Proteomic Strategies to Reveal Tumor Heterogeneity among Urothelial Papillomas *

Julio E. Celis{ddagger},§,, Pamela Celis§, Hildur Palsdottir§,||, Morten Østergaard§, Pavel Gromov{ddagger},§, Hanne Primdahl**, Torben F. Ørntoft**, Hans Wolf{ddagger}{ddagger}, Ariana Celis{dagger},§ and Irina Gromova{ddagger},§

{ddagger} Institute of Cancer Biology and Danish Centre for Human Genome Research, The Danish Cancer Society, DK-2100 Copenhagen
§ Department of Medical Biochemistry, The University of Aarhus, Ole Worms Allé, Building 170, DK-8000 Aarhus
** Departments of Clinical Biochemistry
{ddagger}{ddagger} Urology, Skejby Hospital, DK-8200 Aarhus N, Denmark


   ABSTRACT
TOP
 ABSTRACT
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Proteomics and immunohistochemistry were used to reveal tumor heterogeneity among urothelial papillomas (UPs) with the long term goal of predicting their biological potential in terms of outcome. First, we identified proteins that were deregulated in invasive fresh lesions as compared with normal urothelium, and thereafter we immunostained UPs with a panel of antibodies against some of the markers. Twenty-two major proteins showing variations of 2-fold or more in at least one-third of the invasive lesions were selected. Specific antibodies against several of the proteins were obtained, but only a few reacted positively in immunostaining. A panel consisting of antibodies against keratinocytes (CKs) 5, 13, 18, and 20 and markers of squamous metaplasia (CKs 7, 8, and 14) was used to probe normal urothelium and 30 UPs collected during a period of five years. Four UPs showed a normal phenotype, whereas the rest could be grouped in five major types that shared aberrant staining with the CK20 antibody. Type 1 heterogeneity (n = 4) showed preferred staining of the umbrella cells with the CK8 antibody. Type 2 (n = 11) was typified by the staining of the basal and intermediate layers with the CK20 antibody. Type 3 (n = 7) was characterized by the predominant staining of the basal cell layer with the CK5 antibody. Type 4 (n = 1) showed areas of CK7 negative cells, whereas type 5 (n = 3) showed loss of staining of the basal cells with the CK20. 29% of the patients experienced recurrences, but none progressed to invasive disease. Patients harboring phenotypic alterations in the basal cell compartment (types 3 and 5) showed the highest number of recurrences (4/7 and 2/3, respectively), and all type 3 lesions progressed to a higher degree of dedifferentiation. Even though a long term prospective study involving a larger sample size is required to assess the biological potential of these lesions, we believe that this approach will prove instrumental for revealing early phenotypic changes in different types of cancer.

Bladder cancer is one of the most prevalent type of cancer in Denmark, with 11,509 patients in 1995 and more than 1000 new incidents every year (1). It comprises a broad spectrum of histological heterogeneous tumor types arising predominantly in the epithelium (urothelium) lining of the urinary bladder and the ureters. The urothelium is a multilayer epithelium containing polygonal basal cells that proliferate and differentiate to form the large umbrella-shaped cells of the superficial layer. The latter exhibit a unique trilaminar rigid membrane, termed the asymmetric unit membrane, that is in direct contact with the urine (2).

Tumor types of the urothelium include transitional cell carcinomas (TCCs), 1 squamous cell carcinomas, and adenocarcinomas, as well as other less frequent lesions (3). TCCs comprise more than 90% of the diagnosed bladder tumors in the Western hemisphere and are divided broadly into papillary (growing into the lumen of the bladder) and flat lesions that arise and progress through different genetic alterations (4). Bladder tumors are diagnosed routinely by cytoscopy, which give important information on multifocality, appearance, and size (5, 6). Cytoscopy is essential for resection and provides valuable material for pathological observation and research purposes. At present, pathological staging and grading are the most significant prognostic factors for treatment and outcome.

The biological relevance of low grade papillary tumors is underlined by the fact that about 70% of the bladder lesions are diagnosed as superficial (Ta, T1) at first presentation. These tumors have a high frequency of recurrence (>60%), and about 10–15% of them will progress to life-threatening malignancies (7, 8). Currently, it is not possible to assess with certainty the biological behavior of these tumors based on clinical or morphological criteria alone, and as a result it is urgent to identify early and accurate biomarkers that may predict recurrence, progression, and response to treatment. Moreover, it is important to distinguish those lesions that have no significant effect on the life expectancy, as all tumor-bearing patients are diagnosed with cancer, a fact that has practical and economic implications, as well as a profound psychological effect on the patient (9). To date, several prognostic markers have been identified; these include FDA-approved biomarkers (NMP22, fibrin/fibrinogen degradation product, and basement membrane components), blood group-related antigens (ABH, Lewis antigen), tumor-associated antigens (M344, 19A211, T138, DD23), proliferation antigens (Ki67 antibody, PCNA), oncogenes (c-Erb B2, Ras, c-Myc, mdm2), growth factors (epidermal growth factor, transforming growth factor-ß, fibroblast growth factor, vascular endothelial growth factor), adhesion molecules (cadherins, integrins), cytokeratins (keratin 20), and cell cycle regulatory proteins such as p53, pRb, cyclins, p15, p16, and p21 (10–12). However, with very few exceptions these markers do not predict the biological behavior of low grade lesions. Today, it is becoming increasingly clear that multiple, rather than single markers, may be required to predict accurately prognosis and response to treatment.

