated carcinoma with NED,and finally metastasis. NE metastasis to the liver andlung are particularly common in old mice [8]. The NE-10 allograft model was developed from the prostate ofthe 12T-10 transgenicmouse. This 12T-10prostatic tissuewas subcutaneously transplanted into an athymicnude mouse to form the NE-10 allograft which retainedthe same histological features as seen in the primaryNEtumor of the 12T-10 model [9]. The allograft alsodemonstrates NE metastasis to the lung and liver ofthe athymic mouse hosts [9].

The results of this study suggest that the NEdifferentiation of the prostate follows the same path-way as the endocrine differentiation of b-cells in thepancreas. The transcription factors, Foxa2, mash-1,Ngn3, and Nkx2.2, known to be involved in endocrinedifferentiation, are also expressed in the NE prostatein a sequential manner with the early PIN lesionsexpressing Foxa2 and/or mash-1 and the poorlydifferentiated NE cancers expressing Ngn3 andNkx2.2 as well. Notch signaling prevents aberrantendocrine differentiation of pancreatic beta cells byactivating Hes-1 that represses pro-endocrine geneNgn3 [10,11]. Loss of Notch signaling results in downregulation of Hes-1 [24,25]. RT-PCR results show thatHes-1 expression is lost in these NE prostate tumorsand consequently Ngn3 and Nkx2.2 expression isswitched on. Also, this study shows a difference inthe expression of Ngn3 and Nkx2.2 in the liver andlung NE metastases, with liver metastases expressingNkx2.2 and not Ngn3 and lung metastases expressingNgn3 and not Nkx2.2. This suggests that metastasis ofNE tumors to the liver and lung involves distinctlydifferent cell populations that express a specific geneprofile. This implies that the organ specific metastasisexpress genes that allow them survival advantages inthe surrounding tissue microenvironment.

MATERIALSANDMETHODS

12T-10 TransgenicMouse Line

LPB-Tag transgenic mice lines were establishedwith the 50-flanking region of the rat LPB promoter(�11,500 to þ28 bp) [12] linked to the SV40-Tag genedeletion mutant (d1 2005), which removed the expres-sion of the small Tag. Seven transgenic lines wereestablished and maintained in the CD1 mouse strain

The Prostate DOI 10.1002/pros

NeuroendocrineDifferentiation of Prostate Cancer 51

[13]. The 12T-10 transgenic line showed the slowestneoplastic prostate growth rate of the seven LPB-Tagtransgenic lines and developedNE features [8]. TheNEcarcinomas occurred in the dorsolateral and ventrallobes and were generally androgen receptor negative.The 12T-10 line develops LowGrade PIN (LGPIN) andHigh Grade PIN (HGPIN) in animals ranging from theage of 2–5 months [8]. The HGPIN lesions, in micearound 5 months old, began to show NED. Metastasesare detected in 6–14 months old animals. The mostcommon sites of metastases are regional lymphnodes, liver, and lung. These metastases are poorlydifferentiated neuroendocrine cancer [8].

NE-10AllograftModel

The primaryNE prostate from the 12T-10 transgenicline was implanted subcutaneously (s.c.) in immune-compromised male athymic nude mice. The trans-planted tissuewas recovered after 18weeks and a smallpiece was passed to another nude mice thus maintain-ing the NE-10 line [9]. The histological analysis of theNE allograft showed the same features as the primaryNE prostate from the 12T-10 model [9]. The earlypassages (p1-15) show metastases to liver and micrometastases to lung, which are histologically similar tothe metastasis seen in the 12T-10 mice.

Tissue Preparation

Mice were sacrificed by cervical dislocation after theinhalation of an anesthetic agent according to the policyof the Vanderbilt University Animal Care and UseCommittee. The prostates were generally dissectedinto four different lobes (ventral, lateral, dorsal, andanterior lobe) under a dissecting microscope. When itwasnot possible to separate the lateral anddorsal lobes,the tissue was taken together as the dorsolateral lobe.Regional lymphnodes, liver, lung, spleen, kidney, bone(lumbar vertebrae), adrenal glands, brain, bulboure-thral glands, and s.c.NE-10 tumorswere also harvestedfor histological examination. Tissues were fixed in10% buffered formalin and processed and embedd-ed in paraffin using standard techniques. Paraffin-embedded tissues were cut at 5 mm, and sections wereeither stained with Haematoxylin and Eosin (H&E) orwith relevant antibodies for immunohistochemicalanalysis.

