The Prostate 62:282^289 (2005)

ConstitutiveActivationof gp130 Leads toNeuroendocrineDifferentiation InVitro and InVivo

Jodie Palmer,1,2* Matthias Ernst,3 Annet Hammacher,4 and Paul J. Hertzog2

1Prostate Centreat VancouverGeneralHospital,Vancouver,BC,Canada2Monash Institute of ReproductionandDevelopment,MonashUniversity,

MonashMedical Centre,Melbourne, Australia3Ludwig Institute of Cancer Research,Melbourne, Australia4BernardO’Brien Institute ofMicrosurgery, Fitzroy, Australia

BACKGROUND. Neuroendocrine (NE) differentiation in prostate tumors has been correlatedwith androgen independent disease and increased risk of death. In vitro, IL-6 initiates NEdifferentiation utilizing the signal transduction initiated by the interactionwith IL-6R alpha andgp130. In this study we analysed the NE differentiation process in vitro and in vivo using theLNCaP androgen dependent cell line via ligand independent induction of NE differentiation.METHODS. LNCaP cells were transfected with a constitutively active gp130 subunit,gp130act. Cell proliferation ratewas determined and cloneswere examined for neuroendocrinedifferentiation by morphological change, upregulation of CgA and serotonin and formation ofdense core vesicleswithLNCaPparental cells as the control. Xenograft formationwas examinedand compared in immunocompromised mice.RESULTS. Gp130act expression promoted significant neuroendocrine differentiation in vitroas determined by a NE like morphology change (increased neurite like extension formation),elevated CgA expression and the formation of dense core vesicles (DCV). These measuresconcurred with those examined in LNCaP cells following 100 ng/ml IL-6 treatment. Furtherinvestigation of the LNCaP gp130act cells in vivo, in immunocompromised androgen intactmice, confirmed that theNE likemorphology, as determinedbyhistological andhigh resolutiontransmission electron microscopy, was maintained.CONCLUSIONS. NE differentiation was initiated by the expression of gp130act in a ligandindependent manner, highlighting the importance of gp130 in the neuroendocrine differentia-tion process. Further investigation of upregulated/downregulated gene expression in thesecells may provide valuable insight into the NE differentiation process. Prostate 62: 282–289,2005. # 2004 Wiley-Liss, Inc.

KEY WORDS: prostate cancer; neuroendocrine (NE); gp130; constitutive activation

INTRODUCTION

Prostate cancer (PCa) is the second leading cause ofcancer related death in men of the Western world [1].PCa tumors are heterogeneous and evidence points to arole of neuroendocrine (NE) differentiation in diseaseprogression [2]. NE differentiation in PCa is correlatedwith aggressive disease [3] and the associated morbid-ity in PCa. While NE cells are present in the prostatefrom birth as a minor epithelial population, thereported increase in NE cells to between 10 and 100%of tumor cells [4,5] points to NE differentiation of

Abbreviations: gp130, glycoprotein 130; IL-6, interleukin 6; NE,neuroendocrine; TEM, transmission electron microscopy; DCV,dense core vesicle.

*Correspondence to: Dr. Jodie Palmer, Prostate Centre at VGH, 2660Oak St., Vancouver, BC V6H 3Z6, Canada.E-mail: jpalmer@vanhosp.bc.caReceived 5 January 2004; Accepted 3 June 2004DOI 10.1002/pros.20143Published online 23 July 2004 in Wiley InterScience(www.interscience.wiley.com).

� 2004 Wiley-Liss, Inc.

other resident cell types. The induction of NE differ-entiation in vivo is not well characterized [6]. NEdifferentiated cells are proposed to release growthand/or differentiation factors that may influenceneighboring cells [7–9].

