1-s2.0-S0021997503000069-main

download 1-s2.0-S0021997503000069-main

of 7

Transcript of 1-s2.0-S0021997503000069-main

  • 7/28/2019 1-s2.0-S0021997503000069-main

    1/7

    Cytokeratin and Vimentin Expression in Normaland Neoplastic Canine Prostate

    V. Grieco, V. Patton, S. Romussi* and M. Finazzi

    Dipartimento di Patologia, Igiene e Sanita` Pubblica Veterinaria, Sezione di Anatomia Patologica e Patologia Aviare, and

    *Istituto di Clinica Chirurgica e Radiologia Veterinaria, Facolta` di Medicina Veterinaria Universita` degli Studi diMilano, Via Celoria 10, 20133 Milano, Italy

    SummaryIntermediate filament expression in the canine prostate, unlike that in human prostate, is represented inthe literature by only a few reports. In this study, the expression of cytokeratin (CK) and vimentin wasexamined in three normal canine prostates and 11 canine prostatic carcinomas. Monoclonal antibodiesdirected against vimentin, CK AE1/AE3, CK 18 8 (for luminal epithelial cells), CK 5, CK clone 8.12 andCK 14 (for basal cells) were employed. As in man, normal canine prostatic luminal cells were positive forCK 818. Basal cells were positive for CK 5 and CK clone 8.12 but, in contrast to findings in man, werenegative for CK 14. Luminal cells were vimentin-negative, whereas in man they have been reported as

    vimentin-positive. The majority of carcinomas showed an undifferentiated histological pattern and allwere positive for CK AE1/AE3 and for vimentin. Ten tumours were positive for CK 8 12, but six of themshowed many cells co-expressing CK 14. Moreover, in two of these six cases a large number of neoplasticcells also reacted with CK clone 8.12 antibody, and in one of them co-expression of CK 5 was detectable.This co-expression, of luminal and basal cytokeratins, suggests a possible origin of the tumours from

    prostatic epithelial stem cells. Vimentin expression is an inconstant finding in human prostaticcarcinomas; its almost uniform occurrence in canine carcinomas suggests a lesser degree ofdifferentiation than in the human neoplasm.

    q 2003 Elsevier Science Ltd. All rights reserved.

    Keywords:dog; prostatic cytokeratin; prostatic neoplasia; prostatic vimentin; tumour.

    Introduction

    Spontaneous prostatic carcinoma occurs mostoften in human beings and dogs (Logan, 1947;

    Leav and Ling, 1968). Canine prostatic carcinomais apparently less common than its human counter-part, but it represents an insidious, highlymalignant disease (Bell et al., 1991). Immuno-histochemistry with antibodies against intermedi-ate filament proteins is considered to be a usefuladjunct to the methods of studying normal andneoplastic tissue (Altmannsberger et al., 1981; Mollet al., 1982; Cooper et al., 1985). In the past 20 yearsnumerous studies on cytokeratin (CK) and vimen-tin (VIM) expression in normal and neoplastic

    human prostate have been published (Brawer et al.,1985; Molinolo et al., 1985; Kitajima and Tokes,1986; Purnell et al., 1986; Leong et al., 1988; Ito et al.,

    1995; Hudson et al., 2000; Collins et al., 2001). Bycontrast, there are few reports on intermediatefilament expression in the canine prostate (Leavet al., 2001).

    Two types of epithelial cell are recognizable inhuman and canine normal prostatic alveoli andducts: luminal cells and basal cells (Brawer et al.,1985). The luminal cells, which are highly differ-entiated, are the functionally active prostaticsecretory cells. The basal cells, which are lessdifferentiated, are located at the base of

    J. Comp. Path. 2003, Vol. 129, 78 84doi: 10.1016/S0021-9975(03)00006-9, available online at http://www.sciencedirect.com on

    00219975/03/$ - see front matter q 2003 Elsevier Science Ltd. All rights reserved.