Immunohistochemistry

Slides were deparaffinized by immersing in xylenetwice for 10 min each and hydrated by immersing in aseries of 100%, 95%, 70%, 50% ethanol, and one time indH2O for 5 min each. Slides for histological analysiswere stained with H&E by standard methods, with

generally 3–4 sections reviewed per specimen. ForFoxa2, synaptophysin, T-antigen, AR, mash-1, Ngn3,and Nkx2.2 immunostaining, antigen retrieval wasachieved by micro waving in antigen unmaskingsolution (Catalog # H-3300, Vector Laboratories, Inc.,Burlingame, CA) for 30 min and the slides were thenequilibrated at room temperature for 1 hour. Endo-genous peroxidase activity was blocked by peroxidaseblocking reagent (Dako) 30 min followed by washing inPBS (pH 7.4). After rinsing with PBS, the slides wereplaced in blocking solution (goat, horse or rabbitserum as appropriate) for 20 min to block nonspecificbinding of antibody to the tissue. Sections wereincubated with primary antibody overnight at 48C.The following primary antibodies were used (withthe indicated dilutions in PBS): Foxa2, P19, goatantibody (Santa Cruz Biotechnology Inc., 1:1,000);synaptophysin (Cat# 611880, BD Biosciences Pharm-ingen, 1:500); Tag (SV40 T-Ag, monoclonal mouse IgGOncogene, 1:1,000);AR,N-20 (SantaCruzBiotechnologyInc., 1:1,000); mash-1 (Cat #556604, BD BiosciencesPharmingen, 1:500), Ngn3 (rabbit, 1:6,000) antibodywas received from Dr. Michael German DiabetesCenter in University of California, San Francisco;Nkx2.2 antibody (mouse, 1:40 dilution, DevelopmentalStudies Hybridoma Bank, the University of Iowa). Therespective secondary antibodies were used at a dilutionof 1:200. Staining was visualized using VectastainABC kit (Vector Laboratories, Inc.) and 3, 30-diamino-benzidine tetrahydrochloride (Dako). Slides werecounterstained with hematoxylin, dehydrated andcover slipped.

ReverseTranscriptase-Polymerase Chain Reaction

To further characterize theexpressionofHes-1,mash-1, Ngn3 and Nkx2.2 in the mouse neuroendocrinetumors (NE-10), we performed RT-PCR analysis on10weeks oldnormalCD1DLP, 10weeks old 12T-7fDLPand prostate neuroendocrine tumors were collected.Total RNAwas isolated from themouse tissues using anRNeasy mini kit (Qiagen, Inc., Valencia) with residualgenomic DNA was removed by RNase-Free DNase(Qiagen) treatment. One microgram of total RNA wasreverse transcribedusingSuperscript-IITMreverse trans-criptase (Invitrogen) according to the manufacturer’sinstructions. PCR was performed using sense andantisense primers to produce gene specific fragments.The primer sequences for each of the genes were:

Hes-1: sense 50 TGGAAATGACTGTGAAGCACCTCC30 and antisense 50 ATGATCTGGGTCATGCAACT-GGC 30

mash-1: sense 50 ACGACTTGAACTCTATGGCG-GG-TT 30 and antisense 50 TGACGTCGTTGGCGAGA-AACACTA 30


52 Aparna et al.

Ngn3: sense 50 TTGAGTCGGGAGAACTAGGATGG30 and antisense 50 TTTGCTGAGTGCCAAC-TCGC-TCTT 30

Nkx2.2: sense 50 GGGGTTTTCAGTCAAGGACA 30

and antisense 50 CTTTGGAGAAGAGCACTCGG 30

The conditions for the PCR were: 948C for 5 min(1 cycle), 948C for 30 sec, 608C for 1 min, 728C for 1 min(35 cycles), and 728C for 10min (1 cycle). PCR productswere analyzed by electrophoresis in 1.5% agarose gelscontaining ethidium bromide and photographedunder UV illumination.

RESULTS

Sequential Expression of TranscriptionFactors inNeuroendocrine PIN

Immunohistochemical analysis was performed onthe NE PIN lesions from the 12T-10 mice prostate toestimate the expression of Foxa2, mash-1, Ngn3, andNkx2.2. All the PIN lesions analyzed were positive forsynaptophysin,marker forNE cells (Fig. 1C,D). TheNEPIN lesions were mainly negative for AR expression(occasionally individual faintly AR positive cells aredetected) as shown in Figure 1E,F. These PIN lesionsdid express T-antigen, the transgene used to generatethese prostate tumors (results not shown). The NE PINlesions were always positive for Foxa2 (Fig. 1A,B)as indicated by the arrow, mash-1 expression wasdetected in only about 50% of the samples analyzed(Fig. 1G,H). In the Figure 1 the area surrounding theregion pointed by the arrow shows a faint positivestaining for Foxa2 (Fig. 1B) and also a decrease in ARexpression in these cells (Fig. 1F). Although these cellsdid not express synaptophysin and mash-1 yet, thefaint positive staining for Foxa2 suggests that the cellsare beginning to differentiate into neuroendocrine cellswhere Foxa2 is an early marker. Expression of Ngn3 in