In vitro, the androgen dependent prostate cell line,LNCaP, are induced toNEdifferentiate upon treatmentwith a number of different factors including IL-6 [8,9],IL- 1 [10], and epinephrine [11] as well as androgendepleted conditions [12]. This differentiation is char-acterized by an upregulation of markers including:CgA and serotonin (5HT), which may be stored indense core vesicles (DCV) within the cell body [13],flattening of the cell body in vitro, and a lengthening ofdendritic processes from the cell [14]. These processesmay facilitate interaction with neighboring cells andenable the release of factors storedwithinDCVs locally.Further, NE differentiation is characterized by prolif-eration arrest but not apoptosis.

IL-6 and the gp130 homodimer play an importantrole in the proliferation and NE differentiation ofLNCaP cells. Further, confusion exists in the literatureregarding the proliferation of LNCaP cells in responseto cytokine treatment. The proliferation of LNCaP cellswas demonstrated to be both promoted [15] and in-hibited [16,17] following cytokine treatment. Due to thealternate signaling pathways proposed for IL-6[9,15,18,19], including the MAPK, PI3K, and c-Srcsignaling cascades, a ligand independent approachwas undertaken to determine the importance of gp130in the process. To examine the role of gp130 in NEdifferentiation of LNCaP cells, separate to the liganddependent activation initiated by IL-6, a ligand inde-pendent approachwas examinedvia the transfection ofa constitutively active gp130 subunit, gp130act. Thegp130act construct consisted of a truncated extracel-lular domain with the shared human/murine gp130signal sequence, the Dind b common domain [20,21],to confer factor independent activation, and the intra-cellular sequence of murine gp130 which is over-expressed under the regulation of the elongation factor1a promoter. This is represented in a schematic com-parison of endogenous gp130 (Fig. 1) and the truncatedgp130act.

Given the correlation between aggressive tumorincidence and NE differentiation, in vivo analysis ofLNCaP gp130act cell induced tumor formation wascompared with parental LNCaP cells to examinewhether gp130act expression altered tumorigenic capa-city. It was determined that while the incidence andtumor proliferation rate was significantly decreased bygp130act expression in LNCaP cells, the incidence ofsignificantly altered tumor morphology should pro-vide valuable proof as to the importance of furtherinvestigation into inhibition of NE differentiation.

MATERIALSANDMETHODS

CytokineTreatment

Cells cultured routinely in DMEM (Invitrogen,Burlington, ON, Canada), 5% FBS (PA Biologicals,Sydney, NSW, Australia). Cells were treated with IL-6(a kind gift from Dr. R.J. Simpson, Ludwig Institute forCancer Research,Melbourne, Australia [22]) at 100 ng/ml in routine culture media, with diluent used as thecontrol.

Electroporation of gp130act

pEF1a-MC1 Neo-gp130act was linearised for stabletransfection by restriction enzyme digestion with Sal Iand cellswere electroporatedwith 15 mg linear plasmidusing a Gene Pulser Unit (Biorad Corp., Hercules, CA).Positive clones were obtained after G418 selection(Invitrogen).

Analysis of Genomic Integration of gp130act

Genomic integration of the plasmidwas determinedby PCR utilizing one primer in the EF1a promoter andone primer in the gp130act cDNA section of the plas-mid: EF1a sense: 50 atg cga tgg ttc ccc aca 30 Dindantisense: 50 acg act gct tct ata cat ggt 30 and gp130actantisense: 50 ggt ctc cat tcc aac aga ttc 30.

Fig. 1. InductionofNEdifferentiationbyoverexpressionof a con-stitutively active ligand independentgp130 (gp130act) receptor:Thehumangp130receptorrequiresdimerizationandbinding to an IL-6 -type cytokine receptor ligand complex. The gp130act receptor isproposed to be constitutively active and does not require ligandbinding to initiate signal transduction.Due to this activity, gp130actis hypothesized to be able to induce NE differentiation in prostatecells.

ConstitutiveActivation of gp130 Leads toNEDifferentiation 283

Analysis of Protein Expression of gp130act

Preparation of cell protein lysates and immunopre-cipitation (ip) of gp130 proteins were performed usinganti-gp130 antibody (Santa Cruz Biotech., Santa Cruz,CA). Expression was confirmed using the same anti-body, with gp130act phosphorylation determined bygp130 ip and anti-phosphotyrosine antibody (4G10,Upstate Biotech.) Western blot analysis.

