Patients with advanced CaP experience metastases, for which there is currently no curative treatment. The process of tumor metastasis is highly specific and consists of multiple steps. To metasta-

size successfully, the tumor cells must complete all steps of the process: initial transformation and growth, local invasion, survival of immune defense, adhesion to the secondary site, establishment of micrometastases and growth at secondary sites. The mechanisms involved in dissemination of tumor cells and establishment of metastases at secondary sites are not yet understood. Ideally, animal models of metastasis would parallel the human disease, but it has proven to be very difficult to develop models of spontaneous CaP metastasis that mimic human CaP in all aspects. To improve our understanding of CaP metastases and develop new treatment strategies, better animal models of human CaP demonstrating spontaneous metastasis from the orthotopic site are still needed.

The main sites of metastases of CaP are lymph nodes, bone, and lungs. Researchers have used available xenografts (Table 1) to generate lymph node and lung metastases, with varying results, but with very limited success in generating bone metastases. The following section is an overview of various approaches of generating spontaneous non-osseous and osseous metastases and experimental metastases with the most common CaP cells lines and the CaP xenografts developed in our laboratory.

5.1. Non-Osseous Metastasis Models

Standard procedures of growing xenografts have been used in attempts to generate spontaneous CaP metastases, and it has been found that, generally, metastasis from subcutaneous human CaP xenografts occurs very infrequently (75). Implantation of CaP cells into the prostate (orthotopic site) in mice has been used with more success, and has resulted in lymph node and lung metastases.

5.1.1. LNCaP Xenografts

Sato et al. (76) reported on generation of LNCaP lymph node metastases in 100% of animals and microscopic lung metastases in 40% of animals, using orthotopic injection of LNCaP cells. Metastases from LNCaP orthotopic tumors were also reported by Rembrink et al. (77) and Stephenson et al. (78). Pettaway et al. (79) also used orthotopic implantation of LNCaP cells in athymic mice to develop LNCaP sublines with increased metastatic potential; lymph node metastases were harvested and cells were re-injected into the prostate. This process was repeated three to five times. After implantation into the prostate, LNCaP-LN3 cells produced a higher incidence of regional lymph node metastases than LNCaP-Pro5 or LNCaP cells.

Recently, Scatena et al. (80) reported on generation of metastases from LNCaP-luc-M6 orthotopic and subcutaneous tumors using SCID-beige mice. In this study, sensitive luminescence imaging was used to detect the metastases. Primary tumors were shielded for this purpose. Lung metastases were detected as early as 7 weeks in 2 of 10 animals, and, at sacrifice, 10 of 10 of animals had lung metastases and 7 of 10 animals had rib metastases. Histology confirmed lung metastases but not rib metastases.

A different approach to generating metastases of LNCaP cells was used by Wang et al. (81). In this study, pieces of LNCaP xenografts grown subcutaneously were used for orthotopic implantation into the ventral lateral lobes of the prostate. The authors used this approach to minimize the possibility that lymph node metastases would be generated from leakage of the cells during injection rather than by a true metastatic process. With this approach, 61% of the animals had lymph node metastases and 44% had lung metastases.

5.1.2. PC-3 Xenografts

PC-3 CaP cells have been used extensively to generate metastases and develop cells with increased metastatic potential. Injection of PC-3 (79,82) via the tail vein has led to the establishment of lymph node metastases. Generation of lymph node metastases from PC-3 orthotopic tumors was also reported by Stephenson et al. (78) and Rembrink et al. (77). Rubio et al. (83,84) used intramuscular injection of PC-3 cells, resulting in metastases in animals. Pettaway et al. (79) used PC-3 cells in an attempt to generate sublines of these cells that were more metastatic, parallel to their work with

LNCaP. PC-3M cells were injected orthotopically into athymic mice, lymph node metastases were harvested, and cells were re-injected into the prostate. These experiments yielded PC-3M-Pro4 and PC-3M-LN4. PC-3M-LN4, harvested from a lymph node metastasis, produced more lymph node and bone metastases then cells harvested from prostate PC-3M-Pro4 or parental PC-3M cells.

