Targeting of Bone-Derived Insulin-Like Growth Factor-II by a Human Neutralizing Antibody Suppresses the Growth of Prostate Cancer Cells in a Human Bone Environment

Bone has long been recognized as the most common target organ for prostate cancer metastasis (1, 2). Bone metastasis causes skeletal complications and results in poor prostate cancer outcome in patients (3). Unfortunately, the efficacy of conventional therapies, including androgen deprivation, for the bone metastasis is limited. Despite extensive efforts to develop new therapeutic approach, more efficient treatment of bone metastases from prostate cancer has not been established.

Although the precise reason why prostate cancer cells prefer bone tissue is not fully understood, one hypothesis is that the bone is a source of abundant growth factors required for the growth and survival of prostate cancer cells (4). Previously, we established a bone-specific metastasis model of human prostate cancer to investigate the mechanism of bone metastasis from prostate cancer (5). Using this model, in which human adult bones (HAB) were engrafted in mice, we showed previously that the release of growth factors, such as insulin-like growth factors (IGF), arising from bone resorption induced by the tumor cells was important for bone metastasis from prostate cancer (6, 7).

IGFs are the most abundant growth factors stored in bone matrix (8). The actions of IGFs are inhibited when they bind to IGF-binding proteins. Prostate cancer cells are known to secrete certain IGF-binding protein proteases as represented by prostate-specific antigen (PSA; refs. 9, 10). Thus, the presence of IGF-binding protein proteases could conceivably release IGFs from the complexes and increase the local concentration of free IGFs, which in turn would induce additional growth of prostate cancer cells in the bone microenvironment. In humans, two distinct IGFs are known to exist, IGF-I and IGF-II. IGF-I and IGF-II share a 62% sequence homology and have overlapping but distinct functions: both IGF-I and IGF-II activate the IGF-I receptor (IGF-IR), whereas IGF-II also can activate the insulin receptor. Moreover, IGF-II is the most abundant growth factor in human bone, and the IGF-II level is nine times higher than IGF-I level. In contrast to human bone, IGF-I is three times higher than IGF-II in mouse bone (11). Interestingly, in our bone-transplanted model, human prostate cancer cells metastasized to implanted human bone but not to mouse bone (5). These findings suggest that IGF-II is more supportive of human prostate cancer than IGF-I in the human bone environment, which also supports the concept that IGF-II might be a potentially useful target for the treatment of bone metastases from prostate cancer.

We recently developed a fully human monoclonal antibody (mAb; m610), which is highly specific for IGF-II but not for either IGF-I or insulin (12). To explore the mechanistic role of IGF-II in the growth of prostate cancer cells in human bone and the therapeutic potential of inhibiting IGF-II by a neutralizing antibody, we used our HAB implanted model and a fully human mAb, m610, specific for IGF-II. Our results indicate that IGF-II plays an important role in the prostate cancer cell growth in human bone and that the human mAb m610 and other anti–IGF-II mAbs could be promising candidate therapeutics for treatment of prostate cancer.

An androgen-responsive human prostate cancer cell line, MDA PCa 2b, was purchased from the American Tissue Culture Collection. The cells were maintained in BRFF-HPC1 medium (Athena Enzyme System) containing 20% fetal bovine serum and were incubated at 37°C in a humidified atmosphere of 5% CO₂.

Male nonobese diabetic/severe combined immunodeficient mice were obtained from CLEA Japan. The mice were maintained under specific pathogen-free and temperature-controlled air conditions according to institutional guidelines. The mice used in the experiments were all 6 to 8 weeks old.

After obtaining the informed consent of the patients, HAB was obtained from lung or esophagus cancer patients (age range, 44-82 years; mean, 66.5 years) who had undergone a pulmonary lobectomy or esophagectomy in the Division of Thoracic Oncology, National Cancer Center Hospital East. The implantation of HAB into nonobese diabetic/severe combined immunodeficient mice was done as described previously, with some modifications (5). Briefly, morselized cancellous bone (200 mm³) from HAB was implanted into mammary fat pad (MFP) of the mice through a small skin incision in the right flank. In a preliminary examination, graft survival was ascertained by histologic examination and the detection of human IgG in the mouse sera using Western blotting at 4 weeks after the implantation. At 4 weeks after bone implantation, single-cell suspensions (8 × 10⁶ cells/100 μL serum-free medium) of MDA PCa 2b cells were directly injected into the marrow spaces of the implanted HAB.

