Coexistent intraurothelial carcinoma and muscle-invasive - Figshare

www.elsevier.com/locate/humpath

Human Pathology (2009) 40, 988–997

Original contribution

Coexistent intraurothelial carcinoma and muscle-invasiveurothelial carcinoma of the bladder: clonality and somaticdown-regulation of DNA mismatch repairAlfredo Blanes MD, PhDa, Javier Rubio MD, PhDa, Juan J. Sanchez-Carrillo MD, PhDa,Salvador J. Diaz-Cano MD, PhD, FRCPath a,b,⁎

aDepartment of Pathology, University Hospital of Malaga, 29071 Malaga, SpainbDepartment of Pathology, King's College Hospital and King's College London School of Medicine, SE5 9RS London,United Kingdom

Received 29 July 2008; revised 11 December 2008; accepted 19 December 2008

C2Ir

H

0d

Keywords:Bladder;Carcinoma in situ;Low-grade dysplasia;Urothelial carcinoma;Heterogeneity;Tumor suppressor genes;DNA mismatch repair

Summary Muscle-invasive urothelial carcinomas are heterogeneous neoplasms for which the clonalrelationship with low-grade urothelial dysplasia and carcinomas in situ remains unknown, and bothmonoclonal and field change models have been proposed. Low-grade dysplasia (18) and carcinoma insitu (12) associated with muscle-invasive urothelial carcinoma were microdissected and topographicallyanalyzed (intraepithelial and invasive superficial and deep to muscularis mucosa) for methylationpattern of androgen receptor alleles, TP53, RB1, WT1, and NF1 microsatellite analysis to assess clonalidentity; MLH1 and MSH2 sequencing/immunostaining. Appropriate controls were run. Carcinoma insitu (100%) and invasive urothelial carcinoma (100%) revealed monoclonal patterns, whereas low-gradedysplasia was preferentially polyclonal (80%). Carcinoma in situ showed aneuploid DNA content andmore abnormal microsatellites than the corresponding invasive compartments, opposite to low-gradedysplasia. Absent MLH1 protein expression with no gene mutations were identified in carcinoma in situand nodular-trabecular urothelial carcinoma with high microsatellite abnormalities. Somatic mismatchrepair protein down-regulation and the accumulation of tumor suppressor gene microsatelliteabnormalities contribute to a molecular evolution for monoclonal carcinoma in situ divergent fromcoexistent muscle-invasive urothelial carcinoma. Low-grade dysplasia is however unlikely connectedwith this molecular progression.© 2009 Elsevier Inc. All rights reserved.

1. Introduction

Presented in part as abstracts in the meetings of the United States andanadian Academy of Pathology, Atlanta, GA, 2001; San Antonio, TX,005; and Denver, CO, 2008; and Pathological Society of Great Britain andeland, Maastricht, Netherland, 2001.⁎ Corresponding author. Department of Histopathology, King's College

ospital, SE5 9RS London, United Kingdom.E-mail address: [email protected] (S. J. Diaz-Cano).

046-8177/$ – see front matter © 2009 Elsevier Inc. All rights reserved.oi:10.1016/j.humpath.2008.12.009

Urothelial dysplasia is assumed to be the putative precursorof urothelial carcinoma (UCC) and confers a significant risk forthe development of carcinoma in situ (CIS) and invasive UCC,as reported for intraepithelial breast or melanocytic lesions [1].Although the accumulated molecular data indicate that mostrecurrent and multiple tumors are monoclonal, the controversial

mailto:[email protected]

http://dx.doi.org/10.1016/j.humpath.2008.12.009

989MMR and Microsatellites in VCC

definitions of flat lesions with atypia, that is, reactive atypia,atypia of unknown significance, low-grade urothelial dysplasia(LGUD), and high-grade urothelial dysplasia-CIS [2], havecontributed to create confusion for coexistent lesions [3]. Inaddition, the concept of tumor progression is not consistentlyused in bladder pathologic examination [4], being a cytologicprogression documented in urothelial dysplasias but with adifferent topography for LGUD and CIS. This supports themultifocal distribution of low-grade UCC without proving aclonal identity of LGUD and CIS to sustain the sequenceLGUD —N CIS.

The clonality status of multifocal bladder tumors is stillcontroversially discussed with experimental evidence for bothmonoclonality and field cancerization. Early stage urothelialneoplasms have shown chromosome 9 deletions and FGFR3mutations [5,6], the same genetic alteration being observed incoexistent low-grade papillary superficial UCCs and histologi-cally normal urothelium, whereas 17p13 hemizygosity wasobserved in a minority of urothelial hyperplasias and papillarytumors but not in normal urothelium. This genetic profilesuggests that the earliest molecular alterations in the pathogen-esis of low-grade UCC involve p16/CDKI2 but not TP53 evenin histologically normal areas, but this study does not analyzeurothelial dysplasia and high-grade muscle-invasive UCC [1,7].Clusters of discontinuous deleted segments of tumor suppressorgene loci on chromosomes 13q14 and 17p13 have beenassociated with clonal expansion of in situ bladder preneoplasiausing single nucleotide polymorphisms [8]. The clonal relation-ships between LGUD and CIS associated with muscle-invasiveUCC have not been addressed. Although bladder UCCinfrequently reveals microsatellite instability [9-11], the reducedexpression of mismatch repair (MMR) proteins contributes tothe development of a subset of UCC [12,13].