In our laboratories we are applying gene expression profiling technologies in combination with immunohistochemistry to reveal bladder cancer heterogeneity with the long term aim of predicting the biological behavior of these lesions in terms of recurrence and progression (13–15). An important milestone of these studies has been the establishment of proteomic databases of various bladder tissue compartments (urothelium, TCCs, squamous cell carcinomas, muscle, connective tissue), as well as of urine and plasma (biobase.dk/cgi-bin/celis; see Refs. 16 and 17). The strategy for identifying tumor heterogeneity (18) encompasses a blind and systematic study of the proteome expression profiles of hundreds of fresh biopsy specimens from both normal and tumor origin. First, we identify major proteins that are expressed differentially in invasive lesions as compared with normal urothelium, and thereafter we use specific antibodies against these proteins to immunostain cryostat sections of tumors diagnosed as having the same stage and grade of atypia. Here we report on the analysis of 30 fresh Gr I, Ta, UPs collected at Skejby Hospital in Aarhus over a period of 5 years.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Tumors—
Tumors and other biopsies removed at Skejby Hospital, Aarhus, Denmark (informed consent was obtained from the patient) over a period of 5 years were placed immediately on ice and transported to the laboratory. Tumor pieces for cryostat sections were kept at -80°C. Tumors were classified according to Bergkvist et al. (19). In this classification, Gr I, Ta lesions are defined as papillomas, that is a papillary tumor with a thin fibrovascular stroma and a normal looking urothelium. This corresponds to the World Health Organization group papillary lesions of low malignant potential. The Scientific and Ethical Committee of Aarhus County approved the project.

Labeling of Tumors with [35S]Methionine—
Tumors clean of clots and contaminating tissue were minced in small pieces with the aid of a scalpel and were labeled with [35S]methionine for 14–16 h in a 10-ml sterile plastic conical tube containing 0.2 ml of modified Eagle’s medium lacking methionine, 2% dialyzed (against 0.95% NaCl) fetal calf serum, and 100 µCi of [35S]methionine (SJ204; Amersham Biosciences). At the end of the labeling period the medium was aspirated, and the pieces were dissolved in 0.3 to 0.4 ml of lysis solution with the aid of a 1-ml plastic pipette (20). Samples were stored at -20°C until use.

Labeling of Normal Random Biopsies with [35S]Methionine—
The urothelium from random biopsies diagnosed as normal was dissected with the aid of a scalpel, labeled with [35S]methionine as described above, and processed for 2D-PAGE (21). Samples were stored at -20°C until use.

Two-dimensional PAGE—
2D-PAGE was performed essentially as described by Celis et al. (21). Gels were stained with silver nitrate (22), dried and exposed to x-ray films as described previously (21).

Protein Identification—
Proteins were identified using a combination of procedures that included mass spectrometry, namely matrix-assisted laser desorption ionization time-of-flight, Biflex (Bruker) (22), 2D-PAGE Western immunoblotting (23), and comparison with the master 2D gel images of human keratinocytes and TCC proteins (24) (http:/biobase.dk/cgi-bin/celis).

Antibodies—
Rabbit polyclonal antibodies against CK5, CK13, A-FABP, PA-FABP, maspin, 14–3-3{varsigma}, and hsp28. MRP14 and PCNA were prepared in our laboratory. Monoclonal antibodies against CKs 7, 8, and 18 were purchased from ICCN/Cappel. The monoclonal antibody against CK20 was purchased from Cymbus, and the CK14 antibody was purchased from AH Diagnostics. All antibodies were screened carefully for specificity by 2D-PAGE Western immunoblotting (IEF, non-equilibrium pH gradient electrophoresis) of different cell types and human tissues, including TCCs (23). In addition, they were screened using a battery of human tissues that either expressed or lacked the protein in question.

Indirect Immunofluorescence—
Tumors and random biopsies frozen in liquid nitrogen were sectioned on a Reichert-Jung cryostat. 8-µm sections placed in coverslips were washed three times with Hanks’-buffered saline solution (HBSS) and treated for 10 min with 3.6% formaldehyde (18). After washing extensively with HBSS, the coverslips were covered with 20 µl of the primary antibody and incubated for 45 min at 37°C in a humid environment. The coverslips were washed several times with HBSS and covered with 20 µl of rhodamine-conjugated secondary antibody (diluted 1:50 in HBSS). After 45 min of incubation at 37°C in a humid environment, the coverslips were washed thoroughly with HBSS and mounted in fluorescence mounting medium (DAKO). Observations were made on a Leica photomicroscope equipped with epifluorescence.


   RESULTS
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Identification of Major Proteins Highly Deregulated in Invasive TCCs
Fresh tumor and normal biopsies collected during a period of 5 years were analyzed by 2D-PAGE and autoradiography as described under "Materials and Methods." Biopsies were labeled with [35S]methionine as this improved by at least one order of magnitude the number of polypeptides detected as compared with silver staining, and in addition, avoided the problem associated with the presence of serum proteins, particularly in the case of invasive tumors. Of the several hundred specimens that were received, 122 corresponded to Gr III, T2–4 TCCs at first presentation. These tumors were the most numerous among the invasive lesions and were chosen to generate the markers as they were expected to harbor proteome alterations associated with recurrence and progression.

