Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
In vivo models used in studies of bone metastases
Introduction
In human disease, bone metastasis is a slow process often developing years after the initial tumor is diagnosed and treated. Bone metastases are commonly only detected when patients present with late-stage symptomatic disease, at a point when extensive bone lesions have developed [ , ]. Bone is an inaccessible site that is not routinely biopsied or otherwise sampled, and studies using human material are further complicated by the fact that patients may have undergone extensive therapeutic interventions during the treatment of their primary tumor. As a result, our understanding of the precise biological mechanisms of bone metastasis is incomplete, hampering the development of successful treatment and prevention strategies. Researchers have therefore increasingly turned to the use of animal models, although these generally represent only particular stages of human disease rather than the gradual transition from benign disease through hyperplasia and subsequent aggressive and advanced metastatic disease.
The majority of in vivo studies of bone metastases use mice, as these are widely available, have a short life span, and a relatively large number of animals can be easily investigated. However, mice do not naturally develop bone metastases, suggesting that they may not provide the optimal setting to study this aspect of human cancer, and key molecules and mechanisms that drive human disease may potentially be missed in murine model systems. Despite their limitations, in vivo models of bone metastasis continue to provide valuable insights into the molecules and pathways that may play a role in the development of human disease, as well as the effects of a range of therapies [ , ].
Models used in the studies of breast cancer bone metastases ( Table 4.1 )
Patients with advanced breast cancer are at high risk of developing bone metastases, associated with considerable morbidity and mortality. As covered elsewhere in this volume, the bone lesions arise as a result of complex molecular interactions and involve a range of cell types, including tumor cells, osteoblasts, osteoclasts, immune cells, and cells of the vasculature [ , ]. The net outcome is excessive bone resorption leading to generation of lytic lesions. The following sections describe the main models used to increase our understanding of the processes underlying the development of breast cancer bone metastases [ , ].
Table 4.1
Current main animal models used in studies of breast cancer–induced bone disease.
| Model type | Cells administered | Route of administration | No of cells injected | Duration | Disease characteristics and features of models |
|---|---|---|---|---|---|
| Xenograft immunodeficient animals | MDA-MB-231 (Luc) MDA-MB-435 |
Intracardiac | 1 × 10 5 −1.5 × 10 6 1 × 10 5 |
3–5 weeks | Bone is the main site of colonization low frequency of lung and liver metastases. Osteolytic bone lesions |
| MDA-MB-231 MCF7 |
Intratibial | 1–2 × 10 5 | 2–4 weeks | Extensive osteolytic disease | |
| BO2 | Intravenous | 1 × 10 5 | 2–6 weeks | Cells home specifically to bone, lytic bone disease evident round day 18 | |
| Syngeneic immunocompetent animals | 4T1 Mouse mammary carcinoma cells |
Orthotopic mammary fat pad | 1 × 10 6 | 3–4 weeks | Multifocal disease with metastases in lung, liver, kidney, and bone |
| Human-to-human models Immunodeficient animals | Human breast cancer cell lines Primary human breast cancer cells |
Intravenous | Tumor cells injected 3–4 weeks following implantation of human adult bone Animals culled 8 weeks later |
Cells primarily home to human bone, generating osteolytic disease. No metastases detected in mouse skeleton |
Xenograft models
The majority of in vivo studies of bone metastases have used xenograft models, where human tumor cells are implanted in immunocompromised mice [ ] or rats [ ]. The human tumor cells are now routinely engineered to express green fluorescent protein (GFP), luciferase (Luc), or other markers, allowing noninvasive in vivo imaging of tumor development. Depending on the cell line used, the age and strain of the animal, as well as the route of implantation, tumors will develop in bone within 3–5 weeks, providing the opportunity to study both the biological aspects of bone metastases and effects of therapy [ ].