The relationship (linear versus divergent) between intraur-othelial and invasive compartments (superficial and deep) hasnot been topographically analyzed in muscle-invasive UCCshowing these 3 compartments. This study evaluates in eachtopographic compartment as follows: clonality, MMR proteinexpression/sequencing, and loss of heterozygosity (LOH)-microsatellite profile of tumor suppressor genes (TSG)controlling G1-S transition (TP53, RB1), RAS pathway (NF1),and development (WT1). All these analyses have not beenperformed simultaneously in a series of muscle-invasive andintraurothelial lesions and will inform on both monoclonal/fieldchange models in muscle-invasive UCC and tumor progressionfrom the perspective of the accumulation of genetic alterationsin tumor suppressor genes.

2. Materials and methods

2.1. Case selection and sampling

We reviewed the initial transurethral resection biopsy ofpT2a/b UCC with no special differentiation of the bladder

diagnosed in women (44 cases), treated with cystectomy andlymphadenectomy from 3 reference hospitals (1990-1992,median follow-up 60 months). Transurethral resectionbiopsies were selected because they provided better cellularpreservation, and the results for molecular tests are morereliable as they show lower frequency of artifacts [14,15].Intraurothelial neoplastic lesions were classified according tothe World Health Organization system in low-gradedysplasia (LGUD, 11 patients) and carcinoma in situ (CIS,7 patients) [3], both coexisting in 3 patients. To add power tothe clonality analysis, a combined approach of X chromo-some inactivation and tumor suppressor gene microsatelliteprofile was selected. The X chromosome inactivationrequires samples from female that represent the initialpatient selection; to extend the analysis, cases from malewere selected (28, which revealed LGUD in 7 patients andCIS in 5 patients, coexistent lesions in 2), the selectioncriteria then included matching cases by age and conven-tional histologic features of the muscle-invasive componentto avoid any biases from those aspects. The cases wereclassified independently by 2 pathologists (AB, SDC). Incase of grading disagreement, the lesions were discussedduring simultaneous inspection before final categorization.Reproducibility data were not recorded.

All surgical specimens were completely embedded forhistopathologic diagnosis. The topographic compartmentlimit was the muscularis mucosa (superficial and deep tothe muscularis mucosa), being the same areas analyzed ineach study [16-18]. The muscularis mucosae was selected aslimit because tumors invasive to this level had shown amuch better 5-year survival than tumors invasive through thelevel of the muscularis mucosae, which showed survivalcomparable with patients with tumors invasive of themuscularis propria [19]. This protocol was approved bythe Hospital Research Board and Ethical Committee andcomplied with their requirements.

2.2. Clonality assay and TSG microsatellite analysis

DNA was extracted from the most cellular areas ofintraurothelial, superficial, and deep compartments, aftermicrodissecting at least 100 cells (approximately 0.4 mm2,laser capture; Arturus, Sunnyvale, CA) from two 20-μmunstained paraffin sections/compartment. Appropriate tissuecontrols (histologically normal urothelium, stroma from thelamina propria, and smooth muscle) and quality-assurancecontrols (sensitivity, specificity, positive, and negative) wererun for each test [14,15,20].

DNA was extracted using a modified phenol-chloroformprotocol, precipitated with ice-cold absolute ethanol, andresuspended in 10 μL of Tris-HCl buffer pH 8.4 [15]. DNAwas then used for polymerase chain reaction (PCR) amplifica-tion of TSG intron microsatellites and the hypervariable CAGrepeat in the first exon of the human androgen receptor (seeTable 1 for primer sequences and cycling conditions), using

Table 1 Primer sequences and PCR cycling conditions

Primers Primer sequences Repeats/PCR product PCR cycling conditions

AR-F⁎ 5′-CCG AGG AGC TTT CCA GAA TC-3′ CAG/215-300 bp ⁎AR alleles were amplified using“hot start” protocol, 0.3 μmol/L ofeach primer, and 200 μmol/L ofeach dNTP (including 7-deaza-dGTPinstead of dGTP) (Boehringer-Mannheim,Indianapolis, IN), completing 28 cycleswith an annealing temperature of 55°C.The amplicon was internally labeledwith 0.3 μCi α[32P]-dTTP (800 Ci/mmol,10 mCi/mL; New England Nucleotide,Boston, MA).

AR-R⁎ 5′-TAC GAT GGG CTT GGG GAG AA-3′

TP53(1)-F‡ 5′-AGG GAT ACT ATT CAG CCC-3′ CA/103-135 bp ‡TSG polymorphic regions were amplifiedusing 0.25 μmol/L of each primer,50 μmol/L of each dNTP(Boehringer-Mannheim, Indianapolis, IN),and internally labeled with 0.3 μCiα[32P]-dCTP (3000 Ci/mmol, 10 mCi/mL;New England Nucleotide, Boston, MA).The annealing temperature was 55°C(except for NF1, 52°C), and the numberof cycles was experimentally optimizedto 26.