The single-cell suspensions prepared as mentioned above were injected into the MFP of the age-matched nonobese diabetic/severe combined immunodeficient mice (10-12 weeks old) through a small skin incision in the right flank.

An anti–IGF-II fully human mAb (IgG1 m610), which binds to IGF-II but not to IGF-I and insulin, was identified by screening of a large antibody library in the authors' laboratory (National Cancer Institute-Frederick, NIH; ref. 12). To obtain sufficient quantities of m610, m610 was produced from a permanent cell line developed by transfection of CHO-K1 cells and selection of the highest producing clone (13). A control human antibody (IgG1 m102.4), which was produced from a permanent CHO cell line under the same conditions as m610 (14), was chosen to be as closed in sequence as possible to m610 but with entirely different specificity. Compared with m610, m102.4 has identical Fc, hinge and CH1, very similar CL, and similar frameworks of the variable domains but different complementarity-determining regions to ensure specificity to Hendra and Nipah virus.

In the model involving the induction of bone tumor in HAB (HAB model), m610 (1 or 10 mg/kg) or a control antibody (10 mg/kg; n = 8 per each group) was intraperitoneally administered to the mice at weekly intervals for 4 weeks starting immediately after the inoculation of the MDA PCa 2b cells into the HAB. In the model involving the induction of tumors in MFP (MFP model), m610 (10 mg/kg) or control antibody (10 mg/kg; n = 5 or 4 per group, respectively) was administered to the mice at weekly intervals for 4 weeks after the inoculation of the cells into the MFP. At 4 weeks after the start of the m610 treatment, the mice were sacrificed and the HAB tissues and mouse organs (lung, liver, spleen, kidney, and lumbar vertebrae) in the HAB model and the MFP tissues in the MFP model were harvested. The body weight of mouse was measured at the time of HAB implantation, the start of the m610 treatment, and the time of sacrifice.

Harvested specimens were fixed in 10% neutral buffered formalin and embedded in paraffin. The bone specimens were decalcified with EDTA solution (Wako) before paraffin embedding. Two-μm-thick sections were prepared from each specimen. Then, we analyzed the tumor burden in the HAB and the MFP by histomorphometrically examining sections stained for PSA, as described previously (6). Briefly, tissue sections of the HAB and MFP specimens were cut at three levels far enough apart (>200 μm) to avoid replicating the sampling of a single surface event; immunostaining for PSA was done using the EnVision Plus System HRP kit (Dako) according to the manufacturer's protocol. An anti-PSA polyclonal antibody (1:800; Dako) was used as a primary antibody. Antigen retrieval was not done. The total tissue area and the total tumor area immunostained by the anti-PSA antibody were determined using image analysis software (Image J version 1.38×; NIH). The results were expressed as the absolute tumor area (mm²) and the tumor area as a percentage of the tissue section area (%).

The immunostaining procedure used for IGF-IR, insulin receptor, Ki-67, and cleaved caspase-3 was the same as described above. However, microwaving (95°C, 20 min) in citrate buffer (pH 6.0) or in Target Retrieval Solution High pH Buffer (Dako) was required for Ki-67 antigen retrieval or cleaved caspase-3 antigen retrieval, respectively. An anti–IGF-IR mAb (1:100; Chemicon), an anti–insulin receptor mAb (1:100; Chemicon), an anti-Ki-67 mAb (1:50; Dako), or an anti–cleaved caspase-3 polyclonal antibody (1:200; Cell Signaling Technology) was used as a primary antibody.

To evaluate the proliferative and apoptotic statuses of the tumor cells, we counted the number of Ki-67–positive and cleaved caspase-3–positive cancer cells and the total cancer cells in three high-power fields of the most strongly stained tumor areas and then calculated the percentages of these immunostaining positive cancer cells to the total cells. The necrotic area was not included in the evaluation of apoptosis. The data from samples with the presence of cancer cells were presented.

Blood samples were obtained from the mice before sacrifice. The serum PSA levels were determined using a chemiluminescent enzyme immunoassay kit (Hybritech-PSA; Beckman Coulter) according to the manufacturer's protocol.

