Short-term culturing of low-grade superficialbladder transitional cell carcinomas leads tochanges in the expression levels of severalproteins involved in key cellular activities

Fresh, superficial transitional cell carcinomas (TCCs) of low-grade atypia (3 grade I,Ta; 6 grade II, Ta), as well as primary cultures derived from them were labeled with[35S]methionine for 16 h, between 2 and 6 days after inoculation. Whole proteinextracts were subjected to IEF (isoelectric focusing) two-dimensional polyacrylamidegel electrophoresis (2-D PAGE) followed by autoradiography. Proteins were identifiedby a combination of proteomic technologies that included microsequencing, massspectrometry, 2-D PAGE immunoblotting and comparison with the bladder TCC proteindatabase available on the internet (http://biobase.dk/cgi-bin/celis). Comparison of theIEF 2-D gel protein profiles of fresh tumors and their primary cultures showed that theoverall expression profiles were strikingly similar, although differing significantly in thelevels of several proteins whose rate of synthesis was differentially regulated in at least85% of the tumor/culture pairs as a result of the short-term culturing. Most of theproteins affected by culturing were upregulated and among them we identifiedcomponents of the cytoskeleton (keratin 18, gelsolin and tropomyosin 3), a molecularchaperone (hsp 28), aldose reductase, GST p, metastasin, synuclein, the calreticulinprecursor and three polypeptides of unknown identity. Only four major proteins weredownregulated, and these included two fatty acid-binding proteins (FABP:FABP5 andA-FABP) which are thought to play a role in growth control, the differentiation-associated keratin 20, and the calcium-binding protein annexin V. Proteins that weredifferentially regulated in only some of the cultured tumors included alpha-enolase,triosphosphate isomerase, members of the 14-3-3 family, hnRNPs F and H, PGDH,hsp (heat-shock protein) 60, BIP, the interleukin-1 receptor antagonist, the nucleolarprotein B23, as well as several proteins of yet unknown identity. The suitability of invitro bladder tumor culture models to study complex biological phenomena such asmalignancy and invasion is discussed.

Keywords: Transitional cell carcinomas / Short-term culturing / Proteome profiling EL 3277

Ariana Celis1*

Hanne H. Rasmussen1

Pamela Celis1

Bodil Basse1

Jette B. Lauridsen1

Gitte Ratz1

Bente Hein2

Morten éstergaard1

Hans Wolf2

Torben érntoft3

Julio E. Celis1

1Department of MedicalBiochemistry and DanishCentre for Human GenomeResearch, The University ofAarhus, Aarhus, Denmark

2Department of Urology,Skejby Hospital, Aarhus,Denmark

3Department of ClinicalBiochemistry, SkejbyHospital, Aarhus, Denmark

1 Introduction

Bladder cancer comprises a broad spectrum of tumorsthat include transitional cell carcinomas (TCCs), squa-mous cell carcinomas (SCCs), adenomas and a few othertumor types that are less frequent [1, 2]. TCCs are by farthe more prevalent tumors and represent nearly 90% of allbladder cancers in the Western hemisphere [3]. InDenmark, TCC is a common cancer with an incidenceof 1200±1300 per year and a mortality rate of approx-imately 300 per year. TCCs are subdivided into non-

invasive papillary and nonpapillary invasive carcinomatypes [1, 2] that are believed to orginate from differentgenetic alterations ([4±6]; see Fig. 1 of [7]). Superficialpapillary TCCs correspond to 70% of all TCCs and areusually of low grade and noninvasive at the time ofpresentation. These tumors begin as areas of hyperplasiathat later undergo a process of dedifferentiation (grades I±IV). Invasive tumors may arise from these lesions,although poorly differentiated neoplasms have a highertendency to invade and metastasize [2]. Papillary TCCsare characterized by multiple recurrences; of these, 10±30% will progress to invasive disease. In general, multi-focal recurrent papillary tumors provide a unique modelsystem to study the molecular mechanisms underlying thevarious steps involved in cancer development, and offer avaluable source of material to search for specific markersfor early tumor detection and prognosis. Nonpapillary