TP53(1)-R‡ 5′-ACT GCC ACT CCT TGC CCC ATT C-3′

TP53(2)-F‡ 5′-GAA TCC GGG AGG AGG TTG-3′ AAAAT/140-175 bpTP53(2)-R‡ 5′-AAC AGC TCC TTT AAT GGC AG-3′RB1-F‡ 5′-CTC CTC CCC TAC TTA CTT GT-3′ CTTT(T)/266-306 bpRB1-R‡ 5′-AAT TAA CAA GGT GTG GTG GTA CAC G-3′WT1-F‡ 5′-AAT GAG ACT TAC TGG GTG AGG-3′ CA/approximately 144 bpWT1-R‡ 5′-TTA CAC AGT AAT TTC AAG CAA CGG-3′NF1-F‡ 5′-CAG AGC AAG ACC CTG TCT-3′ CA/171-187 bpNF1-R‡ 5′-CTC CTA ACA TTT ATT AAC CTT A-3′

Abbreviations: F, forward; R, reverse.All reactions were run in duplicate using 1.5 mmol/L of MgCl2 and 1 μL of template, as well as long denaturation (4 minutes) and expansion (90 seconds) inthe first 3 cycles. The appropriate PCR primers were designed using Genrunner software (version 3.02; Hastings Software Inc., Hudson, NY).

990 A. Blanes et al.

HhaI-undigested and digested samples for the X chromosomeinactivation assay that contained XhoI-linearized φX174-RIIphage (Gibco-BRL, Gaithersburg, MD) as mimicker ofdigestion completion checked by gel electrophoresis (Table 1)[16,17,20-22]. The tests were run in a Perkin-Elmer thermalcycler model 480 (Perkin-Elmer, Norwalk, CT). The whole 10-μL PCR volume was electrophoresed into 8% denaturinggradient polyacrylamide gels; dried gels were put insidedeveloping cassettes containing one intensifying screen andpreflashed films (Kodak XAR) [17,23]. The radiographs weredeveloped using an automated processor Kodak-Omat 100(Kodak Co, Rochester, NY).

Interpretation and inclusion criteria included [14,17,20-22]the following: (a) allelic imbalance was densitometricallyevaluated (EC model 910 optical densitometer, EC Appara-tus Corporation, St Petersburg, FL), considering evidence ofLOH only allele ratios 4:1 or greater in any TSG; otherwise,retention of heterozygosity was assigned [17,22]. This ratiowould represent 80% of clonal cells in the sample and wasused to increase the detection specificity [20,21,24]. (b)Additional allele bands present in tumor samples but not in

the corresponding controls were considered evidence ofsomatic microsatellite abnormality by PCR/denaturinggradient gel electrophoresis [17].

2.3. DNA sequencing

All microsatellite extrabands were cut from gels, andDNA was purified using a QIA quick gel extraction kit(Qiagen, West Sussex, UK). The amplified product wasdiluted 20-fold in TE buffer, and 1-μL of the diluted reactionproduct was subjected to a second round of PCR amplifica-tion using the appropriate primers for 30 cycles under theabove conditions. Normal and extrabands from tumor-derived samples were PCR amplified along with thecorresponding controls using a high-fidelity polymerase,Platinum PFX (Life Technologies). PCR products weredirectly sequenced after purification (QIAquick PCRpurification kit, Qiagen). All sequencing was performed onan ABI Prism 3700 automated DNA analyzer, and thesequence data were analyzed using the program Sequencher


(Gene Codes Corporation, Ann Arbor, MI), which reversesand complements the antisense strand. All mutations wereconfirmed by sequencing in both directions and indicated byan “N” in the sequencing chromatogram.

MLH1/MSH2 exons were completely sequenced in caseswith microsatellite abnormalities in at least 40% loci and/orcomplete loss of mlh1/msh2 immunoreactivity, as well as in asample of mlh1/msh2 immunoreactive UCC [20] as controls.

2.4. Immunohistochemical detection of TP53,MLH1, and MSH2

The sections were mounted on positively charged slides(Superfrost Plus, Fisher Scientific, Fair Lawn, NJ), baked at60°C for 2 hours, and processed as described [16,17,25]. Afterroutine dewaxing and rehydration, endogenous peroxidasequenching, and antigen heat retrieval (pressure cooker, citratebuffer [10 mmol/L], for all antibodies), the slides weretransferred to a moist chamber. Nonspecific binding wasblocked with polyclonal horse serum, and sections wereincubated with monoclonal primary antibodies (overnight,4°C) as follows: 2 μg/mL of p53 DO-7, Calbiochem,Cambridge, MA; hMLH1 clones G168 728 and G168-15,

Fig. 1 Coexistent urothelial CIS was associated with nodular-trabecureplacing the muscularis propria) but not with infiltrative UCC (small tumdissecting the smooth muscle fibers) (hematoxylin-eosin, original magni

BD Pharmingen Biosciences, Oxford, UK; hMSH2 cloneFE11, Oncogene Research (Merck Chemicals Ltd., Notting-ham, UK). Then sections were serially incubated withbiotinylated antimouse antibody and peroxidase-labeled avi-din-biotin complex. The reaction was developed undermicroscopic control, using 3,3′-diaminobenzidine tetrahy-drochloride with 0.3% H2O2 as chromogen (Sigma Co, StLouis, MO), and the sections were counterstained withhematoxylin. Both positive (reactive lymph node) and negative(omitting the primary antibody) controls were simultaneouslyrun. Basal cells of the unaffected urothelial mucosa were usedas internal positive controls for mlh1 and msh2.