Only biopsy specimens from invasive tumors exhibiting minor contamination with connective and/or muscle tissue as judged by the expression of vimentin and desmin using immunostaining and 2D gel analysis (13) and that yielded fair separation profiles (30 lesions) were selected for comparison with normal urothelium. A normalized base line of protein expression of "normal" urothelium was attained by visual analysis of 116 random biopsies classified as normal by the pathologist irrespective of whether they were derived from patients bearing low grade or invasive TCCs. Only proteins whose expression was highly deregulated (2-fold or more) in the invasive lesions (see Fig. 1 and Table I) were used as reference. These included A-FABP, PS-FABP, PCNA, CK13, 14–3-3{varsigma}, and MRP14. With the exception of A-FABP and MRP14, which showed deregulation in 20.6 and 14.7% of the random biopsies, respectively, all other markers showed fluctuations in less than 3% of the biopsies. It should be stressed that in the case of A-FABP, more than 50% of the random biopsies showing strong down-regulation were derived from patients bearing invasive tumors.



View larger version (81K):
[in this window]
[in a new window]
  		FIG. 1. Proteins highly deregulated in invasive TCCs. [35S]Methionine-labeled proteins synthesized by normal urothelium and an invasive TCC (Gr III, T2–4) were separated by 2D-PAGE (IEF) and visualized by autoradiography. Polypeptides indicated with red are up-regulated in the tumor whereas those indicated in blue are down-regulated. The position of actin is indicated for reference.

 

View this table:
[in this window]
[in a new window]
  		TABLE I Polypeptides deregulated in invasive TCCs

 
Fig. 1 shows representative proteome expression profiles of fresh biopsies from normal urothelium and an invasive TCC (Gr III, T2–4). Even though the overall expression profiles are remarkably similar, the level of several major polypeptides was altered highly in the invasive lesions. Deregulated polypeptides that showed variations of 2-fold or more in at least 40% of the invasive lesions, as determined by visual inspection of the 2D gels, were selected. The cutoff value of 33% was chosen to include CK20 as the expression of this differentiation marker has been shown to be aberrant in recurrent papillary tumors (9). In reality, however, most of the selected markers were deregulated in more than 50% of the lesions (Table I). Deregulated polypeptides are indicated in red (up-regulated) and blue (down-regulated) in Fig. 1. Their coordinates (apparent Mr and pI) in the TCC database, the frequency at which the change occurred in the invasive lesions, and the putative functions are listed in Table I. Up-regulated proteins included elongation factor 1{gamma}, hsp28, CKs 18 and 20, PCNA, the proteasome {zeta} subunit, and MRP14. Down-regulated proteins comprised the A-FABP, annexin V, the high affinity bile-binding protein (DD2), CKs 5 and 13, lactate dehydrogenase H chain, maspin, plasminogen activator inhibitor 2, PA-FABP, phosphoglycerate kinase, purine nucleoside phosphorylase, a tropomyosin isoform, 14–3-3{varsigma}, (stratifin), and two unknown proteins recorded in the TCC 2D-PAGE database. Changes in the levels of CK5, hsp28, and MRP14 are not observed in the biopsy pairs depicted in Fig. 1, but representative areas of gels from other normal and invasive tumor specimens are shown in Fig. 2.



View larger version (85K):
[in this window]
[in a new window]
  		FIG. 2. Levels of CK5, hsp28, and MRP14 in normal urothelium and invasive TCCs. [35S]Methionine-labeled proteins synthesized by normal urothelium and invasive TCCs (Gr III, T2–4) were separated by 2D-PAGE (IEF) and visualized by autoradiography. A, C, and E, normal urothelium; B, D, and F, invasive lesions. Only a fraction of the IEF 2D-gels are shown.

 
Antibodies against the Deregulated Protein Markers
Antibodies against 13 of the putative protein markers were raised in rabbits (hsp28, PCNA, MRP14, A-FABP, PA-FABP, maspin, 14–3-3{varsigma}, and CKs 5 and 13), purchased from commercial sources (CKs 18 and 20), or procured from colleagues (annexin V, plasminogen activator inhibitor 2). With the exception of two antibodies that showed slight cross-reactivity with other family members (14–3-3{varsigma} and CK5; see Fig. 3, F and G), the rest were highly specific as judged by 2D-PAGE immunoblotting using both isoelectrofocusing (see Fig. 3; blots of PA-FABP and annexin V antibodies are not shown) and non-equilibrium pH gradient electrophoresis (see "Materials and Methods"). Differential expression of the markers was confirmed by 2D-PAGE immunoblotting using representative pairs of normal and invasive lesions (not shown). Only antibodies against CKs 5, 13, 18, and 20 reacted positively with methanol or formaldehyde-fixed cryostat sections and were included in the battery of antibodies to immunostain UPs.



View larger version (34K):
[in this window]
[in a new window]
  		FIG. 3. Antibody specificity as determined by 2D-PAGE Western immunoblotting (IEF) of total cellular proteins from normal human keratinocytes (A–F) and TCCs (G–N). Blots were developed using the enhanced chemiluminescence procedure.

 
Identification of Tumor Heterogeneity among Gr I, Ta Urothelial Papillomas
The study group consisted of 30 patients that were diagnosed with Gr I, Ta UPs at first presentation over a period of 5 years. Clinical features, as well as recurrences, are given in Table II. Patient age at diagnosis ranged from 27 to 88 years old, and the male to female ratio was 2:1.