  • 7/28/2019 1-s2.0-S0021997503000069-main

    2/7

    the epithelium (Verhagen et al., 1988). Differentialexpression of specific cytokeratins in the luminaland basal cells of the human prostate has beenreported (Sherwood et al., 1991). The luminal cellsexpress CK 8 and CK 18 while the basal cells express

    CK 14 and CK 5, and are also recognized by theanti-CK antibody clone 8.12 (Verhagen et al., 1988;Sherwood et al., 1991; Yang et al., 1997). This clone,the first used to recognize basal prostatic epithelialcells, is directed against CK 13, 15 and 16. Never-theless, for prostatic epithelium, only CK 15 and CK13 need be considered, since CK 16 was notdetected in normal or abnormal prostatic tissue(Nagle et al., 1991; Yang et al., 1997). Vimentin hasbeen demonstrated in both the basal and theparanuclear portions of normal luminal epithelialcells of the human prostate (Leong et al., 1988;

    Nagle et al., 1991).The role of basal cells remained poorly under-stood for many years. It was suggested that basalcells could act as stem cells, differentiating intoluminal cells (Mao and Angrist 1966; Verhagenet al., 1988; Sherwood et al., 1991). This suggestionwas recently confirmed by an elegant in-vitro studyin which the authors demonstrated the presence,in the basal compartment, of a stem cell subpopu-lation which, differentiating into luminal cells,underwent a phase of both basal and luminalcytokeratin expression (Van Leenders et al., 2000).The prostatic basal epithelial cells may also be

    important in carcinogenesis. In fact, even thoughhuman prostatic adenocarcinomas usually show aluminal phenotype (Nagle et al., 1991; Yang et al.,1997; Van Leenders et al., 2000) and express CK 8 18 (Sherwood et al., 1991), prostatic tumoursexpressing both luminal and basal CKs haverecently been described (Gil-Diez de Medina et al.,1998), suggesting a possible origin from differen-tiating stem cells. This latter observation may berelevant to canine prostatic epithelial tumoursbecause, although uncommon, they are generallyundifferentiated and highly malignant (Bell et al.,

    1991). In this study the expression of cytokeratinand vimentin in normal and neoplastic canineprostate was investigated and compared withhuman findings reported in the literature.

    Materials and Methods

    Samples

    Three normal canine prostates from intact adultdogs and 11 prostatic carcinomas were studied. The11 tumours were derived from intact dogs rangingin age from 8 to 15 years. The normal prostatic

    samples and five out of the 11 carcinomas were

    fixed in both neutral buffered formalin and inCarnoys fixative. The other six neoplasms, col-lected from archival sources, were fixed exclusivelyin neutral buffered formalin.

    Histopathology and Immunohistochemistry

    All the samples were dehydrated in alcohol,clarified in xylene and embedded in paraffin wax.Sections (4 mm) were cut and stained with haema-toxylin and eosin (HE) for histological examin-ation. For immunohistochemistry, sections werelabelled by the avidinbiotin-peroxidase complex(ABC) procedure (Hsu et al., 1981) with acommercial immunoperoxidase kit (VectastainStandard Elite; Vector Laboratories, Burlingame,CA, USA). Sections were dewaxed, treated withhydrogen peroxide 0.5% in methanol for 20 minand rehydrated. The monoclonal antibodies(Mabs) used are given in Table 1. On Carnoy-fixed sections, all of the antibodies were appliedwithout any pre-treatment. Formalin-fixed sections,however, were pre-treated with pepsin (Zymed, SanFrancisco, USA) (10 min at 37 8C) for CK 818 orwith microwaves (8 min at 750 W) for CK 14.Treatment with CK 5 and CK clone 8.12 antibodieswas applied only to Carnoy-fixed samples becausethese two antibodies do not react on formalin-fixedsamples, even after microwave pre-treatment. Mabswere incubated at 4 8C overnight. After washing,the sections were covered with an anti-mouse IgGbiotinylated antibody (diluted 1 in 200), and left atroom temperature for 30 min. After washing, theperoxidase-conjugate ABC (diluted 1 in 100) wasallowed to react at room temperature for 30 min.The immunohistochemical reaction was developed