Morphometric andDenseCoreVesicle Analysis

For morphometric analysis, LNCaP gp130act cellswere examined in routine growthmedia and comparedto parental controls. Unless otherwise stated, cellswerenot fixed and were photographed in the culture well.Cell size was compared by measuring the longesttotal cell length from captured images using AdobePhotoshop and compared statistically using the stu-dent’s t-test. DCVs were examined using transmissionelectron microscopy by methods described previously[23].

Analysis of Proliferation of gp130act Cells

LNCaP gp130act and LNCaP parental cells wereseeded at 1� 105 cells per well and cultured in normalgrowth medium for 5 days. Total number of cells werecompared by cell count using a haemocytometer andtrypan-blue exclusion.

Analysis of CgAExpression

Cellswere treatedwith IL-6 for 5days and comparedto untreated control cells following fixation, incubationwith anti-CgA antibodies (DAKO) followed by anti-rabbit FITC antibody (DAKO) to detect binding. Cellswere imaged using the Leica DMR fluorescencemicroscope.

Analysis of theGrowthof TumorXenografts inMice

Nu/numutantmice were utilized for the analysis oftumor formation in vivo. LNCaP or LNCaP gp130actcells (1� 106 total) were injected with 50 ml nutri-ent enriched MatrigelTM (Becton Dickinson) sub-cutaneously into the left side of the pelvic region of5 mice per cell line (LNCaP gp130act clone A, LNCaPgp130act clone 408.3, or LNCaP parental cells). Whiletumors were not detected until at least 2 weekspost injection, injection sites were examined, starting1week after injection. Mice were checked every 2 days,with tumors measured using calipers as describedpreviously [24]. Experiments were repeated in SCIDmice to ensure the strain of mice did not alter tumorgrowth and to ensure reproducibility.

TUMORPREPARATION

Histology

The tumor was fixed in fresh 4% paraformaldehyde(pH 7.4), dehydrated and embedded in paraffin wax.Sections (5 mm) were counterstained with Harris’shaematoxylin. Structures were determined as describ-ed previously [25] by light/phase microscopy.

TEMTumorAnalysis

After removal from the animal, the tumor was cutinto approx 1 mm� 1 mm blocks. Tumors were thenfixed overnight at 48C in 3% glutaraldehyde (BDH)diluted in 0.2M sodium cacodylate and then pre-pared and sectioned as described previously [23].Sections were analyzed at 60–80 kV with magnifica-tionsof 2,500–7,500�onaSeimans Joel 1200EXelectronmicroscope.

Statistical Analysis

All statistical analysis was performed using SPSSstatistical analysis software (SPSS, Inc., Chicago, IL).

RESULTS

Expression of gp130actDecreasesthe Proliferation of LNCaPCells

The activity of the gp130act construct was testedprior to transfection in LNCaP cells by M.E.: the IL-6dependent 7TD1 mouse hybridoma cell line wastransfected with the construct (Fig. 2) which led to asignificant proliferation promotion (data not shown).LNCaP cells were electroporated with linear pEF1a-MC1Neo gp130act DNA, subjected to antibioticselection (200 mg/ml G418) and clones with stable ex-pression were confirmed by genomic DNA PCR (datanot shown) and Western blot of total cell lysate. InLNCaP cells, the gp130act protein was expressed as a46 kD protein (Fig. 3A) and appeared to be tyrosinephosphorylated (Fig. 3B) as determined by gp130 ipand Western blot with a phospho-tyrosine antibody.This suggested that the intracellular domain of the

Fig. 2. The gp130act construct Gp130act consists of a truncatedextracellular domain, which is homologous in both the human andmouse and contains the gp130 signal sequence, theDind b commondomain [17] and the intracellular sequence of murine gp130. Thegp130act gene utilizes the elongation factor1a (EF1a) promoter inthis system.