View this table:
[in this window]
[in a new window]
  		TABLE II Urotelial papillomas (Gr I, Ta)

 
Frozen cryostat sections of the 30 UPs fixed in formaldehyde were immunostained with the panel of antibodies described above. The battery was supplemented with antibodies specific for CKs 7, 8, and 14 (Fig. 3), and we and others (18, 25–27) have shown that these proteins are markers for squamous differentiation, a rather common feature of invasive lesions (26, 27). 2 Indeed, the levels of these proteins were deregulated in a significant number of invasive lesions (results not shown).

With the exception of four UPs that displayed identical staining patterns as normal urothelium, the rest could be grouped in five major types that, with the exception of type 1, shared aberrant staining with the CK20 antibody, a predictor of tumor recurrence (9). UPs exhibited only one major type of tumor heterogeneity, but the extent of the heterogeneity was variable and in most cases included variants.

The staining phenotypes of normal urothelium reacted with the panel of antibodies are presented in Fig. 4. Antibodies against CKs 20 (Fig. 4A) and 18 (Fig. 4B) stained the superficial umbrella cells, whereas the probes against CKs 5 (Fig. 4C), 7 (Fig. 4D), and 8 (Fig 4E) reacted with basal, intermediate, and superficial layers in a homogenous fashion. No staining of the umbrella cells was observed with the CK13 antibody, which stained the basal and intermediate layers (Fig. 4F) (18, 27), and as expected, none of the cell layers reacted with the CK14 antibody (Fig. 4G). For reference, a hematoxylin staining of normal urothelium is included in Fig. 4H. Because of space limitations only major staining differences between the subtypes and normal urothelium are highlighted below (see also Table III).



View larger version (39K):
[in this window]
[in a new window]
  		FIG. 4. Immunofluorescence staining of formaldehyde-fixed cryostat sections from normal urothelium reacted with the panel of antibodies against keratins (A–G) and hematoxylin staining of normal urothelium (H).

 

View this table:
[in this window]
[in a new window]
  		TABLE III Antigen expression in normal urothelium and urothelial papillomas

For simplicity, only staining patterns that differ from those observed in normal urothelium are indicated.

 
Type 1 heterogeneity was found in four UPs and was characterized by the predominant staining of the umbrella cells with the CK8 antibody (Fig. 5, A, compare with CK18 staining in B). All urothelial cell layers stained throughout with the CK20 antibody (not shown), although the umbrella cells stained stronger. 2D-PAGE analysis of total protein extracts from these tumors suggested that the increased staining of the superficial cells with the CK8 antibody is most likely not because of increased synthesis of the protein or because of major post-translational modification(s) that affect its charge (compare Fig. 6, A and B).



View larger version (103K):
[in this window]
[in a new window]
  		FIG. 5. Tumor heterogeneity types. Immunofluorescence staining of formaldehyde-fixed cryostat sections from selected UPs showing various types of tumor heterogeneity is shown.

 


View larger version (63K):
[in this window]
[in a new window]
  		FIG. 6. Expression of CKs 8, 14, and 7 in Gr I, Ta in selected UPs. Proteins were separated by 2D-PAGE (IEF). A and B, silver-stained proteins from UPs 521–1 (type 1) and 916–1 (type 2A), respectively. C and D, [35S]methionine-labeled proteins from UP 1250–1 (type 3C) and UP 841–1 (subtype 4), respectively. Only relevant areas of the 2D gels are shown.

 
Type 2 heterogeneity was observed in 11 UPs and was typified by the abnormal expression of CK20 in the basal and intermediate layers (Fig 5C). Only very few umbrella cells stained with this antibody. Variants within this type included tumors that exhibited either scattered cells expressing CK18 in non-umbrella cells (type 2B, Fig. 5D) or areas with CK5 negative cells (type 2C, not shown).

Type 3 heterogeneity was detected in seven UPs and was characterized by the preferred staining of the basal cell layer with the CK5 antibody, as well as by the abnormal staining with the CK20 antibody (type 3A; see Fig. 5, E and F). In some tumors the staining was accompanied by scattered cells expressing CK18 (type 3B; not shown) and, in addition, by basal expression of CK14. The expression of CK14 was confirmed by 2D-PAGE (Fig. 6C).

Type 4 heterogeneity was detected in one UP that showed large areas of CK7 negative cells (Fig. 5G), a fact that was confirmed by 2D-PAGE analysis of total protein extracts (Fig. 6D). The lesion stained with the CK8 antibody but not with the CK14 probe. Even though the pattern was observed in only one tumor we classified it in a separate group as loss of CK7 staining was observed in a significant number of invasive TCCs, as well as in some areas of the urothelium of patients with invasive disease.

Type 5 heterogeneity was observed in three UPs and was characterized by the abnormal expression of CK20, in particular by the lack of staining of the basal cell layer (Fig. 5I). One of the UPs displayed, in addition, CK14 in the basal layer (type 5B; see Fig. 5J).

