Abstract The retinoblastoma (Rb) gene was the first tu-mor suppressor gene to be discovered; however, data onthe influence of Rb inactivation on endometrial carcino-genesis are scarce. We investigated 46 paired primaryhuman endometrial carcinomas and normal tissues to as-sess the frequency of loss of heterozygosity (LOH) in Rband 20 tumor pairs to detect the frequency of p53 LOH.Moreover, expression of the retinoblastoma protein(pRb) was assessed immunohistochemically. Of 44 in-formative cases 8 showed loss of one allele in at leastone Rb marker; Rb LOH frequency thus reached 18%.Two omental metastases of endometrial origin showed aheterogeneity pattern similar to that of the primary tu-mors. We did not find a significant correlation betweenRb LOH and patient age, clinical stage, histologicalgrade or muscle invasion of the tumor. Nevertheless, RbLOH was demonstrated at early (stage I, 5/27, 18%) andadvanced (stages II–IV; 3/9, 33%) clinical stages of theneoplasm, suggesting that LOH at the Rb locus occursbefore the clonal expansion of the tumor. There was asignificant correlation between Rb LOH and weak/absentpRb expression. We noted a single case of p53 LOH atintron 1, but no tumor showed both alterations simulta-neously. Our data suggest that LOH at the Rb locus playsa role in the oncogenesis of a subset of uterine neo-plasms and corresponds with the altered expression ofthe pRb.

Keywords Rb gene · Endometrial cancer · p53 gene ·Loss of heterozygosity · Retinoblastoma protein

Introduction

Development of human cancer results from the accumu-lation of various genetic alterations within oncogenes,tumor suppressor genes (TSG), and DNA repair genes[10]. Multiple genetic abnormalities demonstrated inneoplastic cells led to the concept of a “mutator pheno-type” in cancer, formulated to explain the disparity of thelack of mutations reported in normal cells and the largenumber of genetic abnormalities in various human ma-lignancies [18]. Although genetic alterations may occurrandomly throughout the whole genome, they are prefer-ably reported in tumor suppressor genes functioning innormal cells maintaining the stability of the genome.

The retinoblastoma gene (Rb) was the first TSG to beidentified in humans, representing the classic model forthe recessive tumor suppressor gene, in which both pater-nal and maternal alleles must be inactivated for tumor de-velopment [15]. This TSG was found on human chromo-some 13, band q14 [8, 17]. The Rb gene consists of 27exons, ranging in size from 31 to 1873 bp, and 26 introns,ranging in size from 80 to 70,500 bp [19]. It encodes a110-kDa nuclear phosphoprotein (pRb) which partici-pates in the transcriptional control mechanisms mediatingprogression through the G1 phase of the cell cycle [10].Germline mutations of the retinoblastoma gene in partic-ular are predisposed to cause hereditary malignant eye tu-mors in children under 4 years of age [4, 40]. Somaticpoint mutations have been detected throughout the Rbgene, with exons 3, 8, 18, and 19 preferably altered [10].The fact that some authors have reported the occurrenceof Rb gene abnormalities in many retinoblastoma and os-teosarcoma cases [11, 35, 39] suggests that Rb alterationsplay a crucial role in the development and progression ofthese neoplasms. In addition, Rb inactivation has alsobeen detected in a subset of small-cell lung, bladder, andesophageal carcinomas [3, 23, 41].

A. Semczuk (✉) · J.A. JakowickiIInd Department of Gynecology, Lublin University School of Medicine, 8 Jaczewski St. 20-954,Lublin, Polande-mail: tomek@asklepios.am.lublin.plTel.: +48-81-7425426, Fax: +48-81-7475710

B. Marzec · J. WojcierowskiDepartment of Human Genetics, Lublin Universi-ty School of Medicine, 8 Jaczewski St. 20-954, Lublin, Poland

A. Roessner · R. Schneider-StockDepartment of Pathology, Otto-von-Guericke University, Magdeburg, Germany

Virchows Arch (2002) 441:577–583DOI 10.1007/s00428-002-0695-9

O R I G I N A L A RT I C L E

Andrzej Semczuk · Barbara MarzecAlbert Roessner · Jerzy A. JakowickiJacek Wojcierowski · Regine Schneider-Stock

Loss of heterozygosity of the retinoblastoma gene is correlatedwith the altered pRb expression in human endometrial cancer

Received: 18 April 2002 / Accepted: 2 July 2002 / Published online: 25 September 2002© Springer-Verlag 2002

However, owing to the relatively large size of thegene, various molecular biological techniques [Southernblotting, polymerase chain reaction (PCR), single-strandconformation polymorphism, direct sequencing, or lossof heterozygosity (LOH) testing] have been applied toguarantee a precise analysis of structural Rb gene rear-rangements in different human cancers. Studying the roleof Rb gene alterations, Cairns et al. [3] and Xing et al.[41] have previously reported a frequent LOH at the spe-cific 13q14 intragenic loci in human malignant tumors.To our knowledge, only a limited number of studies haveassessed the role of Rb TSG alterations in carcinomas ofthe female genital tract [5, 14], particularly in endometri-al malignant tumors [6, 12, 22, 28, 43]. Moreover, the re-lationship between Rb gene inactivation and the clinico-pathological variables of endometrial cancer has not yetbeen investigated (Medline database).