284 Palmeret al.

gp130act protein was being phosphorylated by endo-genous tyrosine kinases, such as the janus kinases,which is important in the gp130 signaling cascade. Theproliferation rate of the LNCaP gp130act cells wassignificantly altered by the introduction of gp130act(Fig. 3C) in untreated cells when compared withparental LNCaP cells (tested over >10 cell passages,pEF-Bos transfected clones did not demonstrate alteredproliferation and as such LNCaP parental cells were

used for further comparisons). This decreased prolif-eration rate resembles that seen with high concentra-tions of IL-6 [26].

gp130act Induces aNE-LikeMorphologyin LNCaPCells

LNCaP gp130act cells were examined. IL-6 is pub-lished to induce a significant morphology change inLNCaP cells [8,9,26]. A similar morphology was ap-parent after the introduction of gp130act into LNCaPcells (Fig. 4A) with increased length of dendritic likeprocesses, flattening of the cell body and decrease ofthe cytosolic volume. Data is represented graphically(Fig. 4B) as a percentage of control cell size, with

Fig. 3. gp130act expression in LNCaP cells alters proliferation:LNCaP gp130act cells were examined following transfection.A: LNCaP gp130act cells were lysed and protein was extracted.gp130act and endogenous gp130 protein was immunoprecipitatedwith anti-gp130 antibodies. Immunoprecipitates were separatedby 10% SDS^PAGE, with the gp130act protein detected at 46 kD(endogenous protein was detected at 130 kD). B: Immuno-precipitation (ip) with anti-gp130 antibodies andWesternblotwithanti-phosphotyrosine antibody suggests that gp130act is tyrosinephosphorylated.C:Graphicalrepresentationofproliferationdiffer-ence apparent in LNCaP cells expressing gp130act compared withLNCaPparentalcells:Totalmeancellnumberwas comparedafter a5dayproliferationassay(cellswereinitiallyseededat1�105cellsper/well, n¼ 9). The proliferation of LNCaP gp130act cells was signi-ficantly lower thanparental controls as determinedby the compar-isonofmeancellnumberusingStudent’st-test, **,P< 0.05.

Fig. 4. LNCaP cells areNE differentiatedbygp130act expression(A) LNCaP cells expressing gp130act are compared with parentalcontrolcellsandaremorphologicallydifferentwithlong thindendri-tic like processes as representedby the arrow.B:Mean cell lengthsarerepresentedgraphically,where cell lengthwasmeasuredby theend of the longest process to the corresponding process on theothersideof thecell.Statisticalsignificancewasdeterminedbycom-parisonofmeansusingStudent’st-test, *,P< 0.05

ConstitutiveActivation of gp130 Leads toNEDifferentiation 285

statistical significance determined using raw cell sizedata. To control for experimental error, measurementswere always compared between cells seeded from thesame original population and from the same treatmentperiod.

DenseCoreVesicle Formation is Inducedby gp130act Expression

Expression of gp130act led to DCV formation inLNCaP cells similar to that seen with IL-6 treatment.LNCaP gp130act cells were examined þ/� IL-6 treat-ment by TEM, with little difference apparent pre andpost IL-6 treatment (Fig. 5A,B). In contrast, LNCaPparental cellsþ/� IL-6 had substantial DCV formationfollowing IL-6 treatment, above that seen in control cellpopulations (Fig. 5C,D).

Thus, IL-6 treatment or gp130act expression potenti-ates the formation of DCVs (measured as dense bodiesless than 200 nm). As DCVs are proposed to storemarkers of NE differentiation, these markers wereanalyzed þ/� cytokines. The results of this analysisare described below.

gp130act ExpressionAlters the Basal Expression ofCgAandis Elevatedby IL-6 Treatment

LNCaP gp130act cells were compared with parentalcontrols for theNEmarker, CgA. Basal CgA expressionwas found to be elevated in LNCaP gp130act cells.Further, treatment with IL-6 led to a significantelevation in CgA expression in LNCaP gp130act cellssimilar to that seen in parental LNCaP cells (Fig. 6).These data confirm that gp130act expression leads toNE-like differentiation in LNCaP cells.