Twenty-nine percent of the patients experienced recurrences during the course of the study, but to date none progressed to invasive disease. Patients with UPs harboring types 3 and 5 heterogeneities showed the highest number of recurrences, 4/7 and 2/3, respectively. Interestingly, all recurrent type 3 lesions progressed to a higher degree of dedifferentiation resulting in a higher histological grade (Gr II, Ta). No recurrences have been observed so far for the UPs that exhibited normal immunostaining with the panel of antibodies. It should be stressed, however, that some patients were recruited late in the course of the study, and as a result the recurrence data are limited.


   DISCUSSION
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
To be able to predict which patients with UPs will develop recurrences and/or progress to invasive disease is one of the major challenges that we face today in bladder cancer research. Morphological parameters alone are not able to discern the biological potential and behavior of these tumors, and hence it is imperative to identify markers that may predict patient outcome. Recently, Southgate and colleagues (9) showed that papillary lesions displaying normal distribution of CK20 may not recur, whereas those exhibiting aberrant expression of this keratin, that is expression in all cell layers, had a higher frequency of recurrence. These results have been challenged recently by Alsheikh et al. (28), who compared the WHO/ISUP classification (29, 30) and CK20 expression in predicting the behavior of low grade papillary urothelial tumors. In this study, papillary neoplasms of low malignant potential had a lower recurrence rate than low grade papillary TCCs, and low grade urothelial tumors exhibiting normal CK20 expression pattern recurred less frequently than lesions with an aberrant pattern of CK20 staining. Neither of these differences was significant statistically, however, and recurrences were observed in 20% of patients whose tumors were both classified as papillary neoplasms of low malignant potential and showed normal CK20 immunostaining. The results presented here cannot be compared directly with those described above, because the number of tumors showing normal patterns of CK20 was small, and the recurrence data were sparse. In general, however, they support the contention that aberrant immunostaining of CK20 is an important phenotypic characteristic associated with some non-invasive papillary lesions.

Our approach for identifying tumor heterogeneity among UPs relied on a combination of proteomic technologies and immunohistochemistry and encompassed a blinded and systematic study of the proteome expression profiles of hundreds of fresh biopsy specimens of both normal and tumor origin. Invasive TCCs (Gr III, T2–4) rather than UPs were chosen to generate the putative protein markers as these lesions were expected to harbor proteome alterations associated with recurrence and progression that may be impossible to identify at an early stage. We surmised that areas of urothelial heterogeneity may comprise only a small proportion of the UPs, and as a result neither down- nor up-regulated protein markers may be detected under these conditions.

Although the size of the prospective study group was limited by the necessity of using fresh biopsy material, we were able to identify several types of tumor heterogeneity among 30 patients diagnosed with Gr I, Ta UPs at first presentation over a period of 5 years. The nature of the heterogeneity was defined utilizing only a limited number of proteins whose expression was deregulated by 2-fold or more and for which we were able to obtain specific antibodies that could be used with confidence in immunohistochemistry. This panel was supplemented with antibodies against markers of squamous metaplasia (CKs 7, 8, and 14) as these lesions are a rather common feature of invasive TCCs (26, 27). 3 Low abundance proteins were not targeted here as this study was intended as a proof of concept rather than as a comprehensive analysis of all possible protein markers. Certainly, future work must take into consideration low abundance components as tumor cells may express 6000–7000 primary translation products plus their post-transitionally modified variants (31), many of which may be expressed at levels that escape detection with the current technology. Identification of differentially expressed proteins among low abundance polypeptides will not be an easy task as it will require the use of metabolic labeling with [35S]methionine, enrichment techniques (microdissection (32, 33), subcellular fractionation (34), and purification of specific complexes (35)), as well as development of reliable, high throughput quantitation procedures. Even if the latter requirements are overcome, we still need to address the problems associated with raising specific antibodies as the results presented here have clearly highlighted the limitations we face today regarding preparation, specificity, and reactivity of antibodies. Usually, one searches for heterogeneity that is evident only in very few cells, and as a result one must be confident that these cells express the antigen in question rather than an unknown cross-reactivity.

Of the 30 UPs analyzed by immunofluorescence using the panel of antibodies only four displayed identical staining patterns as normal urothelium. The rest could be grouped in five major types that, with the exception of type 1, shared aberrant staining with the CK20 antibody. As expected, the size of the areas showing heterogeneity varied from UP to UP, but most tumors exhibited only one major type of tumor heterogeneity. Main features of the different types of tumor heterogeneity are discussed below.

Type 1 Heterogeneity
This type of heterogeneity (n = 4) was characterized by the stronger expression of CK8 in the superficial umbrella cells. This phenotype was not accompanied by changes in the expression levels of primary translation product or their modified variants, suggesting that the increased immunoreactivity of the umbrella cells is most likely because of epitope unmasking/masking phenomena that may be related to filament organization (26, 36). Differential staining of the urothelium has been observed using different CK8 antibodies (27), but it has not been reported in the case of paired tumor biopsies that have been stained with the same antibody. It should be emphasized that the CK8 antibody was included in the panel mainly because the levels of this protein decrease in squamous metaplasia (18). Evidently, type 1 heterogeneity is not related to squamous differentiation as none of the four UPs expressed CK14 (also confirmed by 2D gel profiling), a marker that can distinguish squamous differentiation even before it becomes evident by morphological criteria. In addition, we observed normal staining with the CK7 antibody in all the lesions showing this type of heterogeneity. Recently, we reported changes in the levels of CK8 in some invasive TCCs because of protein degradation (14), and we cannot exclude the possibility that this mechanism may play a role in the generation of type 3 heterogeneity.