Intracardiac injection
By injecting human breast cancer cells into the left ventricle of the heart, the cells will bypass the pulmonary vasculature resulting in widespread metastasis formation in several skeletal sites (mainly long bones) within 4–5 weeks [ ]. This method requires a high degree of technical skill, but does provide a model of early bone metastasis, without the complication of potential rapid outgrowth of a primary tumor or development of visceral metastases. However, it does not represent the initial steps in the disease process, where tumor cells leave a primary site (breast), enter the circulation, and subsequently recolonize a distant site (bone). Intracardiac (i.c.) injection of breast cancer cells has been widely used to investigate the role of tumor-derived factors in accelerated bone turnover, the ability of bone-derived molecules to support tumor cell growth as well as the effects of therapeutic agents on bone metastases and the associated bone disease.
Mechanisms of bone metastasis formation
The i.c. injection model was used by Phadke and colleagues to perform a detailed and informative study of the kinetics of breast cancer cell trafficking in bone [ ]. GFP-expressing MDA-MB-435 cells were injected into the left ventricle of female athymic mice and animals were culled at time points ranging from 1 h and up to 6 weeks. This allowed charting the precise localization of breast cancer cells from the initial colonization of bone to advanced disease. The study showed that 1 h following injection, tumor cells can be detected in the highly vascularized metaphysis of the femur and that they are gradually cleared over the next 72 h. One week after injection, small tumor cell foci were detected mainly in the distal metaphysis, and a few of these developed into larger tumors associated with loss of the majority of the trabecular bone during the following 2–3 weeks. This study also showed that increasing numbers of tumor cells were associated with a substantial decrease in osteoblast number through induction of apoptosis. Surprisingly, there was also a decrease in the number of osteoclasts. The osteoblast:osteoclast ratio was reduced from around 40:1 at baseline to 4:1 by 4 weeks, as bone turnover balance shifted toward increased resorption. These data show that breast cancer cells cause a rapid depletion of bone-forming cells by an unknown mechanism, likely to prevent repair of tumor-induced lytic lesions. The implications for human disease are that an effort should be made to develop therapeutic strategies to protect the osteoblast from the detrimental effects of tumor cells, for use in combination with antiresorptive agents.
Improved technologies, including tumor cell labeling and imaging, have made it possible to detect very small tumor foci in vivo and single tumor cells in bone ex vivo . The development of lipophilic fluorescent dyes, which are retained in slow-cycling cells, has allowed studies of tumor cell dormancy. Studies utilized the lipophilic dye Vybrant® DiD (DiD+)-stained breast cancer cells to identify a subpopulation of slow-cycling and mitotically quiescent cells. DiD+ cells were shown to reside in the trabecular region of the bone 5 days after i.c injection [ ]. Mature mice (12 weeks old) were used in this study, as in these the bone microenvironment prevents outgrowth of tumors and supports disseminated tumor cell dormancy. The study of these subpopulations is critical for the development of novel therapies for the prevention of breast cancer recurrence.
Brown and colleagues used the i.c. model to determine the effects of direct tumor cell contact on osteoblasts and osteoclasts during bone metastasis progression [ ]. MDA-MB-231 (GFP+) breast cancer cells were injected i.c. in female BALB/c nude mice and groups of animals were culled between days 10 and 33. Detailed analyses of the tibiae and femora revealed significant tumor burden in bone on day 15, followed by detectable bone loss on day 19. This was the first study to report a significant increase in osteoblast numbers in areas of bone not in direct contact with tumor cells, reflecting a compensatory mechanism in response to tumor-induced increases in osteoclastic bone resorption. In addition, tumor cells induced a significant change in bone cell numbers prior to changes in bone volume. These results suggest that the current clinical practice of initiating therapeutic intervention based on the presence of scan-detectable bone lesions may be too late to affect the key stages of metastatic progression.
By genetic manipulation of the cancer cells prior to i.c. injection, the role of particular molecules in bone metastases formation can be elucidated. This approach has been utilized to study how overexpression of the proto-oncogene c-Src in MDA-MB-231 cells affected the ability of the cells to metastasize [ ] and more recently to demonstrate a role for noggin in breast cancer bone metastasis [ ].