the normal prostate or during early stages of PINin the 12T-10 line could not be detected even by RT-PCR. These results suggest that there is a sequentialexpression of these transcription factors in the NEprostate tumors where Foxa2 and mash-1 are the earlymarkers that get expressed in the PIN lesions andNgn3andNkx2.2 are the late stage markers. Further RT-PCRanalysis was performed on the NE-10 tissue to confirmthe expression of Hes-1, mash-1, Ngn-3, and Nkx2.2 inthe poorly differentiated NE tumors. Figure 2A showsthat Hes-1 expression is lost in these NE tumors (lane3)compared to the prostate from a normal mice (lane 1)and the non-neuroendocrine 12T-7 prostate (lane 2).mash-1 is expressed only in theNE-10 prostate (Fig. 2B)(lane 3). The normal mouse prostate (lane 1) andthe 12T-7f prostate (lane 2) do not express mash-1.SimilarlyNgn3 expressionwas detected only in theNEtumor (Fig. 2C: lane 3). Very faint expression of Nkx2.2(Fig. 2D: lane3) was detected only in the NE10 sample,suggesting that limited number of cells express it or theexpression level is very low. Immunohistochemistry inFigure 2 shows an example of HGPIN that has somefeatures consistent with early stage prostate cancerwhere the ductal structure is still partially retained. TheHGPIN cells in this sample are positive for synapto-physin expression (Fig. 2F), marker for NE cells. Thesecells also express strong levels of Foxa2 (Fig. 2G) andmash-1 (Fig. 2H). AR expression is completely lost inthese cells (results not shown). Nkx2.2 expression is notdetected in any of these cells (Fig. 2I).

Dorsolateral Prostate of12T-10MiceExpresses Pro-EndocrineGenes

Immunohistochemical analysis was performed onthe poorly differentiated NE carcinoma of the dorso-lateral prostate from 56 weeks old 12T-10 mice. TheH&E staining (Fig. 3A) shows that the tumor is a poorly


Fig. 1. ImunohistochemistryshowingpositiveexpressionofFoxa2(A;20�)(B;40�),synaptophysin(C;20�)(D;40�), lossofARexpression(E; 20�) (F; 40�) andmash-1 (G; 20�) (H; 40�)expressionin thelowgradeNEPINlesionsindicatedby thearrows.Bars,50mm.


differentiated cancer with complete loss of ductalstructure. Advanced NE tumors continue to expresssynaptophysin (Fig. 3C), a marker of NE cells. Inaccordance with our previous report [14] the NEcells express Foxa2 (Fig. 3B). Figure 3D shows thatthe NE cells are positive for large T-antigen, thetransgene targeted to the prostate by the probasinpromoter. These NE tumors show loss of androgenreceptor (Fig. 3E). Panels F,G of Figure 3 show positiveimmunostaining of theNE tumor formash-1 andNgn3,respectively. Foxa2 interacts with Ngn3 or NeuroD1 torecruit transcriptional activators to Nkx2.2 promoter[15]. Thus, the expression of Nkx2.2 by NE tumor cellsis consistent with the expression of known activators ofthis gene (Fig. 3H). The homeodomain transcriptionfactor Nkx2.2 is required for the final differentiation ofthe b-cells in the pancreas and for the production ofinsulin. The expression of Nkx2.2 is not uniformthroughout the NE tumor suggesting that certain cellsmay lack factors required to activate the Nkx2.2promoter. The RT-PCR results for Nkx2.2 expression(Fig. 2D) in theNE tumor showsaveryweak expressionsuggesting that not all the cells express Nkx2.2, alsoreconfirming the immunohistochemistry results.

TheVentral Prostate of12T-10MiceDoesnotExpressNkx2.2

Panel A of Figure 4 shows H&E staining on theserial section of an advanced and poorly differentiatedNE tumor derived from the VP of 12T-10mice. Figure 4shows the expression of Foxa2 (B), Synaptophysin (C),Large T-antigen (D), androgen receptor (E), mash-1 (F),Ngn3 (G) andNkx2.2 (H) in the serial sections from theVP of 52 weeks old 12T-10 mice. This NE tumor waspositive for Foxa2, mash-1 and Ngn3 but expression of


Fig. 2. RT-PCRresults showingexpressionofHes-1 (A),mASH-1(B), ngn3 (C), Nkx2.2 (D) and b-actin control (E).Lane1: normalprostate, lane2:prostatefrom12T-7f, lane3:NE-10tumorandlane4: negative control. Immunohistochemistry of the prostateshows expression of Synaptophysin (F), Foxa2 (G), mash-1 (H) andNkx2.2 (1) inHGPIN lesions indicatingneuroendocrine differentia-tionin the12T-10micelineat50weeksof age.Bars,50mm.

Fig. 3. Immunohistochemicalanalysisofpoorlydifferentiateddorsolateralprostateof12T-10neuroendocrineprostate tumor showingH&Estaining (A), positive staining for Foxa2 (B), synaptophysin (C) and T-antigen (D), negative stainingof androgen receptor (E), positive nuclearstaining formash-1 (F),neurogenin-3 (G) andamixedpopulationofcellsbothpositiveandnegative forNkx2.2(H).Bars,50mm.