    Table 1Monoclonal antibodies employed in immunohistochemistry

    (sources and dilutions)

    Mabs Dilution

    Vimentin clone 3B4* 1000Vimentin clone V9 20 000CK AE1/AE3 3000CK 818 100CK 14 2000CK 5k 1500CK clone 8.12 1000

    *Dako Corporation, Carpinteria, USA.Sigma Chemical Company, St Louis, USA.Zymed, San Francisco, USA.

    NeoMarkers, Fremont, USA.kThe Binding Site, Birmingham, UK.

    Canine Prostatic Carcinoma 79

  • 7/28/2019 1-s2.0-S0021997503000069-main

    3/7

    with 3,30 diaminobenzidine (1 min) and sectionswere counterstained with Mayers haematoxylin.

    Results

    Normal Prostate

    Immunohistochemical results are summarized inTable 2. Large spectrum cytokeratin AE1/AE3was expressed by all epithelial cells, namely, basaland luminal cells of both glandular alveoli andducts. The luminal cells of normal alveoli and ductsstrongly expressed CK 8 18 (Fig. 1a) but were

    negative with CK clone 8.12, CK 5 and CK 14antibodies. The basal cells were uniformly positivewith CK clone 8.12 (Fig. 1b) and with CK 5 (Fig. 1c)antibodies, giving a typical necklace appearancearound the negative luminal secretory cells. Alveo-lar and ductular basal cells were negative for CK 14.

    Vimentin was detectable in the stromal fibro-blasts and in the endothelial cells of prostaticvessels. Luminal and basal cells of both glandularalveoli and ducts were negative for vimentin.Nevertheless, vimentin expression was oftendetected in the cytoplasm of desquamated epi-

    thelial cells within glandular lumina.

    Prostatic Adenocarcinomas

    The 11 prostatic carcinomas examined showeddifferent histological patterns (Table 3) and inmost of them poor cellular differentiation wasobserved, only four showing well-differentiatedalveolar or acinar gland-like structures lined bycells with a cuboidal or columnar appearance. Bycontrast, in the majority of the cases, among theheavily distorted pseudoalveolar structures werenumerous single, oval to spindle-shaped neoplastic

    cells. In addition to pleomorphism, these scatteredcells showed a high degree of anisocytosis, rangingin diameter from 10 to 80 mm (monster cells). Thecytoplasm, which was abundant and eosinophilic,frequently contained a large empty vacuole, dis-placing the nucleus (signet-ring cells). The nucleiwere large and oval to spindle-shaped, withmarginated chromatin and prominent nucleoli;

    Table 2Cytokeratin and vimentin expression in both the luminal and

    the basal cells of normal prostatic alveoli and ducts

    Immunolabelling in

    Glandular alveoli Ducts

    MabsLuminal

    cellsBasalcells

    Luminalcells

    Basalcells

    Vimentin clone 3B4 2 2 2 2Vimentin clone V9 2 2 2 2CK AE1/AE3 CK 818 2 2CK clone 8.12 2 2 CK 5 2 2 CK 14 2 2 2 2

    2 , negative immunohistochemical reaction; , positive immuno-histochemical reaction.

    Fig. 1. Normal prostate from adult male dog. Immunolabelling

    with (a) CK 818, (b) CK 8.12 and (c) CK 5. Prostaticluminal cells are strongly positive for CK 818. Basalcells express CK 8.12 and CK 5; note the necklacepattern. 480.