Effects of therapies
I.c. injection of breast cancer cells has also been used to study effects of therapeutics on cancer-induced bone disease (CIBD). Changes in the normal ratio of receptor activator of NF-kappaB ligand (RANKL) and its soluble decoy receptor osteoprotegerin (OPG) may lead to increased osteolysis in human bone metastases. Administration of recombinant osteoprotegerin (OPG-Fc) following i.c. injection of MDA-MB-231-Luc cells inhibited skeletal growth of tumor cells in vivo when given both as a preventative (day 0) and as a therapeutic agent for established bone metastases (day 7) [ ]. OPG-Fc inhibited tumor-induced osteoclastogenesis and osteolysis, accompanied by reduced tumor burden and increased levels of tumor cell apoptosis. This study supports that RANKL is an important regulator of bone metastasis and therefore an attractive therapeutic target, and was the first to show that survival of mice with established bone metastases could be improved by treatment with OPG-Fc, demonstrating the importance of tumor cell–stromal interactions for the development and progression of CIBD.
The i.c model has also been useful to study the effects of the antiresorptive drug zoledronic acid (ZOL) on bone and breast cancer metastasis in the pre- and postmenopausal settings [ ]. In this study, ovariectomy (OVX) of mature (12-week-old) mice caused a decrease in bone volume and turnover, mimicking the postmenopausal setting. Intracardiac injection of MDA-MB-231 cells resulted in a significantly higher occurrence of skeletal tumors in mice that had undergone OVX compared to sham-operated mice (mimicking the premenopausal setting). Administration of ZOL reduced the number of tumors in OVX mice compared to saline-treated (control) OVX mice, providing evidence to support that osteoclast activity was required to trigger growth of disseminated tumor cells in bone to form overt bone metastases. In mice where no tumor was detected by in vivo imaging, ex vivo multiphoton microscopy was used to confirm the presence of DID+ labeled tumor cells in bone, demonstrating that tumor cells had disseminated to bone but remained dormant. This study was the first to describe a model system where tumor cells in bone remained dormant for prolonged periods of time and demonstrated that changes to the bone microenvironment (OVX-induced bone resorption) could trigger the growth of disseminated tumor cells in bone.
Intravenous/intra-arterial injection
Specific “bone-seeking” breast cancer cell lines have been generated by isolation of cells from bone metastases (originally introduced by i.c. injection), expanding the cells in culture and reinoculating them into the left ventricle. Following a number of passages through bone (typically a minimum of 7), clones that home specifically to bone following i.v. or i.a. injection have been generated.
Mechanisms of bone metastasis formation
The group of Clezardin used this approach to develop the B02 model, where intracaudal injection of a bone-homing strain of MDA-MB-231 breast cancer cells results in hind limb tumors [ ]. They subsequently used the model to investigate whether therapeutic targeting of tumor alphavβ3 integrin prevents bone metastasis formation [ ]. Both tumor burden and extent of bone disease were found to be increased in animals injected with alphavβ3-overexpressing cells. A subsequent study assessed the inhibition of autotaxin (ATX), a member of the nucleotide pyrophosphate-pyrophosphatase family which is expressed in various cancers and confers a selective advantage in the progression of metastasis and osteolytic bone lesions [ ]. Furthermore, ATX binds to alphavβ3 integrin and additionally catalyzes the production of LPA which plays a role in cancer progression through the promotion of cell proliferation, motility, invasion, and survival. Following B02 i.a. injection, systemic administration of the ATX inhibitor, BMP22, resulted in significantly reduced osteolytic bone lesions. Interestingly, when comparing wild-type ATX to a mutant ATX (ATX-T209A), the number of B02 bone marrow (BM) colonies was significantly lower in ATX-T209A mice. Utilizing the i.v model has provided further understanding in the mechanism and potential role of LPA in breast cancer bone metastasis [ ].
