ESTABLISHING HUMAN PROSTATE CANCER CELL XENOGRAFTS IN BONE: INDUCTION OF OSTEOBLASTIC REACTION BY PROSTATE-SPECIFIC ANTIGEN-PRODUCING TUMORS IN ATHYMIC AND SCID/bg MICE USING LNCaP AND LINEAGE-DERIVED METASTATIC SUBLINES Tony T. WU 1 , Robert A. SIKES 1 , Quanjun CUI 2 , George N. THALMANN 4 , Chinghai KAO 1 , Cheryl F. MURPHY 3 , Hua Y ANG 1 , Haiyen E. ZHAU 1 , Gary BALIAN 2 and Leland W.K. CHUNG 1 * 1 Molecular Urology and Therapeutics Program, Department of Urology, University of Virginia Health Sciences Center, Charlottesville, VA, USA 2 Department of Orthopedics, University of Virginia Health Sciences Center, Charlottesville, VA, USA 3 Division of Hematology-Oncology, Department of Internal Medicine, University of Virginia, Charlottesville, VA, USA 4 Urologische Universita ¨ tsklinik, University of Bern, Bern, Switzerland LNCaP lineage-derived human prostate cancer cell lines C4-2 and C4-2B 4 acquire androgen independence and osseous metastatic potential in vivo. Using C4-2 and C4-2B 4 the goals of the current investigation were 1) to establish an ideal bone xenograft model for prostate cancer cellsin intact athymic or SCID/bg mice using an intraosseous route of tumor cell administration and 2) to compare prostate cancer metastasis by administering cells either through intravenous (i.v.) or intracardiac administration in athymic or SCID/bg mice. Subsequent to tumor cell administration, prostate cancer growth in the skeleton was assessed by radiographic bone density, serum prostate-specific antigen (PSA) levels, pres- ence of hematogenous prostate cancer cells and histopatho- logic evaluation of tumor specimens in the lymph node and skeleton. O ur resultsshow that whereasLN CaP cellsinjected intracardially failed to develop metastasis, C4-2 cells injected similarly had the highest metastatic capability in SCID/bg mice. Retroperitoneal and mediastinal lymph node metasta- ses were noted in 3/7 animals, whereas 2/7 animals developed osteoblastic spine metastases. Intracardiac injection of C4-2 in athymic hosts produced spinal metastases in 1/5 animals at 8–12 weeks post-injection; PC-3 injected intracardially also metastasized to the bone but yielded osteolytic responses. Intravenous injection of either LNCaP or C4-2 failed to establish tumor colonies. Intrailiac injection of C4-2 but not LNCaP nor C4-2B 4 cells in athymic mice established rapidly growing tumors in 4/8 animals at 2–7 weeks after inoculation. Intrafemoral injection of C4-2 (9/16) and C4-2B 4 (5/18) but not LNCaP (0/13) cells resulted in the development of osteoblastic bone lesions in athymic mice (mean: 6 weeks, range: 3–12 weeks). In SCID/bg mice, intrafemoral injection of LNCaP (6/8), C4-2 (8/8) and C4-2B 4 (8/8) cells formed PSA-producing, osteoblastic tumors in the bone marrow space within 3–5 weeks after tumor cell inoculation. A stepwise increase of serum PSA was detected in all animals. Reverse transcription-polymerase chain reaction (RT-PCR) to detect hematogenously disseminated prostate cancer cells could not be correlated to either serum PSA level or histologi- cal evidence of tumor cells in the marrow space. W e have thus established a PSA-producing and osteoblastic human prostate cancer xenograft model in mice. Int. J. Cancer 77:887–894, 1998.  1998 Wiley-Liss, Inc. Adenocarcinoma of the prostate has been recognized consis- tently as the leading cause of cancer and the second leading cause of cancer death in North American men (Boring et al., 1994). According to 1996 American Cancer Society statistics, the newly diagnosed cases of prostate cancer have risen to 334,500 in the United States, and prostate cancer is the most commonly diagnosed cancer and the second leading cause of cancer deaths in men (Parker et al., 1997). Because of increasing public awareness and screening programs, most of the recently diagnosed prostate cancer patients are clinical stage T1c or organ-confined (Catalona et al., 1991, 1993; Stormont et al., 1993). However, a certain percentage of patients present with disease beyond the confines of the prostate (Catalona et al., 1991). Among these cases, disease progression to distant metastasis is almost unavoidable despite local or systemic therapies (Scher and Chung, 1994). Of those that metastasize, the bone is the most common and debilitating site (Scher and Chung, 1994; Stamey and McNeal, 1992). Prostate cancer is known to grow and induce osteoblastic as well as osteolytic responses when it harbors at the bone marrow space (Frank, 1997). A survey of the literature shows that several human prostate cancer cell lines when injected subcutaneously (s.c.) (Thalmann et al., 1994), orthotopically (Thalmann et al., 1994), intravenously (i.v.) with/without inferior vena cava occlusion (Shevrin et al., 1988; Wang and Stearns, 1991) or intraosseously (Berlin et al., 1993; Soos et al., 1996) induce skeletal metastases or intraosseous tumor growth. A fresh human prostate cancer xeno- graft, when implanted s.c. in SCID mice, has developed microme- tastases in the bone marrow as detected by reverse transcription- polymerase chain reaction (RT-PCR) (Klein et al., 1997). The LNCaP human prostate cancer model described by Thalmann et al. (1994) closely mimics the clinical prostate cancer progression in which a serum marker, prostate-specific antigen (PSA), was elaborated and the tumors progress from androgen dependence to androgen independence with substantial osteoblastic reactions associated with skeletal metastases. Despite the availability of previously described human prostate cancer metastatic models (Klein et al., 1997; Thalmann et al., 1994), there remain significant limitations of each model system. The LNCaP human prostate cancer model (Thalmann et al., 1994) is limited by a lengthy latent period between injection and detectable osseous metastasis. Intra- venous and intraosseous models of PC-3 tumor growth and bone metastasis in immunodeficient animals generate exclusively osteo- lytic responses without reliable means to evaluate tumor volume using serum markers (Shevrin et al., 1988; Wang and Stearns, 1991). While the primary tissue xenograft model results in PSA-producing tumors (Klein et al., 1997), the extent of metastasis is limited and there is no evidence of metastasis-related osteoblas- tic reaction. To improve human prostate cancer metastatic models, this study and others (Pettaway et al., 1996; Sato et al., 1997) used athymic and/or SCID/bg mice as hosts as well as to study interactions Grant sponsor: National Institutes of Health; Grant numbers: CA-63341, CA-64863 and DK-47596. Tony T. Wu and Robert A. Sikes contributed equally to this work. *Correspondence to: Molecular Urology and Therapeutics Program, Department of Urology, Box 422, University of Virginia Health Sciences Center, Charlottesville, VA 22908, USA. Received 14 January 1998; Revised 3 March 1998 Int. J. Cancer: 77, 887–894 (1998)  1998 Wiley-Liss, Inc. Publication of the International Union Against Cancer Publication de l’Union Internationale Contre le Cancer