    V. Grieco et al.80

  • 7/28/2019 1-s2.0-S0021997503000069-main

    4/7

    binucleated or multinucleated cells were common.In one case (no. 11) there was no evidence ofalveolar formation and the neoplasm was com-posed entirely of elongated spindle-shaped cells,organized in closely packed sheets. Three cases(nos 6,7 and 8) showed diffuse squamous metapla-sia of neoplastic cells.

    Immunohistochemically, all the tumours exam-ined were positive for CKAE1/AE3 (Fig. 2a), butvarious degrees of expression were observed.

    Differentiated alveolar and acinar structuresshowed a uniform and strongly positive signal. Onthe other hand, neoplastic cells scattered singly inthe stroma either expressed low and diffuse levelsof CK AE1/AE3 or, not infrequently, showed astrong positive signal limited to a focal central areaof the cytoplasm. Ten cases were positive for CK 818, although in all these tumours the labellingintensity of the neoplastic cells varied (Fig. 3). Theonly CK 818-negative case (no. 11), which wascomposed entirely of spindle-shaped cells, showedpositive labelling for CK AE1/AE3, confirming the

    epithelial origin of the neoplasm. In six cases,numerous neoplastic cells co-expressing CK AE1/AE3, CK 818 and CK 14 were detectable and, intwo of them, many cells reacting with CK 8.12 werealso detected (Fig. 2b). CK 5 was detected only incase no. 9, in which groups of neoplastic cellsscattered in the stroma reacted with all the anti-CKMabs employed (Fig. 4)

    In all the cases examined, vimentin-positiveneoplastic cells were present. In the well-differen-tiated adenocarcinomas, however, such cells wererare, being scattered among the cells lining thealveolar and acinar neoplastic structures (Fig. 5).

    On the other hand, spindle-shaped neoplastic cells,signet-ring cells and all neoplastic cells scatteredsingly in the stroma consistently showed stronglabelling (Fig. 2c).

    Discussion

    In both the human and canine species, the normalprostatic alveolar and ductular epithelium iscomposed of luminal and basal cells. Our study

    demonstrated that the luminal epithelial cells ofthe normal canine prostate express CK 818, asthey do in man. On the other hand, somedifference were noted between man and the dogin respect of cytokeratin expression by basal cells.Thus, while immunolabelling by clone 8.12 and CK5 antibodies in the present study was similar to thatreported in man (Verhagen et al., 1988; Sherwoodet al., 1991; Yang et al., 1997), our study showed thatthe normal prostatic basal cells of the dog were,unlike those of man, uniformly negative for CK 14.The latter finding contrasts, however, with that of

    Leavet al. (2001), who described diffuse expressionof CK 14 in both alveolar and ductular canine basalcells. Nevertheless, the antibody used in that studyrecognized both CK 5 and CK 14; as a result, thepositive signal may have been due to the expressionof CK 5 alone, and not to the co-expression of bothCK 5 and CK 14. In fact, it is assumed that one ofthe rules governing the expression of keratins isbased on the existence of specific keratin pairs,composed of two cytokeratins (one acid and onebasic) whose expression characterizes differentkinds of epithelia. As far as the prostate isconcerned, the two main pairs are represented by

    Table 3Histological pattern and cytokeratin and vimentin expression of the 11 canine prostatic carcinomas examined

    Immunolabelling with the stated antibodies

    Cytokeratin Vimentin

    Case no. Histological pattern AE1/AE3 8 18 8.12 5 14 3B4 V9

    1 Acinar AD nd nd 2 22 Intra-alveolar AD nd nd 2 /23 Intra-alveolar AD nd nd 4 Intra-alveolar AD nd nd 2 25 Intra-alveolar (poorly differentiated) AD nd nd 2 6 Intra-alveolar/squamous AD nd nd 7 Intra-alveolar/squamous AD 2 2 8 Poorly differentiated squamous cell carcinoma ne 9 Undifferentiated carcinoma 10 Undifferentiated carcinoma 2 2 11 Undifferentiated carcinoma 2 2 2 2 2

    AD, adenocarcinoma; nd, not done; ne, not evaluated because of high background staining.