MDA PCa 2b cells (5 × 10⁵) seeded on a 3.5-cm dish were cultured in complete growth medium for 48 h to allow cell adhesion. After changing the medium to serum-free F12K medium (Invitrogen), the cells were treated for 48 h with serum-free medium alone, with recombinant human IGF-II (1, 10, or 100 ng/mL; R&D Systems), with the control antibody (10 μg/mL) plus IGF-II (100 ng/mL), with m610 (0.1, 1, or 10 μg/mL) plus IGF-II (100 ng/mL), or with m610 (10 μg/mL) in the absence of IGF-II. The cells were dislodged by trypsinization and the number of viable cells was counted using the trypan blue (Invitrogen) exclusion assay.

MDA PCa 2b cells (4 × 10⁶) seeded on a 6-cm dish were cultured in complete growth medium for 48 h. The cells were rinsed and cultured in serum-free F12K for 6 h. Then, the cells were treated with the serum-free medium alone, with IGF-I (100 ng/mL; R&D systems), or with IGF-II (100 ng/mL) for 15 min. Immunoprecipitation and Western blot analysis procedures were done as described previously (15). Two hundred micrograms of protein from each cell lysate were, respectively, immunoprecipitated with an anti–IGF-IR polyclonal antibody (2 μg; clone C-20; Santa Cruz Biotechnology) or with an anti–insulin receptor polyclonal antibody (2 μg; clone C-19; Santa Cruz Biotechnology) for 12 h at 4°C, and each immunoprecipitate was collected using protein G Sepharose. Each sample (derived from 20 μg total cell lysate) was fractionated by 7.5% SDS-PAGE and transferred to polyvinylidene fluoride membrane (Millipore). The membranes were probed with an anti-phosphotyrosine mAb (1:2,000; clone 4G10; Upstate Biotechnology) and visualized by an enhanced chemiluminescence (Amersham Biosciences). After stripping the membrane, total IGF-IR or insulin receptor in the immunoprecipitates was detected with C-20 or C-19 antibody, respectively.

In an assay for the effect of IGF-II on phosphorylation of Akt and mitogen-activated protein kinase (MAPK), the cells were treated with the serum-free medium or with IGF-II (1, 10, or 100 ng/mL) for 15 min. In an assay for the inhibitory effect of m610 on the IGF-II–induced phosphorylation of IGF-IR, insulin receptor, and Akt, the cells were preincubated for 30 min with m610 (0.1, 1, or 10 μg/mL) before IGF-II treatment (100 ng/mL). Twenty-microgram proteins from the lysates were subjected to a Western blot analysis. To detect phosphorylated IGF-IR and insulin receptor, a polyclonal antibody specific for both phosphorylated Tyr^1135/1136 of IGF-IRβ and phosphorylated Tyr^1162/1163 of insulin receptor β was used (1:1,000; Biosource International). Phospho-Akt, total-Akt, phospho-MAPK, or total MAPK was, respectively, detected with an anti–phospho-Akt (Ser⁴⁷³) polyclonal antibody, an anti-Akt polyclonal antibody, an anti–phospho-MAPK (Tyr^202/204) mAb, or an anti-MAPK polyclonal antibody (1:1,000 each; Cell Signaling Technology).

Data are expressed as mean ± SE. Variables were compared between experimental and control groups using Student's t test. The relation between the percentage of the tumor area and the serum PSA level was assessed using a linear regression analysis. Statistical calculations were done on a Windows personal computer with the GraphPad PRISM software version 4.03 (GraphPad Software). P < 0.05 was considered statistically significant.

To investigate the effect of m610 on the growth of bone tumors from prostate cancer cells in human bone tissue, m610 (low dose, 1 mg/kg; high dose, 10 mg/kg) or control antibody (10 mg/kg) was administered to the mice with HAB implantation according to the protocol (Fig. 1). At 4 weeks after the treatments, a histomorphologic analysis of the HAB specimens was done. All 8 of the control antibody-treated mice, 7 of the 8 mice in the low-dose group, and 6 of the 8 mice in the high-dose group developed bone tumors. The mean total tumor areas in the low-dose and high-dose groups were 0.8 ± 1.5 and 0.8 ± 1.2 mm², respectively; these values were significantly smaller than that in the control group (2.2 ± 0.9 mm²; P = 0.0397 and 0.0152, respectively; Fig. 2A and B). The mean percentages of the tumor area to the total tissue section area in the low-dose and high-dose groups were 0.7 ± 1.4% and 0.6 ± 0.9%; these values were also significantly lower than that in the control (2 ± 0.8%; P = 0.0438 and 0.0075, respectively; Fig. 2C).