Correspondence: Professor J. E. Celis, Department of MedicalBiochemistry and Danish Centre for Human Genome Research,The University of Aarhus, Ole Worms Alle Building 170, DK-8000Aarhus C, DenmarkE-mail: jec@biokemi.au.dkFax: +45-86131160

Abbreviations: FABP, fatty acid-binding protein; SCC, squa-mous cell carcinoma; TCC, transitional cell carcinoma

� WILEY-VCH Verlag GmbH, 69451 Weinheim, 1999 0173-0835/99/0202-0355 $17.50+.50/0

Electrophoresis 1999, 20, 355±361 355

Pro

teo

mic

san

d2-

DE

* Deceased July 7, 1998

invasive carcinomas, on the other hand, are believed todevelop from carcinoma in situ (CIS), a flat lesion ofuncertain biological behavior that is often associated withhigh-grade invasive tumors [8]. CIS neoplasms areusually of high grade (grade III), and are detected latein the disease, often with invasion.

To date, many genetic alterations have been uncovered inbladder cancer. Cytogenetic and molecular biology datahave indicated that chromosome 9 and 17 are frequentlyaltered in bladder urothelial tumors [5, 9, 10], and there iscompelling evidence that p53 [5, 6, 11] is involved in thedevelopment and progression of TCCs. Spruck et al. [6]have shown that chromosome 9 deletions occur earlyduring progression of superficial papillary tumors leadingto a ªhyperplastic stageº, whereas p53 mutations appearlater in the process confering invasive properties. Thissituation, however, is inverse in CIS, as a large fraction ofthese tumors contain p53 mutations. Chromosome 9alterations in CIS have been shown to occur later, or notat all [6, 12]. Besides pinpointing two divergent pathwaysof bladder tumor progression, these studies have sug-gested that the order in which the genetic changes takeplace is important in determining the outcome of thelesion.

In our laboratories we are interested in identifyingchanges in gene expression that accompany malignanttransformation of the bladder urothelium in an effort togenerate markers that may be used to grade these

tumors in a more objective way and that may identifythose lesions that are at risk of invasive disease [13±16].In due course, these markers will pave the way for futurestudies intended to dissect the molecular mechanismsunderlying progression. So far, we have examined theprotein expression profiles of hundreds of fresh non-cultured TCCs and SCCs using proteomic technology[13±16], but have considered the possibility of analyzingshort-term cultured cells prepared from these lesions, asthese cells could be extensively expanded and manipu-lated in vitro. Here we present data ± based on theanalysis of superficial, noninvasive low-grade TCCs ±showing that short-term culturing (2±6 days) of theselesions results in the altered expression of severalproteins that are involved in key cellular activities.

2 Materials and methods

2.1 Tumor biopsies

Bladder tumors were collected, upon informed consentfrom the patients, according to institutional guidelines.The tumors were obtained fresh from surgery at SkejbyHospital, Aarhus, Denmark, placed on ice, and immedi-ately transported to the Department of Medical Biochem-istry [14].

2.2 Primary cultures

Tumors were minced with the aid of a scalpel and placedin a 10 mL conical plastic tube. After mechanical shaking,

356 A. Celis et al. Electrophoresis 1999, 20, 355±361

Figure 1. IEF 2-D PAGE autoradiograms of [35S]methionine-labeled proteins from (A) a fresh TCC (TCC 532-1), and(B) a primary culture derived thereof. The primary culture was labeled 2 days after plating. Green arrows indicate proteinsthat are downregulated, red arrows indicate proteins that are upregulated, and black arrows indicate proteins that are eitherup-or downregulated in 85% of the tumor/culture pairs.

Electrophoresis 1999, 20, 355±361 Proteome profiles of cultured bladder TCCs 357

Table 1 Major TCC proteins affected by culturing

SSPb) Identity Apparent pI Ratio primary culturea)/tumorNo. molecular MOB MOB MOB MOB MOB MOB MOB MOB

mass 532-1 519-1 521-1 916-1 524-1 603-1 709-1 844-1(kDa) Grade Grade Grade Grade Grade Grade Grade Grade