Thus, using PCR analysis we extended this study toassess the frequency of LOH at four polymorphic mark-ers at the Rb gene in 48 (46 primary and 2 metastatic)carcinomas of the uterine corpus. In addition, we exam-ined the correlation of genetic findings with Rb proteinimmunostaining and with the clinicopathological fea-tures of cancer (patient age, clinical stage, histologicalgrade, histological type, depth of myometrial invasion).We also investigated 20 primary endometrial neoplasmsfor the frequency of LOH at the p53 tumor suppres-sor gene by applying two polymorphic markers: vari-able number of tandem repeats (VNTR) at intron 1 andrestriction fragment length polymorphism (RFLP) at exon 4.

Materials and methods

Patients

We investigated 46 consecutive endometrial cancer patients whohad undergone surgery (total abdominal hysterectomy/bilateralsalpino-oophorectomy) at the IInd Department of Gynecology,Lublin University School of Medicine, Lublin, Poland, between1998 and 2001. Pelvic and para-aortic lymph node dissectionswere carried out when material obtained at dilatation and curettagewas diagnosed as nonendometrioid or poorly differentiated cancer,or when the neoplasm was disseminated outside the uterus at sur-gery. Two omental metastases of endometrial origin were also in-

cluded. For LOH analysis nonneoplastic reference material wasavailable in all cases. The patients had not received chemotherapy,radiotherapy, or hormonal therapy before surgery.

The clinical stage of disease was classified according to thestaging system of the Federation International of Gynecology andObstetrics [21]. Regarding clinical stage, 34 (74%) carcinomaswere of stage I, 7 (15%) were of stage II, 4 (9%) were of stage III,and 1 (2%) was of stage IV. The material was assessed histologi-cally at the Department of Pathology, Lublin University School ofMedicine, and based on the WHO staging system [30]. Regardingthe histological type, 44 (96%) were endometrial endometrioidcarcinomas and 2 (4%) were adenosquamous carcinomas. Twenty(45%) were well-differentiated (G1), 16 (35%) were moderatelydifferentiated (G2), and 10 (20%) were poorly differentiated (G3)endometrial carcinomas. Myometrial invasion was absent or pres-ent, but did not exceed half of the myometrial wall in 28 (61%)cases, while cancer infiltrated more than 50% of the myometriumin 18 (39%) cases.

Lymph node dissection was carried out in 29 patients (63%), ofwhom only one (3.4%; a 72-year-old women with stage III endo-metrial endometrioid adenocarcinoma) showed pelvic lymph nodemetastasis.

DNA isolation

Tissue obtained at surgery was subdivided into two parts. One por-tion was fixed in buffered formalin (pH 7.4) for routine pathologi-cal assessment, while the second was immediately frozen in liquidnitrogen and stored in the deep-freezer. High molecular weightDNA was extracted from tumor and nontumor (muscle, skin, orleukocytes) materials according to a standard protocol [27].

Analysis of Rb LOH

We used four intragenic polymorphic Rb markers, three RFLPs inintrons 4, 17, and 25, and one VNTR polymorphism localized inintron 20 of the gene [39, 43]. Table 1 summarizes the primer se-quence, endonucleases, and annealing temperature used in the ex-periments. All restriction enzymes were purchased from MBI Fer-mentas (Lithuania). PCR reaction was carried out in an automatedthermocycler (Perkin Elmer 2400, USA) in 50 µl, and consisted of250 ng DNA, 20 mM Tris-HCl (pH 8.3), 1.5 mM MgCl2, 200 µMof each deoxyribonucleoside triphosphate, 25 pmol of each prim-er, and 2 U Taq DNA polymerase (MBI Fermentas). An amplifica-tion protocol was carried out in the following manner: 95°C for5 min, followed by 35 cycles at 94°C for 30 s-60 s, at 52°C or57°C for 30–60 s for annealing (Table 1), and at 72°C for 30–60 s.Finally, one additional cycle of elongation at 72°C for 7 min com-pleted the reaction. As a control, 10 µl of the PCR products wassubjected to electrophoresis in 2% agarose gels, stained withethidium bromide, and UV-photographed. For detecting LOH atintrons 4, 17, and 25, PCR products were digested with the appro-

578

Table 1 Primer sequences determining the LOH at the Rb locus

Locus Restriction Product size Annealing Primer sequencesendonuclease (bp) temperature