Fig. 5. Dense core vesicle (DCV) formation following is inducedby IL-6 andgp130act expression: Electronmicrographs ofþ/� IL-6treated LNCaP and LNCaPgp130act cells.LNCaPgp130act expres-singcells [þ/� IL-6: (A)/(B), respectively]haveequableDCVforma-tion and both have condensed nucleoli. LNCaP cells treated with100ng/mlIL-6(D)havegreaternumbersofDCVthanuntreatedcells(C) with a more condensed nucleolus in treated cells. DCV, densecorevesicles;Nu,nucleolus;N,nucleus;C, cytoplasm.

Fig. 6. LNCaP gp130act elevates CgA expression: Expression ofCgAwasdetectedinuntreatedLNCaPgp130actcells andincreaseduponIL-6treatment,whileLNCaPparentalcellshadlittledetectibleexpressionuntil treatmentwith IL-6 occurred.A: LNCaPgp130actcells were treated with100 ng/ml IL-6 and compared to untreatedcellsbyimmunohistochemicalanalysis ofCgAexpression.Photomi-crographs of cells in phase (left) andwith CgA expression is repre-sentedby Fitc binding (right).B: LNCaPparental control cellsweretreated as above. Phase contrast (left) represents morphology ofcells,CgAexpressionisrepresentedbyFitcbinding (right).

286 Palmeret al.

InVivo Proliferation of LNCaPgp130actCells is Altered

The in vivo proliferation of LNCaP gp130act cellswas examined by subcutaneous pelvic injection intoimmunocompromised (androgen intact) male mice.The incidence of tumor formation was significantlydecreased in LNCaP gp130act xenografts when

Fig. 7. LNCaPgp130act xenografts have a lower tumor incidenceand volume thanparental controls.A: Graphical representation oftumor incidence of LNCaP gp130act cell line xenografts (clones 2An¼10 and 408.3 n¼10) and LNCaP parental control (n¼10) xeno-graftsup today60postsubcutaneous(sc) injection(datarepresentsmean tumor incidence from two independent experiments). Inci-dence is significantly lower in LNCaPgp130act xenografts, suggest-ing a decreased tumorigenic potential.B: Graphical representationof representative tumor volumes of LNCaPgp130act cell linexeno-grafts(clones2Aand408.3)andLNCaPparentalcontrolxenograftsatday41,51, and61post sc injection.

Fig. 8. LNCaPgp130actxenografts formNE/small cell carcinomalike tumors (A) Lowresolution lightmicroscopyof LNCaPgp130actxenograft tumors compared with LNCaP parental cell xenografttumors. Parental tumors have a lower nucleus to cytoplasm ratiocomparedwithLNCaPgp130act tumors,whichhavecharacteristicssimilar to smallcellcarcinoma.ETC,epithelialtumorcell;P,pigment.B:Electronmicrographsrepresenting1mmthick leadcitrate/uranylacetate stained sections of LNCaP gp130act xenograft (a) andLNCaP parental xenograft (b) tumors.Distinct DCVs are presentin LNCaPgp130act xenograft cells with a lownucleus to cytoplasmratio concordant with small cell carcinoma. N, nucleus; CN, con-densednuclei;V,vesicle;DCV,densecorevesicles; sizebar,1mm.

ConstitutiveActivation of gp130 Leads toNEDifferentiation 287

compared with LNCaP parental xenografts (Fig. 7A).Further, tumor volume was also significantly decreas-ed (Fig. 7B).

These data suggest that the decreased proliferationapparent in vitro and the decreased cell body size weresustained in vivo.

TumorMorphologyof LNCaPgp130actXenografts

LNCaP gp130act tumors were examined histologi-cally and under high resolution TEM. Expression ofgp130act significantly alters the tumor morphologywhen compared to parental tumors (Fig. 8A), which isapparent in the representative low resolution haema-toxylin stained tumors. TEM micrographs highlightsmall cell tumor morphology in gp130act xenograftscompared with parental controls (Fig. 8B) and ele-vated DCV formation, which is consistent with NEdifferentiation.