II, Ta I, Ta I, Ta I, Ta II, Ta II, Ta II, Ta II, Ta

Upregulated in at least 85 %of the tumor/primary culturepairs

1102 Aldose reductase 37.0 6.8 : Uc) : : : : : :1107 Unknown 41.8 6.6 ; : : : : : : :1305 Unknown 56.3 6.6 ; : : n.d.d) : : : :3206 Unknown 45.0 6.1 ; : : : : : : U4002 Unknown 32.0 5.9 : : : : : : : :4004 hsp28 25.5 6.0 : : U : U : : :4105 Annexin IV 34.2 5.8 : : : : : : : :4610 Gelsolin 88.2 5.8 : : : : : : : :5005 GSTp 25.4 5.6 ; : : : : : : :5007 Metastasin 9.5 5.6 : : : : : : n.d n.d5332 Keratin 18 43.5 5.3 : : : : : : : :8025 Synuclein 13.0 4.6 : U U : : : : :9004 Tropomyosin 3 33.2 4.5 : : : : : : : :9403 Calreticulin precursor 63.0 4.1 : ; : : : ; : n.d9803 Involucrin 123,9 4.5 : : : : ; : : n.d

Downregulated in at least 85 %of the tumor/primary culturepairs

2020 PA-FABP (FABP 5) 12.4 6.3 ; ; ; ; ; ; ; ;3003 A-FABP (FABP 4) 11.7 6.2 ; ; ; ; ; ; n.ee ;5202 Keratin 20 45.7 5.6 ; ; ; ; ; ; ; ;8002 Annexin V 33.5 4.8 ; ; ; U ; ; ; ;

Altered only in a few of thetumor/primary culture pairs

201 a-Enolase 46.8 7.4 ; : : ; : : U ;212 High-affinity hepatic-bile

binding protein 36.7 7.5 : ; ; ; : : : ;310 Unknown 53.9 7.1 ; ; : U ; ; U U1017 GSTmm 31.0 6.7 ; : : n.e U n.e n.e ;1019 Triosephophate isomerase 25.5 6.9 ; U U ; : : U ;1302 Unknown 54.3 6.9 ; U : U : ; U ;1402 Unknown 63.4 6.7 ; : : U : : U U3002 S100c 10.4 6.2 : U U U : : U U4107 Unknown 35.5 5.7 ; U ; U U U U U4303 hnRNP H 52.8 5.9 ; U U U : ; : ;5001 PGDH 25.5 5.7 : ; ; ; ; : : ;6002 Unknown 18.5 5.33 : U U : : : U U6025 Interleukin-1 recpetor anta-

gonist 19.1 5.0 : U U U : : U U7202 hnRNNP F 46.3 5.2 ; U U U : ; : ;7301 hsp 60 56.7 5.2 ; : U : ; ; : ;8005 14-3-3 b 30.3 4.6 ; U U U : U U U8103 Nucleolar protein B23 37.9 4.7 ; U U ; U ; ; U8502 BIP 75.4 4.8 ; ; U : U : : ;9001 14-3-3 z 31.2 4.6 ; U U U : U U U9007 14-3-3 Z 31.1 4.4 ; U U U : : U U

the cell suspension was centrifuged, washed 3 times withDMEM containing 10% fetal calf serum and plated atdifferent cell concentrations in 24-well microtiter plates.Cells bound to the surface of the wells within 16 h ofinoculation.

2.3 [35S]Methionine labeling of tumors andprimary cultures

The procedures for labeling tumors and primary cultureswith [35S]methionine have been described in detail [14].

2.4 Two-dimensional gel electrophoresis

2-D PAGE was carried out as previously described [17].

3 Results

3.1 Proteome expression profiles of freshsuperficial TCCs and primary culturesderived thereof

Fresh superficial TCCs of low-grade atypia (Table 1), aswell as primary cultures derived from them, were labeledwith [35S]methionine for 16 h, between 2 and 6 days afterinoculation. Whole protein extracts were subjected to IEF2-D PAGE; representative 2-D gel autoradiograms of onesuch lesion (TCC-532-1) are shown in Fig. 1A (freshtumor) and B (primary culture, two days after plating).Visual comparison of the IEF 2-D gels showed that theoverall expression profiles of the tumor and the primaryculture were remarkably similar, although they differedsubstantially in the levels of several proteins whose rate ofsynthesis was disregulated as a result of the short-termculturing (Fig. 1, Table 1). For simplicity, TCC 532-1proteins that showed altered expression in at least 85% ofthe tumor/primary culture pairs are marked either red(upregulated) or green (downregulated) in Fig. 1, and are