Intron 4 EcoRI 270 (166/104)a 57°C 5′-TTGACCTAGATGAGATGTCT-3′5′-AGGACACAAACTGCTACCTC-3′

Intron 17 XbaI 945 (630/315)a 57°C 5′-TTCCAATGAAGAACAAATGG-3′5′-GCAATTGCACAATCCAAGTT-3′

Intron20 – 420–470 VNTR 57°C 5′-TGTATCGGCTAGCCTATCTC-3′5′-AATTAACAAGGTGTGGTGG-3′

Intron 25 DraI 167 (129/38)a 52°C 5′-TCCATTTATAAATACACATG-3′5′-TAACGAAAAGACTTCTTGCA-3′

a Enzymatic digestion results in PCR fragments which are shown in the parenthesis

priate endonuclease, run on native 8% polyacrylamide gels, cross-linked with piperazine diacrylamide, and visualized by a silverstaining method [2]. For detecting LOH at intron 20 of the Rbgene PCR products were directly subjected to polyacrylamidegels/silver staining technique.

Analysis of the LOH of p53

LOH was analyzed in intron 1 (VNTR) and exon 4 (RFLP) poly-morphic regions of the p53 TSG in 20 primary human endometrialcarcinomas. PCR amplification protocols have been described pre-viously [29]. RFLP analysis at exon 4 of p53 used cleavage of thePCR products by endonuclease BshI (AGS, Heidelberg, Germa-ny). The gels were silver-stained according to the protocol byBudowle et al. [2].

Detection of LOH

An allelic loss (LOH) was considered when one tumor allele dis-appeared or was reduced by more than 50% compared with con-trol DNA. Evaluation was performed visually and by densitometry(VDS, Pharmacia, Biotech, Germany) in ambiguous cases. To en-sure the reproducibility of the results each case showing LOH wasexamined twice by both PCR and electrophoresis.

Rb protein expression

pRb was immunohistochemically stained on 4-µm-thick sectionsusing a monoclonal mouse anti-human NCL-RB-1 IgM antibody(Novocastra, UK), diluted 1:100, as described previously with mi-nor modifications [31]. Briefly, we applied three cycles of micro-wave oven treatment in 750 W for 10 min each, the avidin-biotinperoxidase detection system (Vectastain ABC, Vector Lab., Bur-lingame, Calif., USA), and 3,3′′-diaminobenzidine tetrahydrochlo-ride containing 0.08% hydrogen peroxidase. Endometrial cancerslide showing enhanced glandular pRb expression served as a pos-itive control, while a negative control was carried out by incubat-ing the adequate sections with control IgM immunoglobulin (di-luted 1:100 from Dako, USA). pRb immunoreactivity was gradedas follows: (–) negative, less than 10% positive cells; (+) 10–50%positive cells; (++) more than 50% of the cells were positive. Theslides showed anti-Rb antibody reactivity with the endothelialcells (internal positive control [25]) incorporated in our study. Toensure the reproducibility of the results all slides showing weak ornegative retinoblastoma reactivity were processed twice.

Statistical analysis

The relationship of Rb LOH and pRb staining with patient age andthe clinicopathological features of cancer was analyzed usingFisher’s exact test or the χ2 test. The correlation of Rb LOH withpRb immunostaining was examined by Spearman’s rank correla-tion test. Statistica software for Windows version 5.1G (StatSoft,Tulsa, Okla., USA) was used for analysis, and a p value below0.05 was considered statistically significant.

Results

LOH at the Rb locus

In total, 44 of 46 primary cases showed heterozygosity inat least one of the polymorphic Rb markers used; the het-erozygosity rate thus reached 96%. The highest infor-mativity was observed at intron 17/XbaI (42 of 44;95%); the heterozygosity rates for intron 4/EcoRI, intron

20/VNTR, and intron 25/DraI were 84% (37 of 44), 84%(37 of 44), and 59% (26 of 44), respectively. Of 44 infor-mative cases 8 (18%) showed a reduction rate (>50%) orcomplete loss of one allele for at least one Rb marker(Table 2). Interestingly, LOH was detected in 2 of 37 tu-mors (for intron 4), in 3 of 42 cases (for intron 17), in 1of 37 tumors (for intron 20), and in 3 of 26 neoplasms(for intron 25; Fig. 1). Two omental metastases showedthe same Rb LOH pattern as did the primary tumors.

A poorly differentiated, stage IIa, endometrial endo-metrioid adenocarcinoma showed Rb LOH at two corre-sponding loci (located at introns 17 and 20) at the coreregion of the gene. However, we never found a completeRb gene loss in the study group, although most positivecases revealed partial losses at the Rb in the proximal ordistal regions of the gene (Table 2).