Type 2 Heterogeneity
This type of heterogeneity was the most common (n = 11) and was characterized by the expression of CK20 in the basal and suprabasal layers as originally described by Harnden et al. (9). The levels of CK20 could not be determined by 2D-PAGE as the protein migrated very closely to a major labeled unknown protein that is down-regulated in invasive lesions (unknown protein close to CK20 in Fig. 1). Variants within type 2 heterogeneity included lesions expressing 1) focus of CK18 positive non-superficial cells, and 2) areas with CK5 negative staining. The latter changes could not be confirmed by 2D-PAGE analysis of UPs as the number of cells showing the phenotype was too small.

Type 3 Heterogeneity
This type of lesion (n = 7) was characterized by the stronger immunostaining of the basal cell layer with the CK5 polyclonal antibody. The antibody showed a weak cross-reactivity with CKs 7 and 8, but this did not interfere with the staining patterns. There is data in the literature, however, indicating that CK5 is expressed basically in normal urothelium as judge by immunostaining using monoclonal antibody 34ßE12 (37). As judged by 2D-PAGE Western immunoblotting this antibody also exhibits a wider cross-reactivity with CKs 14 and 19 (results not shown).

Type 4 Heterogeneity
This type of heterogeneity was observed in only one UP that showed extensive areas with CK7 negative cells, a result that was confirmed by 2D gel analysis of protein extracts prepared from the same tumor. The lesion did not express CK14 and showed normal staining and expression of CK8 suggesting that its origin is most likely not connected to squamous differentiation.

Type 5 Heterogeneity
Even though aberrant expression of CK20 has been reported earlier in papillary tumors (9), the absence of staining of the basal cell layer with CK20 antibodies has not been reported. The characteristic future of type 5 heterogeneity is the absence of CK20 staining in the basal cell layer. In addition, one of the lesions revealing this type of heterogeneity expressed CK14 in the basal layer that was confirmed by 2D gel analysis. The biological potential of this type of heterogeneity is at present unclear, although the expression of CK14 may reveal the early events in squamous differentiation. The origin of the type 5 heterogeneity is at present unclear, although the one expressing CK14 may be undergoing early events in squamous differentiation.

Currently, we are carrying out a systematic analysis of low grade non-invasive and invasive TCCs with the panel of antibodies in an effort to determine the incidence of the various lesions described above at various stages of cancer development. In addition, we are contemplating the idea of using laser capture microdissection to obtain distinct populations of cells exhibiting particular types of heterogeneities for proteomic and microarrays analysis (15, 32, 33). The aim of these studies is to identify additional markers that may prove instrumental for forthcoming studies.

Presently, it is not possible to comment on the correlation, if any, between the presence of a given type of heterogeneity and the rate of recurrence or progression to invasive disease as the number of patients was relatively small, and the follow-up was too short. Twenty-nine percent of the patients experienced recurrences during the course of the study, a number that is very close to those reported in previous studies (8). Interestingly, patients harboring phenotypic alterations in the basal cell compartment (types 3 and 5) showed the highest number of recurrences (4/7 and 2/3, respectively), and all type 3 lesions progressed to a higher degree of dedifferentiation. Evaluation of the biological potential of these and other lesions characterized in this study will require a much larger sample size, as well as a follow-up of up to 10 years. In addition, these studies must take into consideration the biological significance of having combinations of various types of tumor heterogeneities as invasive lesions are known to exhibit areas with different histological grades (38).

Finally, we would like to stress that the approach presented here for revealing tumor heterogeneity has already been used successfully to identify metaplastic and premalignant lesions in bladder squamous cell carcinomas using immunowalking (18). Its full potential, however, would only be realized when a comprehensive set of antibodies against marker proteins is used in combination with tissue arrays (39) in a high throughput fashion.


   FOOTNOTES
 
Received, November 23, 2001, and in revised form, March 21, 2002.

Published, MCP Papers in Press, March 21, 2002, DOI 10.1074/mcp.M100031-MCP200

1 The abbreviations used are: TCC, transitional cell carcinoma; IEF, isoelectric focusing; UP, urothelium papilloma; 2D, two-dimensional; A-FABP, adipocyte fatty acid-binding protein; PA-FABP, psoriatic-associated protein FABP; PCNA, proliferating cell nuclear antigen; HBSS, Hanks’-buffered saline solution; CK, keratinocytes. Back

2 Unpublished results. Back

3 Unpublished observations. Back

* This work was supported in part by grants from the Danish Cancer Society and the Danish Biotechnology Program and by fellowships from the Danish Cancer Society (to M. Ø. and H. P.). Back

{dagger} Deceased July 7, 1998. Back

|| Present address: Dept. of Molecular Membrane Biology, Heinrich Hoffmann Strasse 7, 60528 Frankfurt/Main, Germany. Back

To whom correspondence should be addressed: Inst. of Cancer Biology and Danish Centre for Human Genome Research, The Danish Cancer Society, Strandboulevarden 49, 2100 Copenhagen, Denmark. Tel.: 45-35-25-73-63; Fax: 45-35-25-73-75; E-mail: jec{at}cancer.dk


   REFERENCES
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Cancerregistret, 1999, Danish Cancer Society: www.cancer.dk/tal