DISCUSSION

Analysis of the induction of the NE phenotypeutilizing the ligand independent gp130act modelhighlighted a complex interaction between NE differ-entiation pathways initiated by IL-6 signal inductionand those used by gp130act. These two models pro-vided valuable clues as to the utilization of gp130 inthe NE differentiation process as overexpression ofgp130act initiated a significant proliferation inhibition,a significant morphology change, DCV formation andinduced elevated expression of the NE marker CgA inLNCaP cells. Thesedata highlight the importanceof thespecific gp130 signal in differentiation of LNCaP cellsto NE-like cells. The significance of this finding wasapparent in vivo, where LNCaP and LNCaP gp130actxenografts were compared. A significant differencewas apparent in the tumor incidence, tumor volumeand histology. Interestingly, both LNCaP and LNCaPgp130act tumors were well perfused when harvested,regardless of the size of the tumor, with little detectiblenecrosis in either population. Analysis of these tumorshistologically determined that, as was apparent byin vitro cellmorphology, different tumormorphologiesexisted. LNCaP gp130act xenograft tumors were moredensely packed than LNCaP parental tumors, withgreater nuclear to cytoplasm ratio and higher organiza-tion apparent under microscopy at high magnifica-tion/high resolution.

Together these data confirm that gp130 plays animportant role in the NE differentiation process, whichmay be independent of alternate IL-6 signaling mecha-nisms. Future investigation of the role of NE differ-entiation utilizing the LNCaP gp130act clones willinvolve investigation of the affects of orthotopicinjection of these cells, to analyze interactions between

these NE like cells and normal prostate cells. Analysisof these affects in castrated animalsmay also determinewhether this process is important in late stage disease.Further analysis would also determine the affects onmetastatic ability, where cells would be labeled to trackcell proliferation and metastases, providing a valuabletool to determine whether the induction of NE dif-ferentiation via gp130 signaling plays a role in themetastatic process.

REFERENCES

1. ACS.Cancer facts andfigures.American cancer society, Inc., 2003.

2. Wu JT,AstillME, LiuGH, StephensonRA. Serum chromograninA: Early detection of hormonal resistance in prostate cancerpatients. J Clin Lab Anal 1998;12(1):20–25.

3. Ferrero-Pous M, Hersant AM, Pecking A, Bresard-Leroy M,Pichon MF. Serum chromogranin-A in advanced prostatecancer. BJU Int 2001;88(7):790–796.

4. Cussenot O, Villette JM, Cochand-Priollet B, Berthon P. Evalua-tion and clinical value of neuroendocrine differentiation inhuman prostatic tumors. Prostate Suppl 1998;8:43–51.

5. Grobholz R, Bohrer MH, Siegsmund M, Junemann KP, Bleyl U,Woenckhaus M. Correlation between neovascularisation andneuroendocrine differentiation in prostatic carcinoma. PatholRes Pract 2000;196(5):277–284.

6. Noordzij MA, van Weerden WM, de Ridder CM, et al.Neuroendocrine differentiation in human prostatic tumormodels. Am J Pathol 1996;149(3):859–871.

7. Gkonos P, Krongrad A, Roos B. Neuroendocrine peptides in theprostate. Urol Res 1995;23:81.

8. Deeble P, Cox M, Wu M, Parsons SJ. Neuroendocrine cells andparacrine signaling in prostate cancer progression. Recent ResDevel Endocrinol 2001;2:357–374.

9. Deeble PD, Murphy DJ, Parsons SJ, Cox ME. Interleukin-6- andcyclic AMP-mediated signaling potentiates neuroendocrinedifferentiation of LNCaP prostate tumor cells. Mol Cell Biol2001;21(24):8471–8482.

10. Hoosein NM. Neuroendocrine and immune mediators in pros-tate cancer progression. Front Biosci 1998;3:D1274–D1279.

11. Cox ME, Deeble PD, Lakhani S, Parsons SJ. Acquisition ofneuroendocrine characteristics by prostate tumor cells is rever-sible: Implications for prostate cancer progression. Cancer Res1999;59(15):3821–3830.