A recent study assessed the intracaudal artery (CA) in the tail as a potential route of breast tumor cell injection [ ]. Using fluorescently labeled nanoparticles (near-infrared fluorescence), the authors compared the spread of the particles when injected via the i.v or CA routes. They found that CA injections resulted in a rapid (within 5 s) illumination of capillaries of the lower half of the mouse, whereas a slow and modest illumination was observed when administered by i.v. injection. The group proceeded to inject luciferase-tagged-lung carcinoma (LLC) cells through i.c (2 × 10 5 cells) or CA (1 × 10 6 cells) routes. Bioluminescence intensities were significantly higher in the hind limbs of mice injected by CA than i.c after 30 mins. However, the rate of disease progression was the same between the two models, suggesting that there were no differences in stress exerted on the cells through the two routes of injection. These results were reproduced using breast (MCF7 and MDA-MB-231), prostate (PC3), and other cancer cell lines. This study has not only identified a bone metastasis model that allows for a larger number of cells to be injected without resulting in high rates of mortality, but also demonstrates the bone-homing capabilities of cells that do not otherwise home to bone frequently (such as MCF7). Intrailiac arteries have also been utilized as a route of injection (IIA). This model, like CA, allows injection of a larger number of tumor cells (up to 5 × 10 5 cells depending on the aggressiveness of the cell lines). It also introduces cells into the circulation without causing damage to the local tissue, thus preventing inflammatory and wound-healing responses that may otherwise confound the bone-homing process [ ]. Two studies have consistently shown that the IIA model results in a high rate of hind limb metastases development while avoiding soft-tissue lesions [ , ]. Although the IIA model may prove to be an interesting model for bone metastasis, it is a highly invasive procedure requiring a high level of technical expertise.
Effects of therapies
The antitumor effect of clinical dosing regimens comparing the two bisphosphonates ZOL (ZOL) and clodronate (CLOD) has also been investigated using the B02 model [ ]. Treatment administered prior to injection of tumor cells (preventive protocols) was compared to starting once osteolytic lesions had developed (treatment protocols). Mice receiving a daily preventive regimen of CLOD, or with a daily/weekly preventive regimen of ZOL, showed a decreased tumor burden compared to mice treated with vehicle, whereas a single preventive dose of ZOL had no effect. The authors concluded that daily or repeated intermittent therapy with clinical doses of bisphosphonates inhibits skeletal tumor growth in this model.
In addition to the B02 model, therapeutic targets have also been investigated using the bone-homing MDA-MB-231 models. One such study explored the effects of an IL-1B receptor antagonist, anakinra, on bone metastasis [ ]. Having previously identified IL-1B as a potential biomarker for the increased risk of bone metastasis development [ ], it is overexpressed in a number of cancers, including breast cancer. In this study, two protocols were used to investigate the effects of anakinra on development and progression of bone metastasis [ ]. The first, termed preventative, used daily administration of anakinra for 31 days with intravenous injection of MDA-MB-231(-IV) cells on day 3. In the second protocol, termed treatment, i.v injection of MDA-MB-231(-IV) cells was followed by administration of anakinra commencing 7 days later and continuing daily for a duration of 21 days. No differences was observed in numbers of tumor cells that seeded the skeleton; however, the number of mice that develop overt metastases was significantly lower following both of the anakinra protocols, compared to a placebo group. Tumor volumes were also significantly smaller in animals treated with anakinra compared to the placebo. These data suggest that blocking IL-1B signaling does not impede metastasis to bone but inhibits outgrowth of tumor cells disseminated into this site, retaining these cells in a state of dormancy.
Clezardin's group has also reported that bone metastasis can be targeted in the CA model. In this study, the role of integrin a5, which is highly expressed in patients with poor bone metastasis-free survival, was investigated [ ]. B02 cells with a silenced integrin a5 gene injected into BALB/c mice by the CA route resulted in significantly fewer skeletal micrometastases compared to mice injected with control B02 cells (scrambled). Moreover, pharmacological inhibition of integrin a5 with a monoclonal antibody (M200) resulted in a delayed onset of bone disease with a significantly lower number of osteolytic lesions compared to a vehicle-treated cohort. This study demonstrates that the CA model may be used to identify targetable pathways for the development of novel therapies.