    Canine Prostatic Carcinoma 81

  • 7/28/2019 1-s2.0-S0021997503000069-main

    5/7

    CK 818, expressed in the luminal cells, and CK515, expressed in the basal cells. Therefore, thesimilar signal that we obtained in the basal cellswith CK 5 and CK clone 8.12 antibodies may havebeen due to the fact that clone 8.12 marks CK 15,which in turns forms a pair with CK 5. By contrast,since CK 5 and CK 14 are not considered a keratinpair, it is possible that whenever a cell is labelledpositively with an antibody directed against boththe CKs, in reality it expresses only one of them.

    In our study, normal canine prostatic luminalcells were totally negative with both the anti-vimentin antibodies employed (3B4 and V9). Inthe luminal epithelial cells of the human prostate,the expression of vimentin was demonstrated withthe same two Mabs (Leong et al., 1988; Nagle et al.,1991). Vimentin expression is primarily confinedto mesenchymal cells, although co-expression ofcytokeratin and vimentin is detectable in certainepithelia such as the renal tubular epithelium andthe ovarian epithelium (Holthofer et al., 1983;Leong et al., 1988). Embryologically, these tissuesderive from a part of the mesenchyme that hasundergone epithelial transformation. Therefore,the expression of vimentin is understandable forthese epithelia, whereas, according to Leong et al.(1988) and to Nagle et al. (1991), a readyexplanation is not available for this finding inhuman prostatic epithelium.

    Fig. 3. Prostatic alveolar adenocarcinoma. Immunolabellingfor CK 818. Neoplastic cells show different intensitiesof positive signal. 300.

    Fig. 4. Prostatic undifferentiated carcinoma. Immunolabellingfor CK 5. Positive neoplastic cells scattered in theprostatic stroma. 400.

    Fig. 2. Prostatic undifferentiated carcinoma. Immunolabellingwith (a) CK AE1/AE3, (b) CK 8.12, and (c) VIM.

    Neoplastic cells co-express cytokeratins and vimentin. 200.

    V. Grieco et al.82

  • 7/28/2019 1-s2.0-S0021997503000069-main

    6/7

    As far as the 11 canine prostatic carcinomas wereconcerned, they frequently showed a highlyundifferentiated pattern and consistently demon-strated the expression of vimentin (Table 3). Sincethe expression of vimentin in epithelial cells isgenerally associated with loss of cell-to-cell contact,this finding suggests for canine prostatic tumours amalignant and aggressive behaviour that mayincrease the risk of metastasis. Conversely, theexpression of vimentin in desquamated epithelialcells observed in normal prostate may be related totheir independent existence, following their

    detachment from an epithelial sheet (Leong et al.,1988).

    Independently from the expression of vimentin,all the tumours examined expressed CK AE1/AE3,confirming their epithelial origin. Ten cases werepositive for CK 8 18, demonstrating a perfectanalogy with human prostatic tumours, whichusually show a luminal phenotype (Nagle et al.,1991; Van Leenders et al., 2000). Six of the tumoursexamined also expressed CK 14. This finding inhuman pathology suggests a possible origin frombasal stem cells which, as already mentioned, may

    be able to express both luminal and basal cytoker-atins during the differentiating phase. We demon-strated that canine prostatic basal cells do not showCK 14. In any event, for two out of these six CK 14-positive cases (nos 8 and 9) (Table 3), the possibleorigin from differentiating basal stem cells could bedemonstrated by the co-expression of other basalcytokeratins such as CK 5 and CK clone 8.12. Forthe other four cases (nos 3, 6, 7 and 10) thepresence of CK 14 could be explained by the factthat, although epithelial tissues tend to retaintheir characteristic cytokeratins during malignanttransformation, modulation may occur during

    the development and progression of carcinomas,as observed by Malzahn et al. (1998) in breastcarcinomas. Moreover, cases 6 and 7 showed asquamous pattern, and CK 14 is typically expressedby squamous cells (Chu et al., 2001).