listed at the beginning of Table 1. TCC 532-1 proteinswhose levels are affected in only some of the pairs areindicated in black in Fig. 1 and are listed at the end ofTable 1. Proteins were identified using a combination ofproteomic technologies that included microsequencing,mass spectrometry, 2-D PAGE immunoblotting as well ascomparison with the TCC protein database available onthe internet ([18] and references therein; also see [7];http://biobase.dk/cgi-bin/celis).

3.2 Protein changes observed in at least 85% ofthe tumor/primary culture pairs

Most of the major proteins whose levels were affected inat least 85% of the tumor/culture pairs were upregulatedas judged by visual inspection of the autoradiograms(indicated with red arrows in Fig. 1). Among them weidentified components of the cytoskeleton (keratin 18,gelsolin and tropomyosin 3), a molecular chaperone (hsp28), aldose reductase, GST p, metastasin, synuclein, thecalreticulin precursor and three polypeptides of unknownidentity. Only four major proteins were downregulated,and these included two fatty acid-binding proteins (FABP5 and A-FABP) thought to play a role in growth control,the differentiation-associated keratin 20, and the calcium-binding protein annexin V (indicated with green arrows inFig. 1). For reference, Figs. 2 and 3 display selectedareas of 2-D gels of various tumor/primary culture pairsshowing altered expression of some of the proteinsmentioned above. The striking downregulation of proteinsobserved in the short-term cultured TCCs may notrepresent an exception, as very low levels of FABP 5,A-FABP, and keratin 20 have been observed in a fewbladder cell lines such as T24 (p80, Fig. 4), Hu 609 (p43,not shown), and HCV29 (not shown). So far, we havefound only one bladder cell type, RT4, that expresses highlevels of these proteins (results not shown).

358 A. Celis et al. Electrophoresis 1999, 20, 355±361

Table 1 continued

SSPb) Identity Apparent pI Ratio primary culturea)/tumorNo. molecular MOB MOB MOB MOB MOB MOB MOB MOB

mass 532-1 519-1 521-1 916-1 524-1 603-1 709-1 844-1(kDa) Grade Grade Grade Grade Grade Grade Grade Grade

II, Ta I, Ta I, Ta I, Ta II, Ta II, Ta II, Ta II, Ta

9009 14-3-3 s 30.0 4.4 ; ; ; ; : ; U U

a) TCCs 532-1, 916-1, 603-1, 709-1 and 844-1 were labeled with [35S]methionine 2 days after plating. TCCs 519-1, 521-1and 524-1 were labeled 6 days after plating.

b) SSP, sample spot numberc) U, unchangedd) n.d., not determinede) n.e., not expressed in the tumor

3.3 Proteins showing variations in only some ofthe tumor/primary culture pairs

Proteins differentially regulated in only some of theprimary cultures included alpha-enolase, triosephosphateisomerase, members of the 14-3-3 family, hnRNPs F andH, PGDH, hsp 60, BIP, the interleukin-1 receptorantagonist, the nucleolar protein B23 as well as severalproteins of unknown identity (indicated with black arrowsin Fig. 1, Table 1).

4 Discussion

The results presented in this article reveal strikingdifferences in the proteome expression profiles of low-grade, papillary bladder TCCs cultured for as short asa few days. Some of the changes, in particular thoseinvolving cytoskeletal proteins, are not surprising as theymust reflect changes in cell shape, motility, aggregationand cytoarchitecture that accompany adaptation to cellculture. For example, gelsolin, a protein that caps thebarbed ends of severe actin filaments, has been shown toact as a downstream effector of rac for fibroblast motility[19]. Gelsolin null dermal fibroblasts exhibit a muchreduced ruffling response to serum or epidermal growthfactor (EGF) stimulation and, therefore, the upregulationof this protein in the primary cultures most likely reflectsan increased motility of these cells.