54 Aparna et al.

Nkx2.2 was not detected. Lung metastases from thesame mouse express Ngn3 but do not express Nkx2.2while the liver metastasis from this animal did notexpress either Ngn3 or Nkx2.2 (results not shown).This suggests that the primaryNE tumors could have amixed population of cells; some express Nkx2.2 andsomedonot as is the case in theDLP inFigure 3. It is alsopossible to have an advanced NE tumor with cells notexpressing detectable levels of Nkx2.2 as in Figure 4H.This observation suggests that either NE tumorsadvance to different stages in different mice and/orthe tumors are heterogeneous in nature with differentpatterns of Ngn3 and Nkx2.2 expression.

Neuroendocrine LiverMetastases ExpressNkx2.2

The 12T-10mice usually demonstrate extensive livermetastases with only micro metastases to the lung.Immunohistochemical analysis showed that the NE

cells that reside in the two metastatic sites expressdifferent genes. H&E staining (Fig. 5A) shows livermetastases with rosette histology, typical pathologyof small cell carcinoma/NE cancer. Figure 5 showspositive expression of Foxa2 (B), synaptophysin (C)and large T-antigen (D) in a liver metastasis from a69 weeks old 12T-10 mouse. These metastases loseandrogen receptor expression (E). The liver metastasesexpress mash-1 (F) but do not express Ngn3 (G). Theloss of Ngn3, a marker for differentiated endocrinecells, suggests that cancers have progressed to poorlydifferentiated NE tumors. However, the liver meta-stases can still expressNkx2.2 (H) even in the absence ofNgn3. Based on a 15 mice studied, Nkx2.2 expressionwas not seen in any of the NE lung metastasis but 33%of liver metastasis were positive for Nkx2.2 expression.Thus, even when mice had an Nkx2.2 positive livermetastasis, the lung metastases in the same animals


Fig. 4. Serial sections of ventral prostate of 52 weeks old 12T-10 neuroendocrine prostate showing H& E staining (A), positiveimmunostaining for Foxa2 (B), synaptophysin (C) and T-antigen (D), negative for androgen receptor (E), positive staining formash-1 (F) andneurogenin-3 (G) andnegative forNkx2.2 (H).Bars, 50mm.

Fig. 5. Serial sections of neuroendocrine liver metastasis from a 69 weeks old 12T-10 mice showing H&E (A), positive immunostainingfor Foxa2 (B), synaptophysin (C) and T-antigen (D), negative staining for androgenreceptor (E), positive staining formash-1 (F), negative forneurogenin-3 (G), andpositivenuclear staining forNkx2.2 (H).Bars, 50mm.

NeuroendocrineDifferentiation of ProstateCancer 55

were negative for Nkx2.2. This demonstrates that themetastases to different sites are different in terms of thegene expressionprofile. This could indicate thatNkx2.2is not required for liver metastasis since 66% of themetastasis were negative for this gene. However, cellsexpressing Nkx2.2 did prefer to grow in the liver andnever the lung. None of the liver NE metastasesexpressed Ngn3, suggesting that the 33% that doexpress Nkx2.2 must regulate the Nkx2.2 gene by adifferent pathway.

Neuroendocrine LungMetastasesHaveDifferent Gene Expression Prof|leComparedto LiverMetastases

Figure 6 shows serial sections of NE lung metastasesfrom a 69 weeks old 12T-10 mouse that also had livermetastasis (Fig. 5). TheNE lungmetastases express Foxa2(Fig. 6B), synaptophysin (Fig. 6C), Large T-antigen(Fig. 6D) and mash-1 (Fig. 6E). As in the primary NEtumors, lung NE metastases do not express androgenreceptor (Fig. 6F). Unlike liver metastases, lung meta-stases always express Ngn3 (Fig. 6G) but never expressNkx2.2 (Fig. 6H). This suggests that although Ngn3 isexpressed in the lung metastases, Ngn3 is unable toswitch on the expression of Nkx2.2, probably because ofthe lack of other transcription factors that also control theNkx2.2 promoter. Alternatively, Nkx2.2 expression maybe controlled by inhibitory signals from the surroundingmicroenvironment.

TheNeuroendocrineAllograftModel (NE-10)Reconf|rms the Results Seen in12T-10

NeuroendocrineMiceModel

The neuroendocrine prostate tumor from the 12T-10mice was grafted subcutaneously into athymic nude