    The dog is frequently used as an animal model tostudy human prostatic pathology. Our studydemonstrated that in the dog prostatic adenocarci-nomas may sometime derive from basal stem cells,as recently demonstrated in man. On the otherhand, our results showed that there are somedifferences between cytokeratin and vimentinexpression in the normal canine and humanprostate. Moreover, all canine prostatic carcinomasexamined consistently showed expression of vimen-tin, demonstrating that they are usually moreundifferentiated than their human counterparts.

    These differences should be borne in mind in anycomparative studies of prostatic neoplasia.

    Acknowledgments

    M. Barzani for technical assistance andS. Conversano for English version review aregratefully acknowledged.

    References

    Altmannsberger, M., Osborn, M., Schaufer, A. andWeber, K. (1981). Antibodies to different intermedi-ate filament proteins. Cell type-specific markers onparaffin embedded human tissue. Laboratory Investi-gation, 45, 427434.

    Bell, F. W., Klausner, J. S., Hayden, D. W., Feeney, D. A.and Johnston, S. D. (1991). Clinical and pathologicfeatures of prostatic adenocarcinoma in sexuallyintact and castrated dogs: 31 cases (1970 1987).

    Journal of the American Veterinary Medical Association,199, 16231630.

    Brawer, M. K., Peehl, D. P., Stamwy, T. A. and Bostwick,D. G. (1985). Keratin immunoreactivity in the benign

    and neoplastic human prostate. Cancer Research, 45,36633665.Chu, P. G., Lyda, M. H. and Weiss, L. M. (2001).

    Cytokeratin 14 expression in epithelial neoplasms: asurvey of 435 cases with emphasis on its value indifferentiating squamous cell carcinomas from otherepithelial tumours. Histopathology, 39, 916.

    Collins, A. T., Habib, F. K., Maitland, N. J. and Neal, D. E.(2001). Identification and isolation of human pros-tate epithelial stem cells based on a2b1-integrinexpression. Journal of Cell Science, 114, 38653872.

    Cooper, D., Schermer, A. and Sun, T. T. (1985).Classification of human epithelia and their neoplasmsusing monoclonal antibodies to keratins: strategies,

    Fig. 5. Prostatic alveolar adenocarcinoma. Immunolabellingfor vimentin. Sparse positive neoplastic cells in a largepseudoalveolar structure. 300.

    Canine Prostatic Carcinoma 83

  • 7/28/2019 1-s2.0-S0021997503000069-main

    7/7

    application, and limitation. Laboratory Investigation,52, 243256.

    Gil-Diez de Medina, S., Salomon, L., Colombel, M.,Abbou, C. C., Bellot, J., Thiery, J. P., Radvanyi, F.,Van der Kwast, T. H. and Chopin, D. K. (1998).Modulation of cytokeratin subtype, EGF receptor,and androgen receptor expression during pro-gression of prostate cancer. Human Pathology, 29,10051012.

    Holthofer, H., Miettinen, A., Paasivuo, R., Lehto, V. P.,Linder, E., Alfthan, O. and Virtanen, I. (1983).Cellular origin and differentiation of renal carci-nomas. A fluorescence microscopic study with kidneyspecific antibodies, anti-intermediate filament anti-bodies, and lectin. Laboratory Investigation, 49,317326.

    Hsu, S. M., Raine, L. and Fanger, H. (1981). Use ofavidinbiotin-peroxidase complex (ABC) in immu-noperoxidase techniques: a comparison between ABC

    and unlabeled antibody (PAP). Journal of Histochem-istry and Cytochemistry, 29, 577580.

