SENSITIVE DETECTION OF TRANSITIONAL CELL CARCINOMA OF THE BLADDER BY MICROSATELLITE ANALYSIS OF CELLS EXFOLIATED IN URINE Davide SERIPA 1 , Paola PARRELLA 2,3 , Michele GALLUCCI 4,5 , Carolina GRAVINA 1 , Sara PAPA 1 , Pasquale FORTUNATO 5 , Antonio ALCINI 4 , Gerardo FLAMMIA 4 , Marzia LAZZARI 6 and Vito M. FAZIO 1,2 * 1 Unita ` Patologia Molecolare e Terapia Genica, IRCCS H. “Casa Sollievo Sofferenza”, Opera Padre Pio da Pietrelcina, San Giovanni Rotondo, Italy 2 Laboratory of Molecular Medicine and BioTechnology, Universita ` Campus Bio-Medico, Rome, Italy 3 Department of Otolaryngology—Head and Neck Surgery, Division of Head and Neck Cancer Research, Johns Hopkins University School of Medicine, Baltimore, MD, USA. 4 Department of Urology, Campus Bio-Medico University, Rome, Italy 5 Department of Urology, “Cristo Re” Hospital, Rome, Italy 6 Department of Surgery, Universita ` di Roma “Tor Vergata”, Rome, Italy Transitional cell carcinoma (TCC) is the most common bladder tumor. Urine cytology can identify most high-grade tumors but sensitivity is lower if one includes lesions of all grades. Microsatellite marker alterations have been found in many tumor types including bladder cancer and have been used to detect cancer cells in body fluids including urine. The aim of our study is to further evaluate feasibility and sensi- tivity of microsatellite analysis to detect bladder cancer cells in urine. We studied 55 individuals: 21 with symptoms sug- gestive of bladder cancer, 23 patients with previous history of TCC and 11 healthy subjects. Genomic DNA was extracted from blood lymphocytes, urine sediment, bladder washings and tumor or normal bladder mucosa. Twenty highly infor- mative microsatellite markers were analyzed for loss of het- erozigosity (LOH) and microsatellite instability (MIN) by polymerase chain reaction. Microsatellite analysis of urine identified 33 of 34 (97%) patients with either primary or tumor recurrence, whereas urine cytology identified 27 of 34 (79%) patients (p  0.0001). Detection of microsatellite ab- normalities improved the sensitivity of detecting low-grade and/or stage bladder tumor: from 75–95% for grades G1–G2 and from 75–100% for pTis–pTa tumors. Bladder washings from 25 patients were also analyzed, and in all cases results were identical to those obtained from voided urine. None of the 16 patients without evidence of TCC showed LOH and/or MIN in urine samples or bladder washings. Interestingly, in a patient with persistent bladder mucosa abnormalities, mic- rosatellite alterations were demonstrated 8 months before the histopathologic diagnosis of tumor recurrence. These results further indicate that microsatellite marker analysis is more sensitive than conventional urine cytology in detecting bladder cancer cells in urine and represents a potential clin- ical tool for monitoring patients with low-grade/stage TCC. © 2001 Wiley-Liss, Inc. Key words: genomic instability; tumor progression; diagnosis; prog- nosis Urinary bladder cancers represent the 4th most common malig- nant neoplasia in men and the 9th in women, and the annual incidence is approximately 18 cases per 100,000 in the United States. 1,2 Transitional cell carcinoma (TCC) comprises 90 –95% of bladder tumors, and about 80% of newly diagnosed TCC are confined to superficial mucosa. 3 The risk of recurrence in patients affected by TCC is approximately 70%, and progression to a higher grade and/or stage is estimated from 15–30% of recurrent tumors 3,4 Cystoscopy remains the most accurate method for as- sessing the bladder urothelium but it is an invasive, uncomfortable and expensive procedure. Moreover, small tumors and carcinoma in situ can be missed by cystoscopic examination. The most widely used noninvasive method for detecting bladder tumors is urine cytology. 5,6 This method identifies approximately 80 –90% of high-grade tumor, 7 but sensitivity varies from 50 –70% if all grade lesions are included. 7,8 Cytologic analysis of cells exfoliated in bladder washing fluids, although invasive, is preferable to voided urine samples because of their higher cellular content. 9 It is widely accepted that human neoplasia arise from multiple independent genetic changes that activate proto-oncogenes or in- activate tumor-suppressor genes. 10 Genetic alterations that arise during tumorigenesis have been used for detecting cancer cells in tumor specimens. In addition to specific mutations in oncogenes and tumor-suppressor genes, loss of heterozygosity (LOH) and/or microsatellite instability (MIN) have been reported in many tu- mors, such as colon cancer, 11–13 breast cancer, 14 –16 lung can- cer, 17–19 head and neck cancer, 20 renal tumors 21 and cutaneous 22 and uveal melanoma. 23 The frequency of LOH at specific micro- satellite loci suggests the involvement of genes related to carcino- genesis or tumor progression. However, LOH also occurs in DNA noncoding region, thus microsatellites are likely hot spots for mutagenesis, and mutations within this sequence can be used to detect clonal evolution of neoplastic cells. 24 At least 50% of bladder tumors have lost parts of chromosome 9. 25 Chromosome 9 appears to be involved early in the tumorigenesis as LOH of both chromosomal arms are independent either of stage or grade. 26 –30 Invasive properties appear to be associated with LOH of 11p, 31,32 3p, 33 13q 34 and 17p. 35–37 Finally, a significant association between high tumor grade and stage has been found for chromosome 8 38,39 and chromosome 4. 40 In several studies, microsatellite alterations have been used as clonal markers of neoplasia to detect cancer cells in different body fluids, such as sputum, 41 plasma 41,42 and serum. 43,44 Mao 45 and Steiner 46 identified bladder cancer by microsatellite analysis of DNA extracted from voided urine in 19 of 20 patients with primary TCC and 10 of 11 patients with disease recurrence, respectively. In Grant sponsor: Ministry of Health of Italy, IRCCS “Casa Sollievo della Sofferenza,” San Giovanni Rotondo (FG), Italy; Grant number: RC98; Grant sponsor: “Progetto Sanit` a 96/97, Fondazione Cassa di Risparmio di Verona, Vicenza, Belluno, ed Ancona”; Grant sponsor: University Campus BioMedico, Rome, Italy. The first two authors contributed equally to this work. *Correspondence to: Molecular Pathology and Gene Therapy Unit, IRCCS-II, “Casa Sollievo della Sofferenza,” Viale Cappuccini, I, San Giovanni Rotondo (FG), I-71013, Italy. Fax: +39-882-410575 or +39-882-410794. E-mail: terapiagenica@operapadrepio.it Received 28 February 2001; Revised 5 June 2001; Accepted 18 June 2001 Published online 00 Month 2001 Int. J. Cancer (Pred. Oncol.): 95, 364 –369 (2001) © 2001 Wiley-Liss, Inc. Publication of the International Union Against Cancer