Rb dysfunction and clinicopathological variables of cancer

There was no relationship between Rb LOH and patientage, clinical stage, histological grade, or muscle invasionof the tumor (P>0.05; Fisher’s exact test). Rb LOH wasdetected in 5 of 23 (22%) lymph node negative endome-trial cancers; in one case LOH was reported for twomarkers simultaneously (at introns 17 and 20). Allelicloss was detected at intron 17/XbaI in two cases, at in-tron 20/VNTR in one case, and at intron 25/DraI in threeneoplasms. There was no Rb allelic loss in a lymph

579

Fig. 1 Examples of the Rb LOH (arrows) in carcinomas of theuterine corpus. N Normal DNA; T tumor DNA. Matching pairs ofnormal and neoplastic DNA were located adjacent to each other.An allelic lost was considered when there was a complete disap-pearance or more than a 50% reduction in the signal intensity ofneoplastic bands in comparison with the signals of the normalDNA

node-positive endometrial cancer showing extensive(++) nuclear pRb expression immunohistochemically.

pRB immunostaining in endometrial carcinomas

Immunohistochemical heterogeneous pRb staining wasconfined to the glandular cell nuclei, whereas some epi-thelial and endothelial cells also displayed extensive nu-clear expression. Of 46 cases 39 (85%) demonstrated ex-tensive (++) pRb immunoreactivity, with more than 50%of the neoplastic glandular cells being positive with anti-Rb antibody (Fig. 2). Staining was moderate (10–50%)in six cases (13%). One case of an endometrial adenocar-cinoma displayed lack of anti-Rb reactivity in the glan-

dular cancer cells, although endothelial cells containeddetectable Rb protein (positive internal control). pRb ex-pression was also correlated with the clinical and patho-logical variables of cancer, but none of parametersreached a statistically significant value (P>0.05; data notshown). It is worth pointing out that heterogeneous pRbexpression was reported in all hyperplastic lesions adja-cent to neoplastic endometrial glands (data not shown).Extensive (++) heterogeneous nuclear pRb reactivitywas also detected in two primarily endometrial carcino-mas that were not informative for all polymorphic Rbmarkers.

580

Table 2 LOH at the Rb TSGand pRb expression in 46 pri-mary human endometrial carci-nomas (Inf-abs informative–ab-sence of the LOH, Inf-pres in-formative–presence of theLOH, Noninf noninformative)

Case Rb LOH

Intron 4/EcoRI Intron 17/XbaI Intron 20/VNTR Intron 25/DraI pRb IHC

1 Inf-abs Inf-abs Inf-abs Inf-abs ++2 Inf-abs Inf-abs Inf-abs Noninf ++3 Inf-abs Inf-abs Inf-abs Noninf ++4 Inf-abs Inf-abs Inf-abs Noninf ++5 Inf-abs Inf-abs Inf-abs Inf-abs ++6 Inf-abs Inf-abs Noninf Noninf ++7 Inf-abs Inf-pres Noninf Noninf ++8 Noninf Noninf Inf-abs Inf-abs ++9 Inf-abs Inf-abs Inf-abs Noninf ++

10 Inf-abs Inf-abs Inf-abs Inf-pres ++11 Noninf Inf-abs Inf-abs Inf-abs ++12 Noninf Inf-abs Noninf Inf-abs ++13 Inf-abs Inf-abs Inf-abs Inf-abs ++14 Inf-abs Inf-abs Inf-abs Inf-abs ++15 Inf-abs Inf-abs Inf-abs Noninf ++16 Inf-abs Inf-abs Inf-abs Noninf ++17 Inf-abs Inf-pres Inf-pres Noninf +18 Inf-abs Inf-abs Inf-abs Noninf +19 Inf-abs Inf-abs Inf-abs Noninf +20 Inf-abs Inf-abs Inf-abs Inf-abs ++21 Inf-abs Inf-abs Inf-abs Inf-abs ++22 Inf-abs Inf-abs Inf-abs Inf-abs ++23 Inf-pres Inf-abs Inf-abs Noninf ++24 Inf-abs Inf-abs Inf-abs Noninf ++25 Inf-abs Inf-abs Inf-abs Inf-abs -26 Inf-abs Inf-abs Inf-abs Noninf ++27 Inf-abs Inf-abs Noninf Noninf ++28 Noninf Noninf Noninf Noninf ++29 Noninf Inf-abs Noninf Inf-pres ++30 Inf-abs Inf-abs Inf-abs Inf-abs ++31 Noninf Inf-abs Inf-abs Inf-abs +32 Inf-abs Inf-abs Noninf Inf-abs ++33 Inf-abs Inf-abs Inf-abs Noninf ++34 Inf-abs Inf-pres Noninf Inf-abs +35 Inf-abs Inf-abs Inf-abs Inf-abs ++36 Inf-abs Inf-abs Inf-abs Inf-abs +37 Inf-abs Inf-abs Inf-abs Inf-abs ++38 Noninf Inf-abs Inf-abs Inf-abs ++39 Inf-pres Inf-abs Inf-abs Noninf ++40 Inf-abs Inf-abs Inf-abs Inf-abs ++41 Inf-abs Inf-abs Inf-abs Inf-abs ++42 Inf-abs Noninf Inf-abs Inf-abs ++43 Noninf Noninf Noninf Noninf ++44 Inf-abs Inf-abs Inf-abs Inf-pres ++45 Inf-abs Inf-abs Inf-abs Noninf ++46 Noninf Inf-abs Inf-abs Noninf ++

581

Correlation between Rb LOH and pRb reactivity

pRB expression status was compared with the allelic sta-tus of the Rb gene in 44 informative uterine neoplasms.