To further assess the effect of m610 on tumor growth, serum PSA level, which is widely used for the diagnosis and therapeutic evaluation of prostate cancer, was measured. The serum PSA levels of the m610-treated groups were significantly lower than that of the control group, in parallel with the reduction in the tumor area (low-dose group, P = 0.0384; high-dose group, P = 0.0072; Fig. 2D). Additionally, a strong correlation was observed between the percentage of the tumor area and the serum PSA values (r² = 0.7656; P < 0.0001, linear regression analysis).

Together, these results of the histomorphologic examination and the PSA measurement show that m610 significantly suppressed the growth of bone tumors induced by MDA PCa 2b cells in implanted HAB, although a dose escalation effect of m610 on the response was not observed.

To understand the processes by which the bone tumor growth was suppressed by m610, immunohistologic examinations were done. First, we examined the expression levels of IGF-IR and insulin receptor, which are functional receptors of IGF-II, on the bone tumors by MDA PCa 2b cells in HAB. IGF-IR and insulin receptor were diffusely expressed in both control and m610-treated bone tumors (Fig. 3A and B), and no differences in the expression levels of these receptors were microscopically observed between the bone tumors. These findings indicate that the suppressive effect of m610 on the bone tumor growth was not caused by alterations in the expression levels of these receptors. Next, immunohistologic staining for Ki-67 or cleaved caspase-3 was done to investigate the proliferative and apoptotic statuses of the bone tumors. Compared with the control, an apparent decrease in the number of cancer cells with Ki-67–positive nuclei was observed in the m610-treated bone tumor sections (Fig. 4A), whereas a slightly increased number of cleaved caspase-3–positively stained cancer cells was observed (Fig. 4B). Then, the percentages of positively immunostained cancer cells out of the total cell numbers were calculated in three high-power fields with the most highly stained tumor areas. As shown in Fig. 4C, the percentages of Ki-67–positive cancer cells in the m610-treated groups were significantly lower than that in the control (low-dose group, P = 0.0007; high-dose group, P = 0.0002). Meanwhile, the mean percentages of cleaved caspase-3–positive cancer cells in the m610-treated groups were slightly higher than that in the control but not significantly (Fig. 4D; low-dose group, P = 0.1937; high-dose group, P = 0.0601). These findings show that m610 suppressed the proliferation of MDA PCa 2b cells in HAB but did not lead to a significant increase of apoptosis, although a trend to an increase was observed for the high-dose group.

To examine whether m610 exerts a direct growth-inhibitory effect on MDA PCa 2b cells, m610 (10 mg/kg) or control antibody (10 mg/kg) was administered to the mice without HAB implantation according to the protocol (Fig. 1). Three of the four control antibody-treated mice and four of the five m610-treated mice developed tumors in the MFP tissue. The mean total tumor areas of the control and m610-treated tumors in the MFP model were 1.1 ± 1.5 and 1.3 ± 1.4 mm², respectively (P = 0.8787; Fig. 5A). The mean percentages of the tumor area to the total tissue section area in the control and m610-treated group were 1.3 ± 1.8% and 2 ± 2.1%, respectively (P = 0.6332). The serum PSA levels of the control and m610-treated group were 0.6 ± 0.9 and 2.7 ± 1.2 ng/mL, respectively (P = 0.2912). An immunohistologic examination revealed that IGF-IR and insulin receptor were focally expressed on both control and m610-treated tumors in the MFP tissue and that the expression levels of the receptors were not different between the tumors (Supplementary Fig. S1A and B). Ki-67 immunostaining revealed that the proliferative statuses were similar between control and m610-treated tumors, consistent with the results for tumor size (Fig. 5B; Supplementary Fig. S1C; P = 0.7462). Together, these findings indicate that m610 had no effect on the growth of MDA PCa 2b cells in MFP, unlike in HAB, showing that the inhibitory effect of m610 on tumor cell proliferation is restricted to within the bone.