2.5. Nuclear DNA quantification by slide cytometry

Feulgen-stained sections were used for DNA quantifica-tion [26]. The densitometric evaluation was performed withthe cell analysis system model 200 and the quantitative DNAanalysis software package (Becton Dickinson, Oxford, UK).At least 300 complete, nonoverlapping, and focused nuclei(or the whole lesion if smaller) were measured in every case,beginning in the most cellular area until completion inconsecutive high-power fields (HPFs).

lar UCC (sheets of neoplastic cells with minimal stromal reaction,or nests/thin cords embedded in a prominent desmoplastic reaction,fication ×400).

Fig. 2 Clonality and microsatellite profile in bladder intraurothelial lesions associated with muscle-invasive UCC. Frequency of tumorsuppressor gene microsatellite (TSG MS) abnormalities in LGUD and CIS. Abbreviations: ROH, retention of heterozygosity (green cell);LOH/SNP, loss of heterozygosity/single nucleotide polymorphism (red cell); NI, noninformative (gray cell); TP53, tumor protein p53; RB1,retinoblastoma; WT1, Wilms tumor 1; NF1, neurofibromatosis 1.


External staining calibration was carried out withcomplete rat hepatocytes (Becton-Dickinson; one slide perstaining holder) to normalize the internal controls (lympho-cytes and histologically normal urothelial cells present in thesame tissue section), used for setting the G0/G1 cell limitsand calculating the DNA index of each G0/G1 peak (N10% ofmeasured cells with evidence of G2 + M cells) [27].Proliferation rate (PR = S + G2 + M-phases fraction) wascalculated from the DNA histogram by subtracting thenumber of cells within G0/G1 limits from the total number ofmeasured cells and expressed as percentage [26,27].

External diploid controls were used to determine DNAindices (lymphocytes from reactive lymph nodes) and tostandardize the nuclear area/DNA content analysis (normaltransitional cells) [28].

2.6. Tumor infiltration pattern, grading, andmitotic figure counting

The infiltration pattern was evaluated in deep compart-ments, classifying the tumor by the predominant pattern(N50%) in nodular-trabecular and infiltrative [29]. The

histologic grading evaluated architectural features, nucleargrade, and mitotic figure (MF) counting [13]. MFs werescreened in 50 HPF/compartment (7.140 mm2) or the wholetumor if smaller (3 superficial and 6 deep compartments),beginning in the most cellular area [16]. Both the number ofpositive nuclei/HPF and the number of neoplastic cellsintercepted by the microscope field diameter (n) wereregistered, the latter to estimate the number of neoplasticcells/HPF (N = [nπ/4]2) [30]; results were expressed per1000 cells, calculating average and SD per compartment andpatient. Tumors were graded by 3 independent observers (JR,AB, and SJD-C), being the tumor discussed duringsimultaneous inspection before final categorization in caseof disagreement. Reproducibility data were not recorded.

2.7. Quantification of positive nuclei andstatistical analysis

At least 50 HPF (7.6 mm2) were screened in each pathologicgroup, beginning in the most cellular area. The number ofpositive nuclei was expressed per HPF and per 1000 tumorcells, calculating average and SD for each pathologic condition


and patient [22,30]. The positivity threshold was experimentallyestablished at the positive control in each staining batch. Onlynuclei with staining features similar to those of theircorresponding positive control were considered positive forany marker.

Categorized variables were tested using Fisher exact testsand quantitative variables using Student t tests (if normallydistributed) and nonparametric tests (Mann-Whitney for 2-group comparisons and Kruskal-Wallis for N2-group com-parisons). Differences were considered significant if P b .05in 2-tailed distributions.

3. Results

CIS was identified in patients with nodular-trabecular UCC(Fig. 1) and revealed more abnormal TSG loci than thecorresponding invasive compartment (10/12 patients), TP53loci being involved in all patients (Fig. 2) with expression ofabnormal p53 protein. CIS showed either an additional TSG

Fig. 3 TSG microsatellite pattern in carcinoma in situ (CIS, all monoclsuperficial [sup] and deep compartments) (panel A). Representative gel(monoclonal, TCC1-CIS) and LGUD (monoclonal, TCC7-LGUD, and pNuclear mlh1 and msh2 expression is demonstrated in UCC with no micr(in particular mlh1) in UCC with microsatellite abnormalities.MLH1 andgenes, regardless of the microsatellite pattern.

