ABSTRACTS

Inhibitors of Urokinase Reduce Sizeof Prostate Cancer Xenografts inSevere CombinedImmunodeficient Mice

Jerzy Jankun, Rick W. Keck, EwaSkrzypczak-Jankun, and Rafal Swiercz

From the Departments of Urology[J.J., R.W.K.] and Physiology and Mo-lecular Medicine [J.J., R.S.], MedicalCollege of Ohio, Toledo, Ohio 43699-0008, and Department of Chemistry,The University of Toledo, Toledo, Ohio43606-3390 [E.S-J.].

Cancer Res 57:599–563, 1997.Proteolytic enzymes are required

to mediate tumor cell invasion and me-tastasis. The urokinase plasminogenactivator (uPA) is commonly overex-pressed by many human cancers.Therefore, uPA is a logical target to in-hibit cancer invasion and metastasis.However, uPA inhibitors also reducetumor growth. We used a mutatedform of plasminogen activator inhibi-tor type 1 to conform a correlation be-tween the inactivation of uPA and tu-mor size; we have compared theseresults with the action of p-aminobenz-amidine and amiloride, known inhibi-tors of uPA. Our results show thatblocking uPA by uPA inhibitors re-duces tumor size in experimental ani-mals. Our molecular simulation ofdocking inhibitors to the urokinase re-veals that all tested small molecule in-hibitors bind in proximity of uPA’sspecificity pocket, a critical site for fu-ture search of novel anticancer uPA in-hibitors. (Reproduced with permission ofthe American Association for Cancer Re-search, Inc.)

A Polymorphism of the5a-Reductase Gene and ItsAssociation with Prostate Cancer:A Case-Control Analysis

Philip W. Kantoff, Phillip G.Febbo, Edward Giovannucci, KrishnaKrithivas, Douglas M. Dahl, GloriaChang, Charles H. Hennekens, MylesBrown, and Meir J. Stampfer

From the Division of Solid Tu-mors, Department of Adult Oncology,Dana-Farber Cancer Institute, HarvardMedical School [P.W.K., P.G.F., K.K.,D.M.D., G.C., M.B.]; Division of Pre-

ventive Medicine [C.H.H.] and Chan-ning Laboratory [E.G., M.J.S.] and Di-vision of Urology [D.M.D.], Brighamand Women’s Hospital and HarvardMedical School; and Departments ofEpidemiology [C.H.H., M.J.S.] and Nu-trition [E.G., M.J.S.], Harvard School ofPublic Health, Boston, Massachusetts02115

Cancer Epidemiol Biomarkers Prev 6:189–192, 1997.

Prostate cancer (CaP) is the mostcommonly diagnosed, nondermato-logical cancer in the United States. Thedevelopment and progression of CaP isinfluenced by androgens. 5a-Reduc-tase, type II, converts testosterone todihydrotestosterone and is critical tothe development of the prostate. A TAdinucleotide repeat polymorphism ex-ists in the 38 untranslated region of the5a-reductase type II gene. 5a-Reduc-tase alleles with longer TA repeats aremore common in African-Americans,the group with the highest incidence ofCaP. It has been hypothesized that thelonger TA repeat alleles might be asso-ciated with increased risk of CaP.

We studied this potential associa-tion within the Physician’s HealthStudy, a predominantly Caucasian co-hort study. Using PCR we identifiedthe TA genotype in 590 men with CaPand 802 age-matched controls. The fre-quency of each allele in the controlswas TA(0), 0.87, TA(9), 0.13, andTA(18), 0.01. Homozygotes for thelonger TA alleles, TA(9) and TA(18),were underrepresented among caseswith an odds ratio of 0.47 (confidenceinterval, 0.20–1.12), but this was notstatistically significant (P = 0.08, twotailed). Our analysis does not supportthe prior hypothesis that longer TA al-leles confer an increased risk of CaP ina predominantly Caucasian popula-tion; in fact, longer TA alleles are moreprevalent in men without CaP. (Repro-duced with permission of American Asso-ciation for Cancer Research, Inc.)