The increase in keratin 18 observed in all the cultures ispuzzling as it was not accompanied by an upregulation ofkeratin 8 with whom it pairs to maintain, in part, thestructural integrity of the cells [20, 21]. Keratin 18 hasbeen involved in cell morphology, motility and cellproliferation, and it is a direct target of the ras signalingtransduction pathway [22]. The Ser33 phosphorylatedvariant of keratin 18 binds to the 14-3-3- proteins, one ofwhich, 14-3-3 s (also known as stratifin [23]), has beenshown to be a p53-regulated inhibitor of G2/M [24]. Wehave previously shown that the levels of 14-3-3 s arehighly sensitive to transformation in epithelial cells [23],and the fact that five out of eight tumor/primary culturepairs exhibited a marked downregulation of this proteinsuggests that most of the cultures are proliferating, an

Electrophoresis 1999, 20, 355±361 Proteome profiles of cultured bladder TCCs 359

Figure 2. IEF 2-D PAGE autoradiograms of [35S]methio-nine-labeled proteins from fresh TCCs and a primaryculture derived thereof. (A, C, E) Fresh TCCs 519-1,521-1 and 844-1, respectively. (B, D, F) Primary culturesof TCCs 519-1, 521-1, and 844-1, respectively.

Figure 3. IEF 2-D PAGE autoradiograms of [35S]methio-nine-labeled proteins from (A) a fresh TCC (TCC 532-1),and (B) a primary culture derived thereof. The primaryculture was labeled 2 days after plating.

observation that is also supported by the significantdecrease observed in the levels of the two fatty acid-binding proteins (FABP 5 and A-FABP) that are thought toplay a role in inhibiting cell proliferation [14, 15]. Clearly,these cells show an abnormal expression of severalproteins (aldose reductase, hsp28, annexins IV and V,GST p, metastasin, synuclein, calreticulin precursor) thatplay important roles in diverse cellular activities, andtherefore, they do not represent a suitable model to studychanges in the proteome expression profile of cellsundergoing complex biological phenomena such asinvasion and metastasis, as these require an interplay

between the cancer cells and host tissue. These models,however, may provide a consistent source of cells tostudy well-defined, intermediate steps.

An important outcome of the studies reported here is thatseveral major proteins were up- or downregulated in onlya subset of the tumor/primary culture pairs in spite of thefact that the tumors were reported to have the same gradeof atypia and stage. BIP for example was found to bedownregulated in two out of the five grade II, Ta TCCsanalyzed (TCCs 532-1 and 844-1), unchanged in one(TCC 524-1), and upregulated in the other two (TCCs

360 A. Celis et al. Electrophoresis 1999, 20, 355±361

Figure 4. IEF 2-D PAGE autoradiogram of [35S]methionine-labeled proteins from bladder cell line T24.

603-1 and 709-1; (Table 1). These apparently contra-dictory data, we believe, reflect current limitations in thegrading system which takes into consideration mainlymorphological features. A more objective grading systemmay not only aid prognosis and treatment but will alsofacilitate the interpretation of studies such as the onereported here.

The authors would like to thank Inge Detlefsen for typingthe manuscript. Hanne H. Rasmussen and Mortenéstergaard were supported by fellowships from theDanish Cancer Society. The study was supported bygrants from the Danish Cancer Society and the DanishBiotechnology Program.

Received September 30, 1998

5 References

[1] Friedell, G. H., Nagy, G. K., Cohen, S. M., in: Bryan, G. T.,Cohen, S. M. (Eds.), The Pathology of Bladder Cancer,CRC Press, Boca Raton 1983, Vol. I, pp. 11±42.

[2] Pauli, B. U., Alroy, J., Weinstein, R. S., in: Bryan, G. T.,Cohen, S. M. (Ed.), The Pathology of Bladder Cancer, CRCPress, Boca Raton 1983, Vol. II, pp. 41±140.

[3] Mostofi, F. K., Davis, C. J., Jr., Sesterhenn, I., J. Occ. Med.1990, 32, 793±796.

[4] Heney, N. M., Ahmed, S., Flanagar, M. J., Frable, W.,Corder, M. P., Hafermann, M. D., Hawkins, I. R., J. Urol.1983, 130, 1083±1086.

[5] Simmoneau, A. R., Jones, P. A., World J. Urol. 1994, 12,89±95.