  2. Wu, X. R., Lin, J. H., Walz, T., Haner, M., Yu, J., Aebi, U., and Sun, T. T. (1994) Mammalian uroplakins. A group of highly conserved urothelial differentiation-related membrane proteins. J. Biol. Chem. 269, 13716 –13724[Abstract/Free Full Text]

  3. Bane, B. L., Rao, J. Y., and Hemstreet, G. P. (1996) Pathology and staging of bladder cancer. Semin. Oncol. 23, 546 –570[Medline]

  4. Spruck, C. H., III, Ohneseit, P. F., Gonzalez-Zulueta, M., Esrig, D., Miyao, N., Tsai, Y. C., Lerner, S. P., Schmutte, C., Yang, A. S., and Cote, R. (1994) Two molecular pathways to transitional cell carcinoma of the bladder. Cancer Res. 54, 784 –788[Abstract]

  5. Murphy, W. M., Soloway, M. S., Jukkola, A. F., Crabtree, W. N., and Ford, K. S. (1984) Urinary cytology and bladder cancer. The cellular features of transitional cell neoplasms. Cancer 53, 1555 –1565[Medline]

  6. Grossman, H. B. (1998) New methods for detection of bladder cancer. Semin. Urol. Oncol. 16, 17 –22[Medline]

  7. Zieger, K., Wolf, H., Olsen, P. R., and Hojgaard, K. (2000) Long-term follow-up of noninvasive bladder tumours (stage Ta): recurrence and progression. BJU Int. 85, 824 –828[CrossRef][Medline]

  8. Cheng, L., Neumann, R. M., and Bostwick, D. G. (1999) Papillary urothelial neoplasms of low malignant potential. Clinical and biologic implications. Cancer 86, 2102 –2108[CrossRef][Medline]

  9. Harnden, P., Mahmood, N., and Southgate, J. (1999) Expression of cytokeratin 20 redefines urothelial papillomas of the bladder. Lancet 353, 974 –977[CrossRef][Medline]

  10. Kausch, I., and Bohle, A. (2001) Bladder cancer. II. Molecular aspects and diagnosis. Eur. Urol. 39, 498 –506[CrossRef][Medline]

  11. Adshead, J. M., Kessling, A. M., and Ogden, C. W. (1998) Genetic initiation, progression and prognostic markers in transitional cell carcinoma of the bladder: a summary of the structural and transcriptional changes, and the role of developmental genes. Br. J. Urol. 82, 503 –512[CrossRef][Medline]

  12. van der Poel, H. G., and Debruyne, F. M. (2001) Can biological markers replace cystoscopy? An update. Curr. Opin. Urol. 11, 503 –509[CrossRef][Medline]

  13. Celis, J. E., Ostergaard, M., Basse, B., Celis, A., Lauridsen, J. B., Ratz, G. P., Andersen, I., Hein, B., Wolf, H., Orntoft, T. F., and Rasmussen, H. H. (1996) Loss of adipocyte-type fatty acid binding protein and other protein biomarkers is associated with progression of human bladder transitional cell carcinomas. Cancer Res. 56, 4782 –4790[Abstract]

  14. Celis, J. E., Kruhoffer, M., Gromova, I., Frederiksen, C., Ostergaard, M., Thykjaer, T., Gromov, P., Yu, J., Palsdottir, H., Magnusson, N., and Orntoft, T. F. (2000) Gene expression profiling: monitoring transcription and translation products using DNA microarrays and proteomics. FEBS Lett. 480, 2 –16[CrossRef][Medline]

  15. Orntoft, T. F., Thykjaer, T., Waldman, F. M., Wolf, H., and Celis, J. E. (2002) Genome-wide study of gene copy numbers, transcripts, and protein levels in pairs of non-invasive and invasive human transitional cell carcinomas. Mol. Cell. Proteomics 1, 37 –45[Abstract/Free Full Text]

  16. Rasmussen, H. H., Orntoft, T. F., Wolf, H., and Celis, J. E. (1996) Towards a comprehensive database of proteins from the urine of patients with bladder cancer. J. Urol. 155, 2113 –2119[Medline]

  17. Celis, J. E., Ostergaard, 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., Orntoft, T. F., Celis, P., and Celis, A. (1999) A comprehensive protein resource for the study of bladder cancer: http://biobase.dk/cgi-bin/celis. Electrophoresis 20, 300 –309[CrossRef][Medline]

  18. Celis, J. E., Celis, P., Ostergaard, M., Basse, B., Lauridsen, J. B., Ratz, G., Rasmussen, H. H., Orntoft, T. F., Hein, B., Wolf, H., and Celis, A. (1999) Proteomics and immunohistochemistry define some of the steps involved in the squamous differentiation of the bladder transitional epithelium: a novel strategy for identifying metaplastic lesions. Cancer Res. 59, 3003 –3009[Abstract/Free Full Text]

  19. Bergkvist, A., Ljungqvist, A., and Moberger, G. (1965) Classification of bladder tumours based on the cellular pattern. Preliminary report of a clinical-pathological study of 300 cases with a minimum follow-up of eight years. Acta Chir. Scand. 130, 371 –378[Medline]

  20. O’Farrell, P. H. (1975) High resolution two-dimensional gel electrophoresis. J. Biol. Chem. 250, 4007 –4021[Abstract]