An et al. (85) developed a microsurgical procedure of orthotopic implantation of histologically intact pieces of tumor in SCID mice, to generate reproducible and reliable models of spontaneous metastases of CaP in animals. Intact tissue of the human CaP cell line, PC-3, harvested from a subcutaneous tumor in an athymic mouse, was implanted into the ventral lateral lobes of the prostate gland. A high frequency of lymph node and lung metastasis was noted on histological examination. This was the first report in the literature of the generation of widespread lung metastases after orthotopic implantation of PC-3 cells. In contrast to orthotopic injection of cell suspensions, no multiple meta-static cell selection was necessary to increase the metastatic potential of PC-3. The authors speculated that the stromal tissue architecture maintained in the implanted tumors was important for metastatic potential.

Technical progress has permitted new methodologies to be developed to facilitate detection and quantitative measurements of metastatic spread. This issue is of special importance because the majority of spontaneous metastases of CaP are microscopic. Yang et al. (86) used bits of green fluorescent protein-labeled PC-3 tumors that had been implanted orthotopically in nude mice. Metastases were detected in various organs, including liver, lung, kidney, pleural membrane, and adrenal gland. Interestingly, this approach resulted in skeletal micrometastases as well. Luciferase-labeled PC-3 cells grown intramuscularly were also used to monitor trafficking of these cells to lymph nodes (83).

5.1.3. LuCaP 23.8 and LuCaP 35

We have described another model of generating CaP metastases, using the LuCaP 23.8 (48) and LuCaP 35 (27) CaP xenografts (60). In our efforts to characterize the metastatic potential of xenografts established in our laboratory, we observed small but visible metastases in pelvic lymph nodes of nearly half of the animals harboring various CaP xenografts, along with microscopic lung metastases. We wondered whether the lack of overt metastases in other organs was caused by poor dissemination, poor growth at metastatic sites, or an inadequate period for growth before sacrifice. We addressed this issue by removal of orthotopic tumors after they reached moderate size (250-500 mm3) to allow time for metastases to grow. The procedure is illustrated in Fig. 5. When we used this procedure, 71% of the LuCaP 23.8 animals and 100% of the LuCaP 35 animals had lymph node metastases. Lymph node metastases were macroscopic, and immunohistochemistry confirmed the prostatic origin of metastases. Seventy-one percent of the LuCaP 23.8 animals and 90% of the LuCaP 35 animals had lung metastases.

In summary, LNCaP, PC-3, and other prostate cells can be used to generate non-osseous metastases, including metastases in lymph nodes and lungs. The challenge of using these models to study mechanisms of metastasis or test new treatment modalities is still in the detection of metastases, which is labor intensive and time consuming. Using new imaging technologies with fluorescent or bioluminescent labeling will help to make evaluation of metastatic spread and effectiveness of new therapeutics more efficient.

5.2. Osseous Metastasis Model

Bone is a very common site of CaP metastasis (28,87,88), and bone metastases are responsible for most of the morbidity associated with the advanced disease. In contrast with bone metastases of breast cancer and myeloma, which are mainly osteolytic, a high percentage of CaP metastases exhibit the radiographic appearance of osteoblastic lesions (89). Histomorphometric studies of CaP bone metastases have shown that some of the sclerotic lesions are actually mixed in nature, with increased activities of both osteoblasts and osteoclasts (90,91).

Fig. 5. Scheme of implantation and removal of orthotopic prostate cancer (CaP) tumors. Tumor bits of CaP xenografts were implanted into the coagulating gland (A). After tumors reached a size of 250 to 500 mm3, the animals underwent a second surgery. Tumor and associated seminal vesicle were ligated and removed (B). SV, seminal vesicle; VD, vas deferens; U, ureter; B, bladder; AG, ampullary gland; CG, coagulating gland. (Adapted with permission from ref. 60).

Fig. 5. Scheme of implantation and removal of orthotopic prostate cancer (CaP) tumors. Tumor bits of CaP xenografts were implanted into the coagulating gland (A). After tumors reached a size of 250 to 500 mm3, the animals underwent a second surgery. Tumor and associated seminal vesicle were ligated and removed (B). SV, seminal vesicle; VD, vas deferens; U, ureter; B, bladder; AG, ampullary gland; CG, coagulating gland. (Adapted with permission from ref. 60).