Generation and GeneticCharacterization of ImmortalHuman Prostate Epithelial CellLines Derived from PrimaryCancer Specimens

Robert K. Bright, Cathy D. Vocke,Michael R. Emmert-Buck, Paul H. Du-ray, Diane Solomon, Patricia Fetsch,

Johng S. Rhim, W. Marston Linehan,and Suzanne L. Topalian

From the Surgery Branch [R.K.B.,C.D.V., W.M.L., S.L.T.], Laboratory ofPathology [M.R.E-B., P.H.D., D.S., P.F.]and the Laboratory of Molecular On-cology [J.S.R.], National Cancer Insti-tute, NIH, Bethesda, Maryland 20892

Cancer Res 57:995–1002, 1997.Difficulty in establishing long-

term human prostate epithelial celllines has impeded efforts to under-stand prostate tumorigenesis and todevelop alternative therapies for pros-tate cancer. In the current study, we de-scribe a method that was successful ingenerating 14 immortal benign or ma-lignant prostate epithelial cell culturesfrom primary adenocarcinomas of theprostate resected from six successivepatients. Immortalization with the E6and E7 transforming proteins of hu-man papilloma virus serotype 16 wasnecessary to establish long-term cul-tures. Microscopic examination of freshtumor specimens exhibited a variablemixture of benign and malignant epi-thelium. Thus, single-cell cloning of tu-mor-derived cell cultures was essentialfor defining tumor cell lines. Efforts tocharacterize these cultures using tradi-tional criteria such as karyotype,growth in nude mice, and prostate-specific antigen expression were non-informative. However, allelic loss ofheterozygosity (LOH) represents apowerful alternative method for char-acterizing tumor cell lines originatingfrom primary adenocarcinomas of theprostate. Microdissected fresh tumorsfrom four of six patients revealed LOHat multiple loci on chromosome 8p, asassessed by PCR. LOH on chromosome8p matching the patterns found in mi-crodissected tumors was also observedin a tumor-derived cell line and itsclones, as well as in one clone from atumor-derived cell line from a secondpatient. LOH was not observed in im-mortal lines generated from autolo-gous benign prostatic epithelium,seminal vesicle epithelium, or fibro-blasts. The multifocal nature of pros-tate cancer, as well as the presence ofan entire spectrum of malignant trans-formation within individual prostateglands, necessitates this type of carefulanalysis of derivative cell cultures fortheir validation as in vitro models thataccurately reflect the primary cancers

The Prostate 32:75–76 (1997)

from which they are derived. (Repro-duced with permission of American Asso-ciation for Cancer Research, Inc.)

Prostate-Specific Antigen (PSA) andPSA Density: Racial Differences inMen Without Prostate Cancer

R. Jonathan Henderson, James A.Eastham, Daniel J. Culkin, Michael W.Kattan, Terence Whatley, John Mata,Dennis Venable, Oliver Sartor

From the Department of Urology[R.J.H., J.A.E., D.J.C., T.W., J.M., D.V.],Louisiana State University MedicalCenter, Shreveport, and Overton-Brooks Veterans Affairs Medical Cen-ter, Shreveport; Departments of Urol-ogy and Medicine [O.S.], LouisianaState University Medical Center andOverton-Brooks Veterans AffairsMedical Center; Information Technol-ogy Program [M.W.K.], Baylor Collegeof Medicine, Houston, TX.

J Natl Cancer Inst 89:134–138, 1997.Background: Many physicians

now use serum prostate-specific anti-gen (PSA) to screen for prostate cancerin asymptomatic men. Whether or nota prostate biopsy should also be per-formed depends on an accurate defini-tion of what constitutes a normal PSAvalue. Until recently, studies con-ducted to establish normal serum PSAvalues have involved study popula-tions that have included few African-American men.

Purpose: We sought to compareserum PSA levels and PSA density (i.e.,serum PSA level/prostate volume ra-tio) in African-American and whitemen without histologic evidence ofprostate cancer.