    Hudson, D. L., OHare, M., Watt, F. M. and Masters, J. R.W. (2000). Proliferative heterogeneity in the humanprostate: evidence for epithelial stem cells. LaboratoryInvestigation, 80, 12431250.

    Ito, T., Furusato, M., Akiyama, A., Kato, H. and Aizawa, S.(1995). A clinical and immunohistochemical study ofpapillary adenocarcinoma of the prostate. Prostate, 26,2327.

    Kitajima, K. and Tokes, Z. A. (1986). Immunohistochem-ical localization of keratin in human prostate. Prostate,9, 183190.

    Leav, I. and Ling, G. V. (1968). Adenocarcinoma of thecanine prostate. Cancer, 22, 13291345.

    Leav, I., Schelling, K. H., Adams, J. Y., Merk, F. B. andAlroy, J. (2001). Role of canine basal cells in prostaticpost natal development, induction of hyperplasia, sexhormone- stimulated growth; and the ductal origin ofcarcinoma. Prostate, 47, 149163.

    Leong, A. S.-Y., Gilham, P. and Milios, J. (1988).Cytokeratins and vimentin intermediate filamentproteins in benign and neoplastic prostatic epi-thelium. Histopathology, 13, 435442.

    Logan, M. J. (1947). The pathology of the prostate glandof man and the dog: a review. Cornell Veterinarian, 37,241253.

    Malzahn, K., Mitze, M., Thoenes, M. and Moll, R. (1998).Biological and prognostic significance of stratified

    epithelial cytokeratins in infiltrating ductal breastcarcinomas. Virchows Archiv, 433, 119129.

    Mao, P. and Angrist, M. D. (1966). The fine structure ofthe basal cell of human prostate. Laboratory Investi-gation, 15, 17681782.

    Molinolo, A. A., Meiss, R. P., Leo, P. and Sems, A. I.(1985). Demonstration of cytokeratins by immuno-peroxidase staining in prostatic tissue. Journal of Urology, 134, 10371040.

    Moll, R., Franke, W. W. and Sciller, D. L. (1982). Thecatalog of human cytokeratins: patterns of expressionin normal epithelia, tumors and cultured cells. Cell,31, 1124.

    Nagle, R. B., Brawer, M. K., Kittelson, J. and Clark, V.(1991). Phenotypic relationships of prostatic intrae-pithelial neoplasia to invasive prostatic carcinoma.American Journal of Pathology, 138, 119140.

    Purnell, D. M., Heatfield, B. M., Anthony, R. L. andTrump, B. F. (1986). Immunohistochemistry of the

    cytoskeleton of human prostatic epithelium. AmericanJournal of Pathology, 126, 384395.

    Sherwood, E. R., Theyer, G., Steiner, G., Berg, L. A.,Kozlowski, J. M. and Lee, C. (1991). Differentialexpression of specific cytokeratin polypeptides in thebasal and luminal epithelia of the human prostate.Prostate, 18, 303314.

    Van Leenders, G., Dijkman, H., Hulsbergen-van de Kaa,C., Ruiter, D. and Schalken, J. (2000). Demonstrationof intermediate cells during human prostate epi-thelial differentiation in situ and in vitro using triple-staining confocal scanning microscopy. LaboratoryInvestigation, 80, 12511258.

    Verhagen, A. P. M., Aalders, T. W., Ramaekers, F. C. S.,Debruyne, F. M. J. and Schalken, J. A. (1988).Differential expression of keratins in the basal andluminal compartments of rat prostatic epitheliumduring degeneration and regeneration. Prostate, 13,2538.

    Yang, Y., Han, J., Liu, X., Dalkin, B. and Nagle, R. B.(1997). Differential expression of cytokeratin mRNAand protein in normal prostate, prostatic intraepithe-lial neoplasia, and invasive carcinoma. American

    Journal of Pathology, 150, 693704.

    Received; August 2nd; 2002

    Accepted;January 10th; 2003" #

    V. Grieco et al.84