In cases with intensive pRb nuclear staining Rb LOHwas detected in only 4 of 37 (11%) cancers, whereas 4 of8 (50%) LOH-positive tumors showed weaker or lack ofpRb expression (P=0.01; Fisher’s exact test). Thus RbLOH was also noted in four cases showing extensive nu-clear reactivity (Table 2). Nevertheless, a statisticallysignificant correlation was found between Rb LOH andweaker or absent pRb immunostaining in endometrialcarcinomas using Spearman’s rank correlation test(r=–0.439; P=0.0028).

LOH of the p53 gene

Heterozygosity for intron 1/VNTR and the exon4/BshI/RFLP marker was found in 19 (95%) and 9(45%) of 20 carcinomas, respectively. We noted a singlecase of p53 LOH at intron 1 (stage I well differentiatedendometrial endometrioid adenocarcinoma; Fig. 3).None of the informative cases displayed LOH at exon4/BshI marker. Thus only 1 of 19 (5%) informative casesshowed LOH at the p53 gene, but there was no caseshowing p53 and Rb alterations simultaneously.

Discussion

The Rb pathway consists of four components – retino-blastoma protein, cyclin D1, p16 protein, and cdk4 (cy-clin dependent kinase 4); the derailments of this path-way, which may occur at any level in the hierarchy, mayprovide increasing autonomy for cells undergoing cancerdevelopment [10, 20, 34]. Previously we reported nucle-ar Rb protein immunoreactivity in glandular cancer cellsin 95% (59 of 62) of endometrial tumors [31]. Further-more, we found that Rb and K-ras gene abnormalitiesprobably occur independently of each other during endo-metrial carcinogenesis [32]. Finally, we showed that lossof pRb expression, associated with higher MIB-1 prolif-erative activity, is implicated in the deregulation of thecell-cycle machinery in human uterine cancer [33]. Inthis study the allelic loss of the Rb gene was assessedand compared with altered pRb expression in 46 primaryendometrial carcinomas obtained from women of Polishnationality. We detected Rb LOH in 8 of 44 (18%) infor-

Fig. 2 Heterogeneous extensive (A), moderate (B), and the lackof (C) pRb nuclear immunoreactivity reported in the glandularcells of human endometrial cancer. Original magnification, ×400.pRb was stained on 4-µm-thick sections using a monoclonalmouse anti-human IgM, clone NCL-RB-1, antibody (diluted1:100). We applied the microwave oven retrieval technique withthe avidin-biotin peroxidase 3,3′′-diaminobenzidine tetrahydro-chloride visualization system (see text)

Fig. 3 Examples of genomic instability at intron 1/VNTR p53 lo-cus in carcinomas of the uterine corpus. N Normal DNA; T tumorDNA. Case 1 was not informative for this marker, while p53 LOHwas reported in case 2 (arrow)

mative cases, showing that allelic losses are randomlydistributed throughout the gene (Table 2). Allelic insta-bility was displayed in 9–20% within Rb in primary hu-man endometrial carcinomas, as reported in previouspublications [6, 36, 37, 43], in which only one or two in-tragenic polymorphic Rb markers were analyzed. Tritz etal. [37] reported that LOH at the Rb locus was higher in“usual” tumors (in 3 of 15 informative cases, 20%) ascompared with the prevalence of allelic loss detected in“special variant” uterine neoplasms (in 1 of 11 informa-tive cases, 9%). Sirchia et al. [36] have recently detecteda significant allelic loss (>20%) at the chromosomal re-gions 2q14, 7q35, 10q22.1, 11q13-q14, 15q26, 17p13,and 17q21.3, but only 12% of uterine neoplasms re-vealed LOH at chromosomal locus 13q14.1-q14.2. Thusthe frequency of Rb inactivation reported in this studywas apparently lower than that of LOH in retinoblasto-mas [11], osteosarcomas [39], and bladder [3], ovarian[5], and esophageal [41] carcinomas. In addition, therewas no concomitant allelic imbalance of the Rb and p53genes in endometrial cancers here, and the frequency ofgenetic instability at the p53 TSG reported in the litera-ture was higher (22–32%) than that noted here [13, 24,26]. However, owning to the limited number of cases in-vestigated, further studies on a large number of endome-trial tumors are necessary to ascertain the actual allelicstatus by applying a panel of particular markers spanningthe p53 gene.