  21. Celis, J. E., Ratz, G., Basse, B., Lauridsen, J. B., Celis, A., Jensen, N. A., and Gromov, P. (1997) in Cell Biology: A Laboratory Handbook (Celis, J. E., Carter, N. P., Shotton, D. M., Hunter, T., and Simons, K., eds) 4th Ed., pp.375 –385, Academic Press, San Diego

  22. Shevchenko, A., Wilm, M., Vorm, O., and Mann, M. (1996) Mass spectrometric sequencing of proteins silver-stained polyacrylamide gels. Anal. Chem. 68, 850 –858[CrossRef][Medline]

  23. Celis, J. E., Lauridsen, J. B., and Basse, B. (1997) in Cell Biology: A Laboratory Handbook (Celis, J. E., Carter, N. P., Shotton, D. M., Hunter, T., and Simons, K., eds) 4th Ed., pp.429 –437, Academic Press, San Diego

  24. Celis, J. E., Østergaard, M., Jensen, N. A., Gromova, I., Rasmussen, H. H., and Gromov, P. (1998) Human and mouse proteomic databases: novel resources in the protein universe. FEBS Lett. 430, 64 –72[CrossRef][Medline]

  25. Moll, R., Achtstatter, T., Becht, E., Balcarova-Stander, J., Ittensohn, M., and Franke, W. W. (1988) Cytokeratins in normal and malignant transitional epithelium. Maintenance of expression of urothelial differentiation features in transitional cell carcinomas and bladder carcinoma cell culture lines. Am. J. Pathol. 132, 123 –144[Abstract]

  26. Harnden, P., and Southgate, J. (1997) Cytokeratin 14 as a marker of squamous differentiation in transitional cell carcinomas. J. Clin. Pathol. 50, 1032 –1033[Abstract]

  27. Southgate, J., Harnden, P., and Trejdosiewicz, L. K. (1999) Cytokeratin expression patterns in normal and malignant urothelium: a review of the biological and diagnostic implications. Histol. Histopathol. 14, 657 –664[Medline]

  28. Alsheikh, A., Mohamedali, Z., Jones, E., Masterson, J., and Gilks, C. B. (2001) Comparison of the WHO/ISUP classification and cytokeratin 20 expression in predicting the behavior of low-grade papillary urothelial tumors. World/Health Organization/International Society of Urologic Pathology. Mod. Pathol. 14, 267 –272[CrossRef][Medline]

  29. Epstein, J. I., Amin, M. B., Reuter, V. R., and Mostofi, F. K. (1998) The World Health Organization/International Society of Urological Pathology consensus classification of urothelial (transitional cell) neoplasms of the urinary bladder. Bladder Consensus Conference Committee. Am. J. Surg. Pathol. 22, 1435 –1448[CrossRef][Medline]

  30. Reuter, V. E., Epstein, J. I., Amin, M. B., Mostofi, F. K. (1999) The WHO/ISUP consensus classification of urothelial (transitional cell) neoplasms: continued discussion. Hum. Pathol. 30, 879 –880[Medline]

  31. Celis, J. E., and Gromov, P. (1999) 2D protein electrophoresis: can it be perfected? Curr. Opin. Biotechnol. 10, 16 –21[CrossRef][Medline]

  32. Celis, J. E., and Bravo, R. (eds) (1984) Two-Dimensional Gel Electrophoresis of Proteins, pp.346 –346, Academic Press, Orlando, Florida

  33. Liotta, L., and Petricoin, E. (2000) Molecular profiling of human cancer. Nat. Rev. Genet. 1, 48 –56[CrossRef][Medline]

  34. Cordwell, S. J., Nouwens, A. S., Verrills, N. M., Basseal, D. J., and Walsh, B. J. (2000) Subproteomics based upon protein cellular location and relative solubilities in conjunction with composite two-dimensional electrophoresis gels. Electrophoresis 21, 1094 –1103[CrossRef][Medline]

  35. Puig, O., Caspary, F., Rigaut, G., Rutz, B., Bouveret, E., Bragado-Nilsson, E., Wilm, M., and Seraphin, B. (2001) The tandem affinity purification (TAP) method: a general procedure of protein complex purification. Methods 24, 218 –229[CrossRef][Medline]

  36. Schaafsma, H. E., Ramaekers, F. C., van Muijen, G. N., Lane, E. B., Leigh, I. M., Robben, H., Huijsmans, A., Ooms, E. C., and Ruiter, D. J. (1990) Distribution of cytokeratin polypeptides in human transitional cell carcinomas, with special emphasis on changing expression patterns during tumor progression. Am. J. Pathol. 136, 329 –343[Abstract]

  37. Helpap, B., and Kollermann, J. (2000) Assessment of basal cell status and proliferative patterns in flat and papillary urothelial lesions: a contribution to the new WHO classification of the urothelial tumors of the urinary bladder. Hum. Pathol. 31, 745 –750[Medline]

  38. Cheng, L., Neumann, R. M., Nehra, A., Spotts, B. E., Weaver, A. L., and Bostwick, D. G. (2000) Cancer heterogeneity and its biologic implications in the grading of urothelial carcinoma. Cancer 88, 1663 –1670[CrossRef][Medline]

  39. Nocito, A., Kononen, J., Kallioniemi, O. P., and Sauter, G. (2001) Tissue microarrays (TMAs) for high-throughput molecular pathology research. Int. J. Cancer 94, 1 –5[CrossRef][Medline]