mice. These grafts were harvested and further passag-ed to establish a transplantable allograft model termedNE-10 [9].These subcutaneous grafts retained thecharacteristic rosette pattern of the 12T-10 NE prostatetumors, express the neuroendocrine markers andmetastasize to the lung and liver of the nude mice.We checked the expression of the genes involved inpancreatic differentiation in the NE-10 subcutaneousgrafts as well as the lung and liver metastases. Thesubcutaneous graft and the liver metastases from P13(passage 13) NE-10 tumor were positive for Foxa2 asshown in Figure 7A,E respectively. Further, mash-1expressionwas seen in both subcutaneous (Fig. 7B) andliver metastases (Fig. 7F) samples. The subcutaneoustumor expressed Ngn3 (Fig. 7C) and Nkx2.2 (Fig. 7D)but the liver metastases from the same mice didnot express Ngn3 (G) but did express Nkx2.2 (H),consistent with what is seen in the 12T-10 transgenicmodel. The lung metastases, on the contrary, expressNgn3 and do not express Nkx2.2 (data not shown).Since the subcutaneousNE-10 allograft is heterogenousfor the expression of Nkx2.2 but only lung metastasesexpress Ngn3 and only liver metastases expressNkx2.2, this observation again suggests that differentpopulations of NE cancer cells adapt to the micro-environment of the metastatic site. The NE-10 subcuta-neous tumor, liver and lung metastases express LargeT-antigen and synaptophysin but they do not expressthe androgen receptor (data not shown). The consis-tency between the results seen in the 12T-10 andNE-10model suggests that prostate NED follows the similarmolecular pathway as that involved in pancreaticendocrine cell differentiation. To establish that thispathway is not unique to NED of mouse prostatictumors, the expression of hASH-1 (Fig. 8A) andsynaptophysin (Fig. 8B) was confirmed in human


Fig. 6. Immunohistochemicalanalysis onserial sectionsof lungmetastasis froma69weeksold12T-10mousewithH&Estaining (A), positivestaining forFoxa2 (B), synaptophysin (C),T-antigen (D), negative forAR (E), positivenuclear staining formash-1 (F) andneurogenin-3 (G) andnegative forNkx2.2 (H).Bars, 50mm.

56 Aparna et al.

neuroendocrine prostatic tumors. We have reportedthat the human NE prostate tumors express Foxa2 andgenerally show loss or reduction of androgen receptor[14]. These results suggest that the human NED ofprostate cancer follow the same pathway as the NEmouse models of prostate cancer

DISCUSSION

By definition, neuroendocrine refers to the commu-nication between thenervous and the endocrine systemresulting in the release of secretions. Others and ourgroup have documented that prostatic NE secretionshave growth promoting effects on prostatic cancer cells[7,16]. The embryonic origin of neuroendocrine cells inthe prostate (ectodermal or endodermal) remainsunclear. If, during prostatic development, the NEcells arise from the neural crest, they should haveectodermal features. However, if NE cells are endo-

dermal in origin, then they may share a commonprogenitor cell with the basal and luminal epithelialcells of the prostate. Although the origin of the NEcells has not been a pivotal question, it nevertheless isfundamental to understanding the plasticity of pro-static cancer during tumor progression. In rare cases,human prostate tumors are NE cancers [17,18] andfrequently human prostatic adenocarcinomas undergoneuroendocrine differentiation, both conditions asso-ciated with poor prognosis [19,20]. Numerous studieshave shown that the paracrine effects of NED can causethe adjacent androgen dependent cells to transforminto androgen insensitive cells.Most of the studies havediscussed the effects of NED on prostate cancer butthere has been very littlework done on themechanismsthat can lead to NE differentiation. In this study, wehave shown that prostate neuroendocrine differentia-tion follows the same molecular mechanism involvedin endocrine differentiation of pancreatic b-cells, cells


Fig. 7. Serial sections of subcutaneous neuroendocrine tumors from the xenograft model of 12T-10 mice (NE-10) showing positiveimmunostaining for synaptophysin (A), mash-1 (B), neurogenin-3 (C), and Nkx2.2 (D). Immunohistochemical analysis on the livermetastasis from the same athymic mice showing positive expression for synaptophysin (E), mash-1 (F), negative for neurogenein-3 (G),and positive for Nkx2.2 (H) in the left end of the picture. The right end of the pictures on panels (E^H) shows the host liver. Both thesubcutaneous tumorandthelivermetastasis arepositive forFoxa2andT-antigenandnegative forAR(resultsnot shown).Bars,50mm.

Fig. 8. Humanneedlebiopsysamplespositive for synaptophysin (A) andhash-1 (B).Bars, 50mm.

NeuroendocrineDifferentiation of ProstateCancer 57

that are endodermal in orgin. Further, we report thatnormal prostatic NE cells as well as NE prostaticcancers express Foxa1 and Foxa2 [14,21], two forkheadproteins frequently associated with endodermallyderived tissues [22].

Figure 9A shows the sequence of transcriptionfactors that are expressed in pancreatic endocrineprecursor cells. First, endodermal transcription factors,HNF6, Foxa2, and HNF1 are expressed. These dif-ferentiating endocrine cells express Notch ligands(delta and serrate) on the surface and activate Notchsignaling in adjacent cells, this prevents neighboring