Wada et al. [38], investigating Rb LOH by applyingseven polymorphic markers located between 13q11 and13q32 in 236 bladder carcinomas, noted a correlation be-tween allelic imbalance and stage and grade of tumors, afinding that was statistically significant for 13q14.1. RbLOH in epithelial ovarian carcinomas is reported to bemore common in invasive grade 3 and grade 4 cancersthan in invasive low-grade tumors (P<0.001) [5]. Also,LOH at the Rb locus was detected in 2 of 13 (15%) het-erozygous cases, both of which were stage II poorly dif-ferentiated endometrial carcinomas [6]. In the currentstudy, however, the prevalence of allelic imbalance at13q14 was not correlated with the clinicopathologicalfeatures of cancer, including patient age, clinical stage,histological grade, and depth of myometrial invasion.Nevertheless, LOH at 13q14 was demonstrated at early(stage I, 5/27; 18%) and advanced (stages II–IV, 3/9;33%) clinical stages of the neoplasm, suggesting that Rballelic imbalance occurs before the clonal expansion ofthe tumor. This evidence is also supported by the factthat 22% of lymph node negative endometrial cancers re-vealed LOH at the Rb locus.

Unlike the random prevalence of allelic imbalancewithout an obvious causal role, LOH, which is frequent-ly detected at specific loci, is used to investigate chro-mosomal regions containing TSG(s) involved in thepathogenesis of cancer. Indeed, the frequently detectedallelic imbalance at chromosomal locus 13q14, wherethe Rb gene is spanned, is associated with weaker oreven absent pRb staining, as detected by immunohisto-chemistry in several types of human malignant tumors

[9, 41, 42]. In contrast, some authors report a lack ofconcordance between LOH at 13q14 and pRb expres-sion in ovarian, prostate, and head and neck squamouscell carcinomas [5, 16, 44]. As a consequence (an)otherTSG(s), located close to the Rb 13q14 locus, is (are)mapped in this region [1, 7]. In this study one-half ofthe LOH-positive endometrial tumors showed weaker orabsent pRb expression, and the association between RbLOH and altered protein expression revealed a signifi-cant value (P=0.0028; Spearman’s rank correlation test).Thus, our results allow for the hypothesis that decreasedpRb staining corresponds with allelic imbalance at theRb locus in human endometrial carcinomas. However,the absence of weaker pRb expression without concomi-tant LOH at 13q14 implies that different genetic mecha-nisms (minor or major deletions within coding regions,point mutations, or hypermethylation of unmethylatedCpG islands within the promotor region of the gene)might also be responsible for the Rb gene rearrange-ments in a subset of human uterine cancers. On the oth-er hand, in four endometrial cancers with LOH at the Rblocus, normal pRb expression was identified immuno-histochemically. As an explanation, these tumors mayrepresent the hemizygous cases at the Rb locus, with aone allele being normal. Also, as suggested previouslyby Xu et al. [42], large tumor areas may contain Rb-pos-itive as well as Rb-negative tumor cells, and the areasused for immunohistochemical and molecular analysismay differ significantly in respect to the Rb status.Therefore further studies are necessary, particularly inthe pRb-positive endometrial carcinomas harboringLOH to ascertain the exact mechanism of Rb-genetic in-stability in those cases.

Acknowledgements The authors are grateful to the staff of theDepartment of Pathology, Lublin University School of Medicine,Lublin, Poland, for the histological assessment of the material. Wealso thank Mr. B. Wuesthoff for the careful editing of the manu-script, and Mrs. E. Firlej and Mrs. I. Jaskowska for their excellenttechnical assistance. A.S. was supported by a grant from the Deut-scher Akademischer Austauschdienst.

References

1. Brown AG, Ross FM, Dunne EM, Steel M, Weir-Thompson E(1993) Evidence of a new tumor suppressor locus (DBM) inhuman B-cell neoplasia telomeric to the retinoblastoma gene.Nat Genet 3:67–72

2. Budowle B, Chakraborty R, Giusti AM, Eisenberg AJ,Allen RC (1991) Analysis of the variable number of tandemrepeats locus D1S80 by the polymerase chain reaction fol-lowed by high resolution polyacrylamide gel electrophoresis.Am J Hum Genet 48:137–144

3. Cairns P, Proctor AJ, Knowles MA (1991) Loss of heterozy-gosity at the RB locus is frequent and correlates with muscleinvasion in bladder carcinoma. Oncogene 6:2305–2309

4. Cowell JK, Hogg A (1992) Genetics and cytogenetics of reti-noblastoma. Cancer Genet Cytogenet 64:1–11

5. Dodson MK, Cliby WA, Xu H-J, DeLacey KA, Hu SX, Keeney GL, Li J, Podratz KC, Jenkins RB, Benedict WF(1994) Evidence of functional RB protein in epithelial ovariancarcinomas despite loss of heterozygosity at the RB locus.Cancer Res 54:610–613