locus involved (8 patients; TP53 in 5, RB1 in 2, andWT1 in 1)or a combined pattern of superficial and deep compartments (2patients). The other 2 cases showed TP53 abnormality only ordifferent microsatellite abnormalities at the same loci inintraurothelial and invasive compartments. In contrast, LGUDrevealed LOH in 2 patients, one at RB1 (monoclonalmethylation of androgen receptor alleles) and the second atWT1-NF1 loci (polyclonal pattern), respectively (Fig. 2). CIS(6; 100%), invasive UCC (13; 100%), and LGUD (2; 20%)from informative females revealed unbalanced methylationpattern of androgen receptor alleles, whereas polyclonalpatterns were observed in LGUD only (8; 80%; Figs. 2and 3). Discordant pattern of AR allele was observed in onecase, the larger allele being methylated in LGUD and thesmaller allele in CIS-invasive UCC. The UCC microsatelliteprofile of superficial and deep compartments was provenstatistically different from CIS profile at TP53 locus only (P =0.042; Fig. 4), showing similar topographic heterogeneity inUCC invasive compartments, regardless of the presence orabsence of CIS.

onal), LGUD (mainly polyclonal), and muscle-invasive UCC (froms of the methylation allele pattern of androgen receptor from CISolyclonal, TCC2-LGUD) (panel B). Mismatch protein expression.osatellite abnormalities and at least one of these proteins was absentMSH2 exon sequencing. Normal sequence is demonstrated for these

Fig. 4 Clonality and microsatellite profile of muscle-invasive UCC by topographic compartments. The genetic pattern was statisticallydifferent in UCC with dysplasia compared with UCC without dysplasia. Microsatellite abnormalities were more frequent in superficialcompartments (RB1, P b .001; NF1, P = .004) and deep compartments (RB1, P = .003) of UCC without dysplasia. Different patterns werenoted in the superficial and deep compartments from a given patient in 5 patients with urothelial dysplasia (20%). This topographic geneticheterogeneity was due to the presence of additional abnormalities in superficial compartments (NF1, 1 patient) or deep compartments (TP53, 2patients; RB1, 1 patient) or discordant pattern (1 patient). Frequency of TSG microsatellite abnormalities in superficial and deep compartmentsof muscle-invasive UCC. Abbreviations: ROH, retention of heterozygosity (green cell); LOH/SNP, loss of heterozygosity/single nucleotidepolymorphism (red cell); NI, noninformative (gray cell); TP53, tumor protein p53; RB1, retinoblastoma; WT1, Wilms tumor 1; NF1,neurofibromatosis 1; S, superficial; D, deep.


Nodular-trabecular UCCs were more frequently aneuploid(27/28; 96%) and high grade (26/28; 93%) than infiltrativeUCCs (9/16, 56%, and 12/16, 75%, respectively). Nodular-trabecular UCCs and superficial compartments showed sig-nificantly higher values for both mitotic figure counting andproliferation rate (Table 2). The number of diploid (6 cases) andlow-grade (5 cases) UCCs precluded any statistical compar-isons of these features. Nodular-trabecular UCCs revealed moreabnormal loci than infiltrative UCCs (Fig. 2; P = .0001).Discordant genetic patterns by tumor compartments wereobserved in only 2 infiltrative UCCs precluding any statisticalassessment, but all showed nuclear TP53 expression and moreLOH/single nucleotide polymorphisms (SNPs) in the deepcompartment (WT1 LOH/SNP in 1 and NF1 LOH/SNP in 1).

Immunostaining for mlh1/msh2 revealed statistically sig-nificant reduction of at least one of the proteins (especiallymlh1) in CIS and the deep compartment of trabecular UCC

with 2 or more TSG microsatellite abnormalities (Fig. 3). Nosignificant difference was observed in the mlh1/msh2 immu-noexpression in UCC with less than 2 TSG geneticabnormalities (MS stable or MS instable-low) but revealeddeficient MMR system at the deep compartment. NormalMLH1/MSH2 exons sequences were observed in all UCCanalyzed, regardless of immunoexpression and microsatellitestatus (Fig. 3).

4. Discussion

Microsatellite analysis of TSG supports that coexistentCIS and muscle-invasive UCC evolves independently,contributing to intratumoral heterogeneity, despite havinga common progenitor (monoclonal proliferation). Thesomatic MMR protein down-regulation contributes to the

Table 2 Kinetic features of muscle-invasive UCC by infiltration patterns (nodular-trabecular versus infiltrative) and tumor compartments(superficial versus deep)

Nodular-trabecular pT2a/b UCC Infiltrative pT2a/b UCC Significance

Superficial Deep Superficial Deep

MF counting 9.0 ± 5.1 4.1 ± 3.1 5.9 ± 3.5 2.0 ± 1.8 P = .012Proliferation rate 33.74 ± 7.60 19.69 ± 6.15 26.46 ± 8.36 14.57 ± 4.58 P = .009


accumulation of genetic alterations, which suggests anindependent evolution rather than a precancerous lesion/early neoplasm. LGUD is mainly a polyclonal lesionrevealing low incidence of TSG microsatellite abnormalities,suggestive of a nonneoplastic condition.