cells from differentiating into the same cell type. Notchsignaling mediates its response through activationof Hes genes [23]. Hes genes act as transcriptionalrepressors and Hes1 inhibits the expression of pro-endocrine gene, Ngn3 [10,11] (Fig. 9A). Previousstudies have shown that activation of the Notch-1receptor leads to activation of a Hes-1 promoterconstruct as well as activation of the endogenousHes-1 gene [24]. Further recent work using antisenseoligos against Notch-1 in cell culture and overexpres-sion of Notch-1 in transgenic mice indicates thatNotch-1 regulatesHes-1 expression [25].Hes1 is knownto bind to the N box-related sequence of the mash-1promoter and repress mash-1 transcription [26].Animals lacking Ngn3 fail to develop endocrine cells[27] and ectopic expression of Ngn3 causes prematuredifferentiation of pancreas into endocrine cells [10,28].Once Ngn3 is activated in the progenitor cell, that cellis destined to become an endocrine cell. The Ngn3promoter contains inhibitory Hes1 binding sites prox-imal to the TATA box [29]. In addition to Hes-1, theNgn3promoter contains binding sites forHNF1, Foxa2,and HNF6 [29,30]. Figure 9A shows that Nkx2.2 is adownstream target of Foxa2 and Ngn3 which interactsynergistically to recruit transcriptional activators toNkx2.2 gene promoter [15]. Nkx2.2 expression in thepancreas marks the final development of the matureinsulin secreting b-cells.

Figure 9Bdemonstrates that the transcription factorsinvolved in pancreatic endocrine differentiation aresequentially expressed in prostate NE cancer cells.Foxa2, an endodermal transcription factor is express-ed early in the PIN lesions. In the HGPIN lesionsFoxa2 and synaptophysin are all expressed, but onlyoccasional HGPIN samples express mash-1. However,once NE cancer appears, the cells routinely expressmash-1. Similarly, hASH-1 has been reported tobe highly expressed in human medullary thyroidcancer and small cell lung cancer, neuroendocrinecancer [31–33].

As these HGPIN lesions grow to an advanced andundifferentiated NE cancer, there is loss of Hes-1expression. Thus, Hes-1mediated inhibition of Ngn3 isremoved and these NE cells express the pro-endocrinegene, Ngn3. As during pancreatic development [15],Ngn3 and Foxa2 can recruit other transcriptionfactors to the Nkx2.2 promoter permitting expressionof Nkx2.2 in some of the advanced NE tumors.Figure 9B shows that the NE lung metastases expressNgn3 and Foxa2 but never express Nkx2.2 while livermetastases express Foxa2 and Nkx2.2 but they donot express Ngn3. Therefore, expression of Nkx2.2 inNE tumors is not absolutely dependent upon theexpression of both Ngn3 and Foxa2. Further, thedifference in Ngn3 and Nkx2.2 expression between


Fig. 9. A: Diagrammatic representation of gene expression inpancreatic endocrine cell differentiation. Endodermal factors arethe initial ones to be expressed. Notch signaling prevents cellsfrom differentiating into endocrine cells and maintains normalcell differentiation. Notch signaling mediates its response throughHes-1 (hairy/enhancer of split) that represses pro-endocrine gene,neurogenin3 (Ngn3).Hes-1also downregulatesmASH-1.Ngn3 andFoxa2 interact synergistically to recruit transcriptional activatorsto Nkx2.2 promoter. Nkx2.2 is expressed in pancreatic b-cells.B: Diagrammatic representation of sequential gene expressionwith progression of neuroendocrine prostate cancer from LGPIN(low-gradeprostaticintraepithelialneoplasia)toHGPIN(highgradePIN) to poorly differentiated NE cancer followed by lung and livermetastases.

58 Aparna et al.

differentmetastatic sites indicates that a uniquepatternof gene expression exists in NE cells that either allowstheir growth at a given site and/or is controlled by thetissue microenvironment at the site of metastases.Foxa2 can bind to the distal and proximal Ngn3promoter and activate gene expression, Hes-1 on thecontrary specifically inhibits the Ngn3 promoter [29].This observation can be explained through the Foxa2positive early PIN lesions in 12T-10 prostates, whichstill express Hes-1 that prevents the expression ofNgn3. In the 12T-10 transgenic and NE-10 allograftmodel ofNEprostatic cancer,we can confirm the loss ofNotch signaling by the loss ofHes-1 shown byRT-PCR.Similarly, the loss of Notch signaling has been reportedin lung NE cancers [34,35], neuroblastomas [36], achildhood tumor originating from cells of developingsympathetic nervous system as well as in the neuro-endocrine phenotype in gastrointestinal carcinoids[37]. These results suggest that the genes involved inpancreatic endocrine differentiation are expressed inthe same sequential manner during NED of prostaticcells. This study presents markers for molecularmechanism that can either be used to detect early stageNE tumors or to target therapy to a subset of NEprostate tumors.

ACKNOWLEDGMENTS

The authors thankDr.MichaelGerman (Departmentof Medicine, Diabetes Center, University of California,San Francisco) for providing the neurogenin-3 anti-body. This research is supported byNational Institutesof Health grants to RJM (R01-CA76142 and R01-AG023409) and Frances Williams Preston Laboratoriesof the T.J. Martell Foundation.

Liao et al. reported that pten -/- prostates haveincreased NED after castration [38].

REFERENCES

1. Abrahamsson PA, Falkmer S, Falt K, Grimelius L. The courseof neuroendocrine differentiation in prostatic carcinomas. Animmunohistochemical study testing chromogranin A as an‘‘endocrine marker’’. Pathol Res Pract 1989;185(3):373–380.