12. Burchardt T, Burchardt M, Chen M-W, Cao Y, De La Taille A,ShabsighA,HayekO,Dorai T, ButtyanR. Transdifferentiationofprostate cancer cells to a neuroendocrine cell phenotype in vitroand in vivo. J Urol 1999;162:1800–1805.

13. Bang YJ, Pirnia F, FangWG, KangWK, Sartor O,Whitesell L, HaMJ, Tsokos M, Sheahan MD, Nguyen P. Terminal neuroendo-crine differentiation of human prostate carcinoma cells inresponse to increased intracellular cyclic AMP. Proc Natl AcadSci USA 1994;91(12):5330–5334.

14. Spiotto MT, Chung TD. STAT3mediates IL-6-induced neuroen-docrine differentiation in prostate cancer cells. Prostate 2000;42(3):186–195.

15. Ueda T, Bruchovsky N, Sadar MD. Activation of the androgenreceptor N-terminal domain by interleukin-6 via MAPK andSTAT3 signal transduction pathways. J Biol Chem 2002;277(9):7076–7085.

288 Palmeret al.

16. Chung TD, Yu JJ, Spiotto MT, Bartkowski M, Simons JW.Characterization of the role of IL-6 in the progression of prostatecancer. Prostate 1999;38(3):199–207.

17. Giri D, OzenM, IttmannM. Interleukin-6 is an autocrine growthfactor in human prostate cancer. Am J Pathol 2001;159(6):2159–2165.

18. Qiu Y, Ravi L, Kung HJ. Requirement of ErbB2 for signalling byinterleukin-6 in prostate carcinoma cells. Nature 1998;393(6680):83–85.

19. Lin DL, Whitney MC, Yao Z, Keller ET. Interleukin-6 inducesandrogen responsiveness in prostate cancer cells through up-regulation of androgen receptor expression. Clin Cancer Res2001;7(6):1773–1781.

20. Hannemann J, Hara T, Kawai M, Miyajima A, Ostertag W,Stocking A. Sequential mutations in the interleukin-3(IL-3)/Granulocyte-macrophage colony stimulating factor/IL-5 recep-tor beta-subunit genes are necessary for the complete coversionto growth autonomy mediates by a truncated beta C subunit.Mol Cell Biol 1995;15(5):2402–2412.

21. Militi S, Riccioni R, Parolini I, Sposi NM, Samoggia P, Pelosi E,Testa U, Peschle C. Expression of interleukin 3 and granulocyte-macrophage colony-stimulating factor receptor common chain

betac, betaIT in normal haematopoiesis: Lineage specificity andproliferation-independent induction. Br J Haematol 2000;111(2):441–451.

22. Hammacher A, Richardson RT, Layton JE, Smith DK, Angus LJ,Hilton DJ, Nicola NA,Wijdenes J, Simpson RJ. The immunoglo-bulin-like module of gp130 is required for signaling byinterleukin-6, but not by leukemia inhibitory factor. J Biol Chem1998;273(35):22701–22707.

23. Lewis PR, Knight Elsevier DP. Cytochemical stainingmethods for Electron Microscopy Series: Practical methods inelectron microscopy, Vol. 14. Editor: Audrey M. Glauert. 1992:321p.

24. Johns T, Mackay I, Callister K, Hertzog P, Devenish R, LinnaneA. Antiproliferative potencies of interferons on melanoma celllines and xenografts: Higher efficacy of interferon beta. J NatCancer Inst 1992;84:1185–1190.

25. Copenhaver WM, Kelly DE, Wood RL. Bailey’s textbook of his-tology, 17th edn. Baltimore,USA:Williams&WilkinsCompany;1979.

26. Palmer J,HertzogPJ,HammacherA.Differential expression andeffects of gp130 cytokines and receptors in prostate cancer cells.Int J Biochem Cell Biol 2004; In Press.

ConstitutiveActivation of gp130 Leads toNEDifferentiation 289