Methods: We reviewed the medi-cal records of 826 consecutive men whounderwent one or more prostate biop-sies at the Veterans Affairs MedicalCenter in Shreveport, LA, from Janu-ary 1993 through December 1995. Inthis retrospective review, we recordedpatient’s age, race, serum PSA level,digital rectal examination result, ultra-sound-determined prostate volume, in-dications for biopsy, and biopsy re-sults. Data from a total of 752 consecu-tive men who were either white orAfrican-American and whose indica-tion for biopsy included a serum PSAof greater than 4.0 ng/mL and/or an

abnormal digital rectal examinationwere analyzed. To examine possibledifferences in serum PSA level, PSAdensity, prostate volume, and patientage, the two-sided Student’s t test wasemployed. Multivariate linear regres-sion analysis was used to determine ifserum PSA levels were associated withthe patient’s age, race, or prostate vol-ume in men without prostate cancer.

Results: Of the 752 men includedin this analysis, 254 had histologic evi-dence of prostate cancer and 498 didnot. Of the 498 men without prostatecancer, 367 (74%) men were white and131 (26%) were black. There were noracial differences in age or calculatedprostate volume. Serum PSA levels andcalculated PSA density, however, weresignificantly (both P<.0001) higher inAfrican-American men than in whitemen. A multivariate linear regressionanalysis indicated that race and pros-tate volume were independent vari-ables associated with serum PSA level.For African-American and white men,serum PSA values of greater than 4 ng/mL were associated with prostate can-cer with sensitivities of 89.5% and81.9%, respectively, and specificities of38.2% and 52.3%, respectively.

Conclusion: Among biopsied menwithout histologic evidence of prostatecancer, African-Americans have a sig-nificantly higher PSA level and PSAdensity than similarly aged white men.

Implications: Published criteriafor normal PSA level and density havebeen derived primarily from whitemen and may not be directly applicableto other populations. Race-specific dataare needed to fully optimize PSA as atumor marker in racial populationsthat are at high risk for prostate cancerdeath. (Reproduced with permission ofOxford University Press.)

CD44 Is a Metastasis SuppressorGene for Prostatic Cancer Locatedon Human Chromosome 11p13

Allen Chuan Gao, Wei Lou, Jin-Tang Dong, and John T. Isaacs

From the Johns Hopkins OncologyCenter, James Buchanan Brady Uro-logical Institute, Department of Urol-ogy, The Johns Hopkins UniversitySchool of Medicine, Baltimore, Mary-land 21231 [A.C.G., W.L., J.T.I.], and

Department of Pathology, Universityof Virginia, Charlottesville, Virginia22908 [J-T.D.]

Cancer Res 57:846–849, 1997.We have used microcell fusion-

mediated chromosomal transfer to in-troduce normal human chromosomesinto highly metastatic rodent prostaticcancer cells to map the location of ametastasis suppressor gene(s). Usingthis approach, several chromosomal re-gions have been identified that harborsuch metastatic suppressor genes, in-cluding human chromosome 11 be-tween p11.2–13 (T. Ichikawa et al., Can-cer Res., 52: 3486–3490, 1992, 54: 2299–2302, 1994; N. Nihei et al., GenesChromosomes & Cancer, 14: 112–119,1995; C.W. Rinker-Schaeffer et al., Can-cer Res., 54: 6249–6256, 1994). Usingpositional cloning, a metastatic sup-pressor gene, termed KAI1, was identi-fied, which is located at human chro-mosome 11p11.2 (5). Overexpression ofKAI1 results in metastasis suppressionin certain highly metastatic DunningR-3327 rat prostatic cancer sublines,such as AT6.1, without metastasis sup-pression in other highly metastatic sub-lines, such as AT3.1. This suggests thatan additional metastasis suppressorgene is located within the human chro-mosome 11p11.2–13 region. The CD44gene is located on human chromosome11p13 and encodes an integral mem-brane glycoprotein that participates inspecific cell-cell and cell-extracellularmatrix interactions. Down-regulationof CD44 expression both at the mRNAand protein levels correlates withmetastatic potential within the Dun-ning system of rat prostatic cancer sub-lines. Transfection-induced enhancedexpression of the Mr 85,000 standardform of CD44 in the highly metastaticAT3.1 rat prostatic cells greatly sup-presses their metastatic ability to thelungs without suppression of their invivo growth rate or tumorigenicity.These results suggest that CD44 is ametastasis suppressor for prostatic can-cer and that decreased expression ofthe standard form of CD44 is involvedin the progression of prostatic cancer toa metastatic state. (Reproduced with per-mission of the American Association forCancer Research, Inc.)

76 Abstracts