2. Bostwick DG, Dousa MK, Crawford BG, Wollan PC. Neuro-endocrine differentiation in prostatic intraepithelial neoplasiaand adenocarcinoma. Am J Surg Pathol 1994;18(12):1240–1246.

3. AbrahamssonPA.Neuroendocrine cells in tumourgrowthof theprostate. Endocr Relat Cancer 1999;6(4):503–519.

4. di Sant’Agnese PA. Neuroendocrine differentiation in prostaticcarcinoma: An update on recent developments. Ann Oncol2001;12 (Suppl 2):S135–S140.

5. Weinstein MH, Partin AW, Veltri RW, Epstein JI. Neuro-endocrine differentiation in prostate cancer: Enhanced predic-tion of progression after radical prostatectomy. Hum Pathol1996;27(7):683–687.

6. Roudier MP, True LD, Higano CS, Vesselle H, Ellis W,Lange P, Vessella RL. Phenotypic heterogeneity of end-stageprostate carcinoma metastatic to bone. Hum Pathol 2003;34(7):646.

7. Jin RJ,Wang Y,Masumori N, Ishii K, Tsukamoto T, Shappell SB,Hayward SW, Kasper S, Matusik RJ. NE-10 neuroendocrinecancer promotes the LNCaP xenograft growth in castratedmice.Cancer Res 2004;64(15):5489–5495.

8. MasumoriN, ThomasTZ,ChaurandP,CaseT, PaulM,Kasper S,CaprioliRM,TsukamotoT, Shappell SB,MatusikRJ.Aprobasin-large T antigen transgenic mouse line develops prostateadenocarcinomaandneuroendocrine carcinomawithmetastaticpotential. Cancer Res 2001;61(5):2239–2249.

9. MasumoriN,TsuchiyaK,TuWH,LeeC,KasperS, TsukamotoT,Shappell SB, Matusik RJ. An allograft model of androgenindependent prostatic neuroendocrine carcinoma derivedfrom a large probasin promoter-T antigen transgenic mouseline. J Urol 2004;171(1):439–442.

10. Apelqvist A, Li H, Sommer L, Beatus P, Anderson DJ, Honjo T,Hrabe de Angelis M, Lendahl U, Edlund H. Notch signallingcontrols pancreatic cell differentiation. Nature 1999;400(6747):877.

11. Jensen J, Pedersen EE, Galante P, Hald J, Heller RS, Ishibashi M,Kageyama R, Guillemot F, Serup P, Madsen OD. Control ofendodermal endocrine development by Hes-1. Nat Genet 2000;24(1):36–44.

12. Yan Y, Sheppard PC, Kasper S, Lin L, Hoare S, Kapoor A,Dodd JG, Duckworth ML, Matusik RJ. Large fragment ofthe probasin promoter targets high levels of transgene ex-pression to the prostate of transgenic mice. Prostate 1997;32(2):129–139.

13. Kasper S, Yan Y, PettigrewN, BorowskyAD, Prins GS, Dodd JG,Duckworth ML, Matusik RJ. Development, progression, andandrogen-dependence of prostate tumors in probasin-large Tantigen transgenic mice: Amodel for prostate cancer. Lab Invest1998;78(3):319–333.

14. Mirosevich J, Gao N, Gupta A, Shappell SB, Jove R, Matusik RJ.Expression and role of Foxa proteins in prostate cancer. Prostate2006;66(10):1013–1028.

15. Watada H, Scheel DW, Leung J, German MS. Distinct geneexpression programs function in progenitor and mature isletcells. J Biol Chem 2003;278(19):17130–17140.

16. UchidaK,MasumoriN,TakahashiA, ItohN,KatoK,MatusikRJ,Tsukamoto T. Murine androgen-independent neuroendocrinecarcinoma promotes metastasis of human prostate cancer cellline LNCaP. Prostate 2006;66(5):536–545.

17. Ketata S, Ketata H, Fakhfakh H, Sahnoun A, Bahloul A,Boudawara T, Mhiri MN. Pure primary neuroendocrine tumorof the prostate: A rare entity. Clin Genitourin Cancer 2006;5(1):82–84.

18. Okada H, Gotoh A, Ogawa T, Arakawa S, Ohbayashi C,Kamidono S. Two cases of small cell carcinoma of the prostate.Scand J Urol Nephrol 1996;30(6):503–508.

19. Yashi M, Terauchi F, Nukui A, Ochi M, Yuzawa M, Hara Y,Morita T. Small-cell neuroendocrine carcinoma as a variantform of prostate cancer recurrence: A case report and shortliterature review. Urol Oncol Semin Orig Investig 2006;24(4):313.

20. Turbat-Herrera EA, Herrera GA, Gore I, Lott RL, Grizzle WE,Bonnin JM. Neuroendocrine differentiation in prostatic carcino-mas. A retrospective autopsy study. Arch Pathol Lab Med 1998;112(11):1100–1105.