582

6. Enomoto T, Fujita M, Nakazawa-Miyamoto O, Tanizawa O,Nomura T (1993) Alterations of the Rb gene and its associa-tion with Ki-ras activation and p53 inactivation in endometrialadenocarcinoma. Mol Carcinog 8:132–137

7. Fiedler U, Ehlers W, Meye A, Fussel S, Faller G, Schmidt U,Wirth MP (2001) LOH analyses in the region of the putativetumour suppressor gene C13 on chromosome 13q13. Antican-cer Res 21:2341–2350

8. Friend SH, Bernards R, Rogelj S, Weinberg RA, Rapaport JM,Albert DM, Dryja TP (1986) A human DNA segment withproperties of the gene that predisposes to retinoblastoma andosteosarcoma. Nature 323:643–646

9. Gouyer V, Gazzeri S, Brambilla E, Bolon I, Moro D, Perron P,Benabid AL, Brambilla C (1994) Loss of heterozygosity at theRB locus correlates with loss of RB protein in primary malig-nant neuro-endocrine lung carcinomas. Int J Cancer 58:818–824

10. Hesketh R (ed) (1997) The oncogene and tumour suppressorgene. FactsBook. Academic, San Diego

11. Horowitz JM, Park S-H, Bogenmann E, Cheng JC, YandellDW, Kaye FJ, Minna JD, Dryja TP, Weinberg RA (1990) Fre-quent inactivation of the retinoblastoma anti-oncogene is re-stricted to a subset of human tumor cells. Proc Natl Acad SciUSA 87:2775–2779

12. Inoue M (1993) Genetic alterations and transformations in de-velopment and establishment of uterine endometrial carcino-mas. Nippon Sanka Fujinka Gakkai Zasshi 45:751–762

13. Kihana T, Hamada K, Inoue Y, Yano N, Iketani H, Murao S,Ukita M, Matsuura S (1995) Mutation and allelic loss of thep53 gene in endometrial carcinoma. Incidence and outcome in92 surgical patients. Cancer 76:72–78

14. Kim JW, Lee CG, Han SM, Kim KS, Kim JO, Lee JM,Kim IK, Namkoong SE (1997) Loss of heterozygosity of theretinoblastoma and p53 genes in primary cervical carcinomaswith human papillomavirus infection. Gynecol Oncol 67:215–221

15. Knudson AG (1993) Antioncogenes and human cancer. ProcNatl Acad Sci USA 90:10914–10921

16. Latil A, Bieche I, Pesche S, Volant A, Valeri A, Fournier G,Cussenot O, Lidereau R (1999) Loss of heterozygosity atchromosome arm 13q and RB1 status in human prostate can-cer. Hum Pathol 30:809–815

17. Lee W-H, Bookstein R, Hong F, Young L-J, Shew J-Y, LeeEY-HP (1987) Human retinoblastoma susceptibility gene:cloning, identification, and sequence. Science 235:1394–1399

18. Loeb LA (2001) A mutator phenotype in cancer. Cancer Res61:3230–3239

19. McGee TL, Yandell DW, Dryja TP (1989) Structure and par-tial genomic sequence of the human retinoblastoma suscepti-bility gene. Gene 80:119–128

20. Michalides RJAM (1999) Cell cycle regulators: mechanismsand their role in etiology, prognosis, and treatment in cancer.J Clin Pathol 52:555–568

21. Mikuta JJ (1995) Preoperative evaluation and staging of endo-metrial cancer. Cancer 76:2041–2043

22. Milde-Langosch K, Riethdorf L, Bamberger A-M, Loning T(1999) P16/MTS1 and pRB expression in endometrial carcino-mas. Virchows Arch 434:23–28

23. Mori N, Yokota J, Akiyama T, Sameshima Y, Okamoto A,Mizoguchi H, Toyoshima K, Sugimura T, Terada M (1990)Variable mutations of the RB gene in small-cell lung carcino-ma. Oncogene 5:1713–1717

24. Niederacher D, An HX, Camrath S, Dominik SI, Gohring UJ,Oertel A, Grass M, Hantschmann P, Lordnejad MR, Beck-mann MW (1998) Loss of heterozygosity of BRCA1, TP53and TCRD markers analysed in sporadic endometrial cancer.Eur J Cancer 34:1770–1776

25. Niemann TH, Yilmaz AG, McGaughy VR, Vaccarello L(1997) Retinoblastoma protein expression in endometrial hy-perplasia and carcinoma. Gynecol Oncol 65:232–236

26. Saegusa M, Okayasu I (1997) Bcl-2 is closely correlated withfavorable prognostic factors and inversely associated with p53protein accumulation in endometrial carcinomas: immunohis-

tochemical and polymerase chain reaction/loss of heterozygos-ity findings. J Cancer Res Clin Oncol 123:429–434