A linear progression model for UCC would expect tofind progressive accumulation of genetic alterations in thetransition intraurothelial-superficial invasive-deep invasivecomponents of UCC. However, this gradient is not found.CIS displays microsatellite alterations similar to UCC deepcompartments [1,16,17] but with more genetic abnormal-ities in CIS (Figs. 2,4,5). This and the differential TSGmicrosatellite pattern of UCC superficial compartments(NF1-defective) [16,17] do not support the sequence

Fig. 5 Muscle-invasive UCC can evolve through pathways with and wgrowth pattern (trabecular-nodular with coexistent CIS versus infiltrat(interstitial DNA loss versus no DNA loss).

coexistent CIS—Nsuperficial UCC—Ndeep UCC [4].Several intraepithelial foci show more alterations thanmatched invasive foci, suggesting a more extensive geneticevolution for the former and supporting multifocality andindependent clonal evolution of these coexistent carcino-mas [30,31]. The accumulation of genetic abnormalities incoexistent CIS is consistent with an independent progres-sion of bladder carcinoma (Fig. 2) [32-37]. Geneticalterations centered around RB1 may represent an incipientevent in bladder neoplasia. However, the inactivation ofRB1 occurred later and was associated with the onset ofsevere dysplasia/carcinoma in situ [17]. There is alsoevidence for the presence of critical alternative candidategenes mapping to the 13q14 region that are involved in

ithout microsatellite abnormalities that correlate with the invasiveive) and TSG (NF1/RB1/TP53) regulation of the G1-S transition


clonal expansion of neoplasia within the bladder ante-cedent to the inactivation of the RB1 gene [38]. Finally, weperformed high-resolution mapping using single nucleotidepolymorphism markers within one region on chromosome13q14, containing the model tumor suppressor gene RB1,and defined a minimal deleted region associated withclonal expansion of in situ neoplasia. These analysesprovided new insights on the involvement of severalnoncoding sequences mapping to the region and identifiednovel target genes, termed forerunner (FR) genes, involvedin early phases of cancer development [39]. In addition,the invasive compartment microsatellite pattern of UCCwith intraurothelial lesions revealed a significant decreaseof RB1 and NF1 abnormalities (Fig. 4), which correlatedwith a nodular-trabecular pattern and high cellular turnover[13], proving the topographic genetic heterogeneity ofmuscle-invasive UCC [16,17].

LGUD shows much lower incidence of genetic abnorm-alities, making the direct connection with the linearprogression unlikely. Although a clonal relationship hasbeen suggested by LOH and/or comparative genomichybridization analyses, supporting the hypothesis that flaturothelial hyperplasias can display many genetic alterationscommonly found in bladder cancer [40], the kinetic featuresof LGUD makes unlikely this progression. The combinedgenetic-kinetic studies are needed for a full assessment [7,16].LGUD showed low incidence of TSG microsatelliteabnormalities, no TP53 alterations, and polyclonal patterns.These results support the existence of 2 transformationpathways for bladder UCC, TP53 alterations in high-gradeUCC (frequently muscle-invasive) and p16 in low-gradeUCC (often pT1) [7]. Our results also disprove the clonalidentity of coexistent LGUD and CIS, and their dissimilargenetic profiles and topography question the sequence LGUD—Nco-existent CIS. The hypotheses of tumor evolution andoligoclonality as derived fromLOH data need to be supportedby deletion-independent clonality studies as X chromosomalinactivation analysis [20,21,41]. Multifocality and recurrenceare clinically important features of urothelial carcinomas ofthe urinary bladder. Combination of molecular data withhistopathologic bladder mapping suggested a monoclonaldevelopment of the multifocal lesions mostly via intraur-othelial migration. Recent molecular genetic studies havesuggested that multifocal urothelial carcinomas are mono-clonally derived from an identical transformed progenitor cell[31,42].

MMR protein down-regulation, normal MLH1/MSH2sequences, and microsatellite abnormalities characterized CIS(12 cases, P = .0053) and nodular-trabecular UCC.Microsatellite profiles of bladder UCC have showninfrequent instability [9-11], which can be an independentprognostic marker for assessing risk of recurrence insuperficial tumors irrespective of the grade [43]. Likewise,reduced expression of the MMR proteins may have animportant contribution in the development of a subset ofUCC and is a potentially useful prognostic marker [12,13],

in particular for upper urinary tract tumors [44]. However,the pattern of microsatellite instability depends on thelocation and elevated microsatellite alterations at selecttetranucleotides being reported more frequently in bladder[44]. Our results add other type of microsatellite abnorm-alities (extrabands due to single nucleotide substitutions) inmuscle-invasive UCC revealing solid-trabecular growthpattern and coexistent CIS. MMR gene inactivation (byeither mutation or protein down-regulation) leads to mutationaccumulation and molecular progression not necessarilyindependent from chromosomal instability [20].

In conclusion, TSG microsatellite patterns support anonlinear and independent genetic evolution of coexistentCIS and invasive UCC (Fig. 5). The combined assessment ofclonal identity (LOH of tumor suppressor genes and X-chromosome inactivation) and molecular progression (definedas accumulation of genetic abnormalities) is comprehensive andincorporates some kinetic features that allow the persistence andtransmission of genetic alterations to descendant cells. SomaticMMR protein down-regulation, the accumulation of TSGmicrosatellite abnormalities, and monoclonal pattern suggest adivergent molecular evolution for CIS and coexistent muscle-invasive UCC (in particular those with trabecular growthpattern). UCC with infiltrative pattern should follow alternativepathways and LGUD are most likely unrelated with themechanisms of somatic down-regulation of mismatch repairsystem. In contrast, LGUD shows polyclonal pattern and lowincidence of TSG microsatellite abnormalities, suggestive of anonneoplastic condition.