[6] Spruck III, C. H., Ohneseit, P. F., Gonzales-Zuelta, M.,Esrig, D., Miyao, N., Tsai, Y. C., Lerner, S. P., Schmütte, C.,Yang, A. S., Cote, R., Dubeau, L., Nichols, P. W., Hermann,G. G., Steven, K., Horn, T., Skinner, D. G., Jones, P. A.,Cancer Res. 1994, 54, 784±788.

[7] Celis, J. E., éstergaard, M., Rasmussen, H. H., Gromov, P.,Gromova, I., Varmark, H., Palsdottir, H., Magnusson, N.,Andersen, I., Basse, B., Lauridsen, J. B., Ratz, G., Wolf, H.,érntoft, T. F., Celis, P., Celis, A., Electrophoresis 1999, 20,300±309.

[8] Lamm, D. L., Urol. Clin. North Am. 1992, 19, 499±508.

[9] Perucca, D., Szepetowski, P., Simon, M. P., Gaudray, P.,Cancer. Genet. Cytogenet. 1990, 49, 143±156.

[10] Presti, J. C., Jr., Reuter, V. E., Galan, T., Fair, W. R.,Cordon-Cardo, C., Cancer Res. 1991, 51, 5405±5409.

[11] Sidransky, D., von Eschenbach, A., Tsai, Y. C., Jones, P.,Summerhayes, I., Marshall, F., Paul, M., Green, P.,Hamilton, S. R., Frost, P., Vogelstein, B., Science 1991,252, 706±709.

[12] Rosin, M. P., Cairns, P., Epstein, J. I., Schoenberg, M. P.,Sidransky, D., Cancer Res. 1995, 55, 5213±5216.

[13] Celis, J. E., Rasmussen, H. H., Vorum, H., Madsen, P.,HonorØ, B., Wolf, H., érntoft, T. F., J. Urol. 1996, 155,2105±2112.

[14] Celis, J. E., éstergaard, M., Basse, B., Celis, A., Lauridsen,J. B., Ratz, G. P., Andersen, I., Hein, B., Wolf, H., érntoft, T.F., Rasmussen, H. H., Cancer Res. 1996, 56, 4782±4790.

[15] éstergaard, M., Rasmussen, H. H., Nielsen, H. V., Vorum,H., 0rntoft, T. F., Wolf, H., Celis, J. E., Cancer Res. 1997,57, 4111±4117.

[16] Celis, J. E., éstergaard, M., Celis,P., Rasmussen, H. H.,Wolf, H., Bio-Radiations 1998, 101, 24±27.

[17] Celis, J. E., Ratz, G., Basse, B., Lauridsen, J. B., Celis, A.,Jensen, N. A., Gromov, P., in: Celis, J. E., Carter, N. P.,Shotton, D. M., Hunter, T., Simons, K. (Eds.) Cell Biology:A Laboratory Handbook, Academic Press, San Diego 1998,Vol. IV, pp. 375±386.

[18] Celis, J. E., éstergaard, M., Jensen, N. A., Gromova, I.,Rasmussen, H. H., Gromov, P., FEBS Lett. 1998, 430,64±72.

[19] Azuma, T., Witke, W., Stossel, T. P., Hartwig, J. H.,Kwiatkowski, D. J., EMBO J. 1998, 17, 1362±1370.

[20] Quinlan, R., Hutchison, C., Lane, B., Protein Profile 1994, 1,779±991.

[21] Quinlan, R., Hutchison, C., Lane, B., Protein Profile 1995, 2,795±952.

[22] Ku, N. O., Liao, J., Omary, M. B., EMBO J. 1998, 17, 1892±1906.

[23] Leffers, H., Madsen, P., Rasmussen, H. H., HonorØ, B.,Andersen, A. H., Walbum, E., Vandekerckhove, J., Celis,J. E., J. Mol. Biol. 1993, 231, 982±998.

[24] Hermeking, H., Lengauer, C., Polyak, K., He, T. C., Zhang,L., Thiagalingam, S., Kinzler, K. W., Vogelstein, B., Mol. Cell1997, 1, 3±11.

Electrophoresis 1999, 20, 355±361 Proteome profiles of cultured bladder TCCs 361