27. Sambrook J, Fritsch EF, Maniatis T (ed) (1989) Molecularcloning: a laboratory manual, 3rd edn. Cold Spring HarborLaboratory, Cold Spring Harbor

28. Sasano H, Comerford J, Silverberg SG, Garrett CT (1990) Ananalysis of abnormalities of the retinoblastoma gene in humanovarian and endometrial carcinoma. Cancer 66:2150–2154

29. Schneider-Stock R, Walter H, Radig K, Rys J, Bosse A, Kuhn-en C, Hoang-Vu C, Roessner A (1998) MDM 2 amplificationand loss of heterozygosity at Rb and p53 genes: no simulta-neous alterations in the oncogenesis of liposarcomas. J CancerRes Clin Oncol 124:532–540

30. Scully RE, Bonfiglio TA, Kurman RJ, Silverberg SG, Wilkinson EJ (ed) (1994) Histological typing of female genitaltract tumours, 2nd edn. Springer, Berlin Heidelberg New York

31. Semczuk A, Schneider-Stock R, Miturski R, Skomra D, Tomaszewski J, Roessner A, Jakowicki JA (2000) RB proteinexpression in human endometrial carcinomas – an immunohis-tochemical study. Pathol Res Pract 196:41–46

32. Semczuk A, Schneider-Stock R, Berbec H, Marzec B, Jakowicki JA, Roessner A (2001) K-ras exon 2 point muta-tions in human endometrial cancer. Cancer Lett 164:207–212

33. Semczuk A, Skomra D, Cybulski M, Jakowicki JA (2001) Im-munohistochemical analysis of MIB-1 proliferative activity inhuman endometrial cancer. Correlation with clinicopathologi-cal parameters, patient outcome, retinoblastoma immunoreac-tivity and K-ras codon 12 point mutations. Histochem J33:193–200

34. Sherr CJ (2000) The Pezcoller Lecture: cancer cell cycles re-visited. Cancer Res 60:3689–3695

35. Shimizu T, Toguchida J, Kato MV, Kaneko A, Ishizaki K,Sasaki MS (1994) Detection of mutations of the RB1 gene inretinoblastoma patients by using exon-by-exon PCR-SSCPanalysis. Am J Hum Genet 54:793–800

36. Sirchia SM, Sironi E, Grati FR, Serafini P, Garagiola I,Rossella F, Dulcetti F, Pardi G, Garsia S, Simoni G (2000)Losses of heterozygosity in endometrial adenocarcinomas:positive correlations with histopathological parameters. Can-cer Genet Cytogenet 121:156–162

37. Tritz D, Pieretti M, Turner S, Powell D (1997) Loss of hetero-zygosity in usual and special variant carcinomas of the endo-metrium. Hum Pathol 28:607–612

38. Wada T, Louhelainen J, Hemminki K, Adolfsson J, WijkstromH, Norming U, Borgstrom E, Hansson J, Sandstedt B, Stein-eck G (2000) Bladder cancer: allelic deletions at and aroundthe retinoblastoma tumor suppressor gene in relation to stageand grade. Clin Cancer Res 6:610–615

39. Wadayama B, Toguchida J, Shimizu T, Ishizaki K, Sasaki MS,Kotoura Y, Yamamuro T (1994) Mutations spectrum of the reti-noblastoma gene in osteosarcomas. Cancer Res 54:3042–3048

40. Wiggs J, Nordenskjold M, Yandell D, Rapaport J, Grondin V,Janson M, Werelius B, Petersen R, Craft A, Riedel K, LiberfarbR, Walton D, Wilson W, Dryja TP (1988) Prediction of the riskof hereditary retinoblastoma, using DNA polymorphisms with-in the retinoblastoma gene. N Engl J Med 318:151–157

41. Xing EP, Yang G-Y, Wang L-D, Shi ST, Yang CS (1999) Lossof heterozygosity of the Rb gene correlates with pRb proteinexpression and associates with p53 alteration in human esoph-ageal cancer. Clin Cancer Res 5:1231–1240

42. Xu H-J, Cairns P, Hu S-X, Knowles MA, Benedict WF (1993)Loss of RB protein expression in primary bladder cancer cor-relates with loss of heterozygosity at the RB locus and tumorprogression. Int J Cancer 53:781–784

43. Yaginuma Y, Katayama H, Kawai K, Duenas JC, Ishikawa M(1996) Analysis of the retinoblastoma gene in human endome-trial carcinoma. Obstet Gynecol 87:755–759

44. Yoo GH, Xu H-J, Brennan JA, Westra W, Hruban RH,Koch W, Benedict WF, Sidransky D (1994) Infrequent inacti-vation of the retinoblastoma gene despite frequent loss ofchromosome 13q in head and neck squamous cell carcinoma.Cancer Res 54:4603–4606

583