To represent a good model of the human disease, animal models of CaP bone metastasis should exhibit two major characteristics: metastasis to bone, and osteoblastic response in the bone. However, unlike the human disease, human CaP xenograft models rarely metastasize spontaneously to bone from the orthotopic site of primary tumor growth, and, in the few instances in which this has been noted, they did not yield osteoblastic lesions. Although much effort has been spent on the establishment of new CaP xenografts that spontaneously metastasize to bone yielding an osteoblastic reaction, none exists that is reasonably reproducible and at a sufficient frequency to be experimentally useful. Therefore, to study the biology of CaP bone metastases and test new treatment modalities, experimentally induced metastatic models have been used as alternatives. The most commonly used alternatives are intravenous or intracardiac injection of tumor cells, and direct injection of cells into bone.

5.2.1. Intravenous and Cardiac Injection

Shevrin et al. (82) used an injection of the CaP cell line PC-3 into the tail veins of athymic mice while the inferior vena cava was occluded. This technique diverted cells into the vertebral venous plexus. Bone lesions developed in 3 of 16 experimental mice. Two tumor sublines were established from explant cultures of bone lesions, and re-injection of these cells into the tail vein resulted in bone metastasis in 19 of 36 animals. The main sites of bone metastasis were lumbar vertebrae, pelvis, and femurs. Wang and Stearns (92) selected highly invasive PC-3 sublines based on enhanced capacities to migrate across Matrigel in vitro, and, when these cells were injected intravenously in the tail vein of SCID mice, they metastasized to a wide variety of tissues. Four distinct sublines were isolated that metastasized preferentially to various skeletal sites in approx 80% of animals. Angelucci et al. (93) used PC-3 cells that were injected into the left cardiac ventricle of athymic mice, and followed meta-static spread and growth by luminescence. Sixty-four percent of animals developed osteolytic bone metastases. In this study, Angelucci et al. (93) isolated PC-3 cells from bone, named PCb2, that exhibited a more invasive phenotype, resulting in higher numbers of bone metastases.

The LNCaP cell line has also been used for studies of interactions between CaP and bone cells (94), although these cells do not metastasize spontaneously to the bone or colonize bone. Thalmann et al. (30) developed sublines of LNCaP, of which, one, C4-2, was able to metastasize to bone, although at low frequency (2 of 20). C4-2 sublines, designated B2, B3, B4, and B5, have a higher propensity to metastasize to bone and cause osteoblastic lesions, but, again, the rate of bone metastases was rather low (31,32,95).

In summary, intravenous and cardiac injections can be used to generate experimental bone metastases of CaP. However, these models have critical limitations:

1. PC-3 and its sublines result in osteolytic bone lesions, and do not produce PSA, a marker of prostate epithelial cells, whereas human CaP bone metastases exhibit mainly osteoblastic characteristics.

2. The C4-2b human CaP models (30-32) result in bone metastases after injection of the cancer cells, but only after a long delay, and the frequency of these metastases is still too low for preclinical studies.

5.2.2. Direct Bone Injections

Direct injection of tumor cells into the bone marrow has been developed as an alternative to spontaneous metastases for the study of interactions between CaP and bone cells (24,96-102).

5.2.3. SCID-hu Model

The possibility exists that the low frequency of bone metastasis in animal models is caused by incompatibilities between tumor and host environment. Nemeth et al. (24) used human fetal bone grafted subcutaneously into SCID mice as a substrate for the growth of human CaP cells PC-3, DU 145, and LNCaP. After 4 weeks of bone growth, tumor cells were injected intravenously or directly into the implanted bone tissues. Only intravenously injected PC-3 cells readily colonized human bone in 5 of 19 animals (Fig. 6) whereas no metastases were observed in implanted human lung or intestinal tissues or mouse bone, demonstrating tissue specificity of the process. Direct injection of tumor cells into the bone implants resulted in tumors of DU 145, PC-3, and LNCaP in 75 to 100% of animals. PC-3 and DU 145 lesions were primarily osteolytic, whereas LNCaP lesions were both osteoblastic and osteolytic (Fig. 7). Davies et al. (46) used LAPC-4 CaP cells in the SCID-hu model to establish cells that were more aggressive with stronger metastatic character. LAPC-4 cells were injected near to bone, and LAPC-4 (squared) was selected after growth in the bone environment. These cells form tumors, develop androgen independence, and metastasize to the human bone implants after orthotopic implantation.

5.2.4. Human Adult-Bone Model

The SCID-hu model adapted by Cher's group used human fetal bone as a substrate for human CaP bone metastases. Because CaP affects mainly older men, and bone remodeling is different in fetal and

Was this article helpful?

0 0

Post a comment