21. Mirosevich J, Gao N, Matusik RJ. Expression of Foxa tran-scription factors in the developing and adult murine prostate.Prostate 2005;62(4):339–352.

22. Ang SL, Wierda A, Wong D, Stevens KA, Cascio S, Rossant J,Zaret KS. The formation and maintenance of the definitiveendoderm lineage in the mouse: Involvement of HNF3/fork-head proteins. Development 1993;119(4):1301–1315.

23. Jan YN, Jan LY. HLH proteins, fly neurogenesis, and vertebratemyogenesis. Cell 1993;75(5):827–830.

24. Jarriault S, Le Bail O, Hirsinger E, Pourquie O, Logeat F, StrongCF, Brou C, Seidah NG, Israel A. Delta-1 activation of notch-1signaling results in hes-1 transactivation. Mol Cell Biol 1998;18(12):7423–7431.

25. Shan L, Aster JC, Sklar J, Sunday ME. NOTCH-1 regulatespulmonary neuroendocrine cell differentiation in cell lines andin transgenic mice. Am J Physiol Lung Cell Mol Physiol 2007;292(2):L500–L509.

26. Chen H, Thiagalingam A, Chopra H, Borges MW, Feder JN,Nelkin BD, Baylin SB, Ball DW. Conservation of the Drosophilalateral inhibitionpathway inhuman lung cancer:Ahairy-relatedprotein (HES-1) directly represses achaete-scute-homolog-1expression. Proc Natl Acad Sci 1997;94(10):5355–5360.

27. GradwohlG,DierichA, LeMeurM,Guillemot F.Neurogenin3 isrequired for the development of the four endocrine cell lineagesof the pancreas. Proc Natl Acad Sci 2000;97(4):1607–1611.

28. Schwitzgebel VM, Scheel DW, Conners JR, Kalamaras J, Lee JE,Anderson DJ, Sussel L, Johnson JD, German MS. Expression ofneurogenin3 reveals an islet cell precursor population in thepancreas. Development 2000;127(16):3533–3542.

29. Lee JC, SmithSB,WatadaH,Lin J, ScheelD,Wang J,MirmiraRG,German MS. Regulation of the pancreatic pro-endocrine geneNeurogenin3. Diabetes 2001;50(5):928–936.

30. Jacquemin P, Durviaux SM, Jensen J, Godfraind C, Gradwohl G,Guillemot F, Madsen OD, Carmeliet P, Dewerchin M, Collen D,Rousseau GG, Lemaigre FP. Transcription factor hepatocyte

nuclear factor 6 regulates pancreatic endocrine cell differ-entiation and controls expression of the proendocrine genengn3. Mol Cell Biol 2000;20(12):4445–4454.

31. Ball DW, Azzoli CG, Baylin SB, Chi D, Dou S, Donis-KellerH, Cumaraswamy A, Borges M, Nelkin BD. Identificationof a human achaete-scute homolog highly expressed inneuroendocrine tumors. Proc Natl Acad Sci 1993;90(12):5648–5652.

32. Ito T,UdakaN,Okudela K, YazawaT, KitamuraH.Mechanismsof neuroendocrine differentiation in pulmonary neuroendocrinecells and small cell carcinoma. Endocr Pathol 2003;14(2):133–139.

33. Jiang S-X,KameyaT,AsamuraH,UmezawaA, SatoY, Shinada J,KawakuboY, Igarashi T,Nagai K, Okayasu I. hASH1 expressionis closely correlated with endocrine phenotype and differ-entiation extent in pulmonary neuroendocrine tumors. ModPathol 2003;17(2):222.

34. Ball DW. Achaete-scute homolog-1 and Notch in lung neuro-endocrine development and cancer. Cancer Lett 2004;204(2):159–169.

35. Collins BJ, Kleeberger W, Ball DW. Notch in lung developmentand lung cancer. Semin Cancer Biol 2004;14(5):357–364.

36. AxelsonH. TheNotch signaling cascade in neuroblastoma: Roleof the basic helix-loop-helix proteins HASH-1 and HES-1.Cancer Lett 2004;204(2):171.

37. Nakakura EK, Sriuranpong VR, Kunnimalaiyaan M, Hsiao EC,Schuebel KE, Borges MW, Jin N, Collins BJ, Nelkin BD, Chen H,Ball DW. Regulation of neuroendocrine differentiation ingastrointestinal carcinoid tumor cells by notch signaling. J ClinEndocrinol Metab 2005;90(7):4350–4356.

38. Liao CP, Zhong C, Saribekyan G, Bading J, Park R, Conti PS,Moats R, Berns A, Shi W, Zhou Z, Nikitin AY,Roy-Burman P.Mouse models of prostate adenocarcinoma with the capacity tomonitor spontaneous carcinogenesis by bioluminescence orfluorescence. Cancer Res 2007;67(15):7525–7533.


60 Aparna et al.

neuroendocrine differentiation in the 12t-10 transgenic prostate mouse model mimics endocrine...

Documents