References

[1] Rosin MP, Cairns P, Epstein JI, Schoenberg MP, Sidransky D. Partialallelotype of carcinoma in situ of the human bladder. Cancer Res 1995;55:5213-6.

[2] McKenney JK, Gomez JA, Desai S, Lee MW, Amin MB. Morphologicexpressions of urothelial carcinoma in situ: a detailed evaluation of itshistologic patterns with emphasis on carcinoma in situ withmicroinvasion. Am J Surg Pathol 2001;25:356-62.

[3] Eble JN, Sauter G, Epstein JI, Sesterhenn IA, editors. World HealthOrganization classification of tumours: pathology and genetics oftumours of the urinary system and male genital organs. Lyon: IARCPress; 2004.

[4] Diaz-Cano SJ. General morphological and biological features ofneoplasms: integration of molecular findings. Histopathology 2008;53:1-19.

[5] Lindgren D, Liedberg F, Andersson A, Chebil G, Gudjonsson S, BorgA, et al. Molecular characterization of early-stage bladder carcinomasby expression profiles, FGFR3 mutation status, and loss of 9q.Oncogene 2006;25:2685-96.

[6] Eiber M, van Oers JMM, Zwarthoff EC, van der Kwast TH, Ulrich O,Helpap B, et al. Low frequency of molecular changes and tumorrecurrence in inverted papillomas of the urinary tract. Am J SurgPathol 2007;31:938-46.

[7] Baithun SI, Naase M, Blanes A, Diaz-Cano SJ. Molecular and kineticfeatures of transitional cell carcinomas of the bladder: biological andclinical implications. Virchows Arch 2001;438:289-97.

[8] Tuziak T, Jeong J, Majewski T, Kim MS, Steinberg J, Wang Z, et al.High-resolution whole-organ mapping with SNPs and its significanceto early events of carcinogenesis. Lab Invest 2005;85:689-701.


[9] Bonnal C, Ravery V, Toublanc M, Bertrand G, Boccon-Gibod L,Henin D, et al. Absence of microsatellite instability in transitional cellcarcinoma of the bladder. Urology 2000;55:287-91.

[10] Ericson KM, Isinger AP, Isfoss BL, Nilbert MC. Low frequency ofdefective mismatch repair in a population-based series of upperurothelial carcinoma. BMC Cancer 2005;5:23.

[11] Mongiat-Artus P, Miquel C, van der Aa M, Buhard O, Hamelin R,Bangma C, et al. Infrequent microsatellite instability in urothelial cellcarcinoma of the bladder in young patients. Eur Urol 2006;49:685-90.

[12] Catto JW, Xinarianos G, Burton JL, Meuth M, Hamdy FC. Differentialexpression of hMLH1 and hMSH2 is related to bladder cancer grade,stage and prognosis but not microsatellite instability. Int J Cancer2003;105:484-90.

[13] Rubio J, Blanes A, Sanchez-Carrillo JJ, Diaz-Cano SJ. Microsatelliteabnormalities and somatic down-regulation of mismatch repaircharacterize nodular-trabecular muscle-invasive urothelial carcinomaof the bladder. Histopathology 2007;51:458-67.

[14] Diaz-Cano SJ. Are PCR artifacts in microdissected samples pre-ventable? HUM PATHOL 2001;32:1415.

[15] Diaz-Cano SJ, Brady SP. DNA extraction from formalin-fixed,paraffin-embedded tissues: protein digestion as a limiting step forretrieval of high-quality DNA. Diagn Mol Pathol 1997;6:342-6.

[16] Blanes A, Rubio J, Martinez A, Wolfe HJ, Diaz-Cano SJ. Kineticprofiles by topographic compartments in muscle-invasive transitionalcell carcinomas of the bladder: role of TP53 and NF1 genes. Am J ClinPathol 2002;118:93-100.

[17] Diaz-Cano SJ, Blanes A, Rubio J, Matilla A, Wolfe HJ. Molecularevolution and intratumor heterogeneity by topographic compartmentsin muscle-invasive transitional cell carcinoma of the urinary bladder.Lab Invest 2000;80:279-89.

[18] Ro JY, Ayala AG, el-Naggar A. Muscularis mucosa of urinary bladder.Importance for staging and treatment. Am J Surg Pathol 1987;11:668-73.

[19] Younes M, Sussman J, True LD. The usefulness of the level of themuscularis mucosae in the staging of invasive transitional cellcarcinoma of the urinary bladder. Cancer 1990;66:543-8.

[20] Diaz-Cano SJ, Blanes A, Wolfe HJ. PCR techniques for clonalityassays. Diagn Mol Pathol 2001;10:24-33.

[21] Diaz-Cano SJ. Designing a molecular analysis of clonality in tumours.J Pathol 2000;191:343-4.

[22] Diaz-Cano SJ, de Miguel M, Blanes A, Tashjian R, Galera H, WolfeHJ. Clonality as expression of distinctive cell kinetics patterns innodular hyperplasias and adenomas of the adrenal cortex. Am J Pathol2000;156:311-9.