Using a cell proliferation assay, we confirmed that m610 inhibits the IGF-II–induced proliferation of MDA PCa 2b cells. Exogenous IGF-II dose-dependently induced the proliferation of MDA PCa 2b cells (Fig. 6A, lanes 1-4). The control antibody had no inhibitory effect on the IGF-II–induced cell proliferation (Fig. 6A, lane 5). m610 at doses of 1 and 10 μg/mL significantly inhibited the IGF-II–induced cell proliferation, compared with the control antibody group (Fig. 6A, lanes 7 and 8; P = 0.0049 and 0.0039, respectively), but m610 had no effect on the proliferation of MDA PCa 2b cells without exogenous IGF-II stimulation, indicating that m610 did not have a direct effect on the cells (Fig. 6A, lane 9).

To confirm that m610 prevents signal transduction mediated by receptor interaction with IGF-II in MDA PCa 2b cells, we firstly investigated whether IGF-II activates IGF-IR and insulin receptor in the cells by immunoprecipitation-Western blot analysis. On ligand binding, IGF-IR and insulin receptor undergoes autophosphorylation of tyrosine residues in their β-subunits. Western blot analysis showed that IGF-II induces phosphorylation of tyrosines in each receptor in MDA PCa 2b cells (Fig. 6B). Note that IGF-II induces significantly higher level of insulin receptor phosphorylation than of the IGF-IR phosphorylation. In contrast, IGF-I induces much higher levels of phosphorylation of the IGF-IR than of the insulin receptor. Tyrosine phosphorylation of IGF-IR or insulin receptor leads to phosphorylation of the downstream kinases Akt and MAPK. We next examined whether IGF-II induces the downstream kinases in MDA PCa 2b cells by a Western blot analysis. IGF-II induced Akt phosphorylation in the cells but not MAPK phosphorylation (Supplementary Fig. S2). We finally investigated whether m610 prevents the IGF-II–induced IGF-IR and insulin receptor signaling in MDA PCa 2b cells by a Western blot analysis. To detect the total amount of activated statuses (IGF-IR plus insulin receptor), which is most relevant for the degree of induced cell proliferation, an antibody specific for phosphorylated tyrosines of IGF-IR and insulin receptor was used in the analysis. As shown in Fig. 6C, the m610 dose-dependently prevented the IGF-II–induced tyrosine phosphorylation of IGF-IR/insulin receptor as well as serine phosphorylation of Akt in MDA PCa 2b cells.

IGF-II is the most abundant growth factor in human bone (8) and, unlike in mouse bone, is nine times more abundant than IGF-I (11). Our previous study revealed the role of IGFs in prostate cancer growth in human bone, as shown by the suppressive effect of a neutralizing antibody (KM1468), which cross-reacts with human IGF-I and IGF-II and mouse IGF-II, on the growth of MDA PCa 2b cells in implanted HAB (6). Moreover, another study of ours indicated that tail vein–injected prostate cancer cells preferentially metastasized to implanted HAB rather than to implanted mouse bone or native bone (5). These findings suggested that IGF-II may support human prostate cancer more efficiently than IGF-I in human bone environment, in accordance with the “seed and soil” theory by Paget (16) for metastasis. Therefore, we became interested in specifically targeting IGF-II as a potential therapy for bone metastasis from prostate cancer.

Although clinical studies have suggested that IGF-II contributes to the development and progression of several cancers including prostate cancer (17–19), only a few studies by others and us have reported an in vivo antitumor effect of targeting IGF-II. O'Gorman et al. reported that the overexpression of the IGF-II receptor, which is a clearance receptor for IGF-II, on choriocarcinoma cells reduced the cell growth in vitro and in vivo, indicating that inhibiting IGF-II is a potential target for cancer therapy (20). In another strategy using neutralizing antibodies, Miyamoto et al. reported that an IGF-II neutralizing antibody partly suppressed the development of liver metastasis induced by the intrasplenic injection of colon cancer cells, whereas IGF-I neutralizing antibody significantly suppressed the development of liver metastasis (21). Because IGF-II production in rodents becomes attenuated in most tissues soon after birth unlike IGF-I production (22), the effect of IGF-II as an endocrine and/or paracrine on tumor growth may be underestimated in models using mouse-native organs.