[23] Blanes A, Sanchez-Carrillo JJ, Diaz-Cano SJ. Topographic molecularprofile of pheochromocytomas: role of somatic down-regulation ofmismatch repair. J Clin Endocrinol Metab 2006;91:1150-8.

[24] Diaz-Cano SJ. Clonality studies in the analysis of adrenal medullaryproliferations: application principles and limitations. Endocr Pathol1998;9:301-16.

[25] Diaz-Cano SJ, de Miguel M, Blanes A, Tashjian R, Galera H, WolfeHJ. Clonal patterns in phaeochromocytomas and MEN-2A adrenalmedullary hyperplasias: histological and kinetic correlates. J Pathol2000;192:221-8.

[26] Bibbo M, Bartels PH, Dytch HE, Wied GL. Cell image analysis. In:Bibbo M, editor. Comprehensive cytopathology. Philadelphia: Saun-ders Co; 1991. p. 965-83.

[27] Dressler LG. Controls, standards, and histogram interpretation in DNAflow cytometry. Methods Cell Biol 1990;33:157-71.

[28] Sherwood SW, Schimke RT. Cell cycle analysis of apoptosis usingflow cytometry. Methods Cell Biol 1995;46:77-97.

[29] Jimenez RE, Gheiler E, Oskanian P, Tiguert R, Sakr W, Wood Jr DP,et al. Grading the invasive component of urothelial carcinoma of thebladder and its relationship with progression-free survival. Am J SurgPathol 2000;24:980-7.

[30] Koch M, de Miguel M, Hofler H, Diaz-Cano SJ. Kinetic profiles ofintraepithelial and invasive prostatic neoplasias: the key role of down-regulated apoptosis in tumor progression. Virchows Arch 2000;436:413-20.

[31] Denzinger S, Mohren K, Knuechel R, Wild PJ, Burger M, WielandWF, et al. Improved clonality analysis of multifocal bladder tumors bycombination of histopathologic organ mapping, loss of heterozygosity,fluorescence in situ hybridization, and p53 analyses. 2006;37:143-51.

[32] Cairns P, Proctor AJ, Knowles MA. Loss of heterozygosity at the RBlocus is frequent and correlates with muscle invasion in bladdercarcinoma. Oncogene 1991;6:2305-9.

[33] Cappellen D, Gil Diez de Medina S, Chopin D, Thiery JP, Radvanyi F.Frequent loss of heterozygosity on chromosome 10q in muscle-invasive transitional cell carcinomas of the bladder. Oncogene 1997;14:3059-66.

[34] Habuchi T, Devlin J, Elder PA, Knowles MA. Detailed deletionmapping of chromosome 9q in bladder cancer: evidence for twotumour suppressor loci. Oncogene 1995;11:1671-4.

[35] Habuchi T, Ogawa O, Kakehi Y, Ogura K, Koshiba M, Hamazaki S,et al. Accumulated allelic losses in the development of invasiveurothelial cancer. Int J Cancer 1993;53:579-84.

[36] Habuchi T, Ogawa O, Kakehi Y, Ogura K, Koshiba M, Sugiyama T,et al. Allelic loss of chromosome 17p in urothelial cancer: strongassociation with invasive phenotype. J Urol 1992;148:1595-9.

[37] Habuchi T, Yoshida O, Knowles MA. A novel candidate tumoursuppressor locus at 9q32-33 in bladder cancer: localization of thecandidate region within a single 840 kb YAC. HumMol Genet 1997;6:913-9.

[38] Kim MS, Jeong J, Majewski T, Kram A, Yoon DS, Zhang RD,et al. Evidence for alternative candidate genes near RB1 involvedin clonal expansion of in situ urothelial neoplasia. Lab Invest 2006;86:175-90.

[39] Majewski T, Lee S, Jeong J, Yoon DS, Kram A, Kim MS, et al.Understanding the development of human bladder cancer by using awhole-organ genomic mapping strategy. Lab Invest 2008;88:694-721.

[40] Obermann EC, Junker K, Stoehr R, Dietmaier W, Zaak D, Schubert J,et al. Frequent genetic alterations in flat urothelial hyperplasias andconcomitant papillary bladder cancer as detected by CGH, LOH, andFISH analyses. J Pathol 2003;199:50-7.

[41] Stoehr R, Hartmann A, Hiendlmeyer E, Murle K, Wieland W,Knuechel R. Oligoclonality of early lesions of the urothelium asdetermined by microdissection-supported genetic analysis. Pathobio-logy 2000;68:165-72.

[42] Hartmann A, Rosner U, Schlake G, Dietmaier W, Zaak D, HofstaedterF, et al. Clonality and genetic divergence in multifocal low-gradesuperficial urothelial carcinoma as determined by chromosome 9 andp53 deletion analysis. Lab Invest 2000;80:709-18.

[43] Vaish M, Mandhani A, Mittal RD, Mittal B. Microsatellite instabilityas prognostic marker in bladder tumors: a clinical significance. BMCUrol 2005;5:2.

[44] Catto JW, Azzouzi AR, Amira N, Rehman I, Feeley KM, Cross SS,et al. Distinct patterns of microsatellite instability are seen in tumoursof the urinary tract. Oncogene 2003;22:8699-706.

Documents

Coexistent intraurothelial carcinoma and muscle-invasive - Figshare