The present study showed that inhibiting IGF-II by m610 effectively suppresses the growth of prostate cancer cells in a human bone environment. The following results support this conclusion: (a) m610 significantly suppressed the growth of bone tumors from MDA PCa2b cells in implanted HAB; it even decreased the number of mice developing tumors from 8 to 6, although larger number of mice are needed to confirm this effect; (b) Ki-67 immunostaining revealed that the proliferative status of m610-treated bone tumor was apparently suppressed; (c) in the absence of HAB, m610 had no effect on the growth of tumors from MDA PCa 2b cells, indicating that the suppressive effect of m610 on the tumor cells was restricted to those within HAB; and (d) in vitro assays confirmed that m610 prevents the exogenous IGF-II–induced proliferation of MDA PCa 2b cells. These results provide clear evidence of the important role of IGF-II for tumor growth in the HAB model and of an in vivo antitumor effect of m610 on metastatic bone tumor from prostate cancer through a mechanism involving the inhibition of IGF-II. They also underscore the notion that IGF-II levels in local tissue may be more relevant in tumor promotion than its plasma levels and that a paracrine mechanism of IGF-II may play a critical role in tumor growth.

The potency of m610 on the growth inhibition of MDA PCa 2b cells in the HAB model is 65%, whereas that of the previously published antibody, KM1468, is 97% compared with the respective controls: the antitumor effect of inhibiting IGF-II alone is lower than that of inhibiting both IGF-I and IGF-II in the HAB model. Despite the lower antitumor effect of m610 in the HAB model, targeting IGF-II by m610 might provide certain clinical benefits in cancer therapy for the following reasons. (a) Growth hormone feedback is not known for IGF-II, but IGF-I is regulated by this feedback. Lowering IGF-I concentration triggers feedback upregulation of the growth hormone; the growth hormone compensates for the reduced IGF-I levels. Thus, targeting IGF-I might require high concentrations of anti–IGF-I antibodies. It should be noted that KM1468 is not reactive with mouse IGF-I; therefore, its use in our HAB model does not trigger the growth hormone feedback on the IGF-I and the tumor growth. (b) Because m610 is a fully human antibody, its clinical use is less likely to induce immune reactions compared with murine antibodies.

Targeting IGF-II might provide additional therapeutic benefit in combination with other treatments. IGF-IR activation by IGF-I and IGF-II has been shown to stimulate the growth of a wide range of cancer cells (23, 24). Currently, potent mAbs against the IGF-IR are being tested in clinical trials against multiple tumor types including prostate, breast, and colon cancers and Ewing's sarcoma (25). Importantly, it is becoming increasingly evident that insulin receptor activation by IGF-II enhances the growth of Ewing sarcoma and breast cancer in addition to the IGF-IR activation: cotargeting IGF-IR and insulin receptor is likely to be more effective than targeting the IGF-IR alone (26–28). Recently reported immunohistologic examinations of primary human prostate cancer show that IGF-IR as well as insulin receptor are both commonly expressed on the tissues (29). Our results from the immunohistologic examinations (Fig. 3) and Western blot analyses (Fig. 6B and C) suggest that insulin receptor activation by IGF-II plays an important role in the prostate cancer cell growth in bone in addition to the IGF-IR activation. Based on these findings, IGF-II could be a promising candidate target in therapeutic strategies for cotargeting IGF-IR and insulin receptor.

If m610 were capable of suppressing the growth of prostate cancer in bone without any adverse reactions, m610 therapy might considerably improve the quality of life of patients with bone metastases of prostate cancer. In the present study, the administration of m610 did not affect the body weights of mice during the 4-week treatment period (Supplementary Fig. S3), and no adverse findings were observed in the histologic examination for the mouse organs.

In conclusion, the present study showed that an IGF-II–specific antibody, m610, can sufficiently suppress the growth of bone tumors from MDA PCa 2b cells in a human bone environment and that this effect is caused by the suppression of the proliferative status of the tumor cells. These results suggest that the targeting of bone-derived IGF-II using a neutralizing antibody offers a new therapeutic strategy for bone metastasis from prostate cancer.

No potential conflicts of interest were disclosed.

We thank Sachiko Fukuda and Hiroko Hashimoto for technical assistance; Motoko Suzaki for secretarial support; and Maki Touko, Hiroyuki Maeda, Takako M. Oki, and Hiroyuki Yonou for suggestions.

Grant Support: Grant-in-Aid for the Third-Term Comprehensive 10-Year Strategy for Cancer Control from the Ministry of Health and Welfare of Japan and Research Resident Fellowships from the Foundation for Promotion of Cancer Research in Japan (T. Kimura, S. Ashimine, M. Yamazaki, and C. Yamauchi). The Intramural Research Program of the NIH, National Cancer Institute, Center for Cancer Research (Y. Feng and D.S. Dimitrov).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.