Intraosseous implantation
Bone metastases can also be modeled by implanting cancer cells directly into bone, most commonly by direct intratibial (i.t.) injection [ ]. This is generally used to investigate late-stage disease, as a large number of tumor cells are introduced directly into bone, bypassing the early steps of homing and colonization. Holes are drilled in the tibiae of anesthetized animals, the BM flushed out and replaced by a suspension of tumor cells. This will inevitably cause some damage to the bone around the injection site, complicating analysis of tumor-induced changes of bone structure. The majority of the injected tumor cells will die, but sufficient numbers survive to generate bone tumors within 3–4 weeks. It is assumed that the tumor cells will preferably colonize areas associated with active bone resorption, although the importance of the putative stem cell niche in tumor cell colonization has become an area of increased study [ ]. The following section describes examples where i.t. injection of breast cancer cells has been used to investigate molecular pathways involved in bone metastasis and effects of therapy.
Mechanisms of bone metastasis formation
Fisher and colleagues used i.t. implantation of MCF-7 cells overexpressing PTHrP and OPG to examine the effect of local tumor production of OPG on the ability of breast cancer cells to establish and grow in bone [ ]. There was increased tumor growth and osteolysis in mice receiving MCF-7 cells overexpressing PTHrP and OPG, compared to that seen in mice receiving the parental MCF-7 cells. In marked contrast, administration of recombinant Fc-OPG reduced tumor growth and limited osteolysis even in mice injected with OPG-overexpressing tumor cells. The data suggest that there may be a difference between the biological actions of tumor cell–derived OPG, compared to therapeutically administered recombinant constructs like OPG-Fc. This model has also been used to demonstrate that increased bone turnover due to dietary calcium deficiency promotes tumor growth in bone, suggesting that breast cancer patients may benefit from treatments to normalize calcium both in the adjuvant and advanced setting [ ].
Effects of therapies
The potential antitumor effect of OPG and the antiresorptive agent ibandronate (IBN) has been investigated following i.t. implantation of MDA-MB-231 breast cancer cells [ ]. Tumors could be detected 10 days following implantation, and mice were then treated with vehicle, OPG, IBN, or IBN and OPG for a week. The development of osteolytic lesions was inhibited by all treatments compared to vehicle controls, accompanied by a reduction in tumor area. OPG and IBN, alone and in combination, produced a similar increase in cancer cell apoptosis and a decrease in cancer cell proliferation. The authors concluded that there was no additional benefit combining OPG and IBN compared to using the single agents, as both agents mainly affect tumor volume indirectly by reducing tumor-induced bone disease.
The effects of combination therapy with doxorubicin and ZOL on tumor growth in bone have been studied using an i.t. model [ ]. MDA-MB-436-GFP cells were injected into the tibiae of female CD1 mice that received either vehicle, doxorubicin, ZOL or doxorubicin followed 24 h later by ZOL, weekly for 6 weeks. Combination therapy caused a significant reduction in tumor burden in bone, accompanied by induction of apoptosis and suppression of tumor cell proliferation. Although this intensive schedule reduced tumor growth and preserved bone integrity, the tumor burden remained significant, indicating that repeated cycles of treatment are required to eradicate established bone metastasis.
Syngeneic models
Xenograft models do not include the initial steps of metastasis, and syngeneic models have therefore been developed, using implantation of murine tumor cells (like 4T1) into the mammary fat pad of immunocompetent BALB/c mice, resulting in development of bone and visceral metastasis within 3–4 weeks. 4T1 is a mammary carcinoma originally derived from a spontaneously arising mammary tumor in BALB/cfC3H mice [ ]. This model includes the growth of the initial mammary tumor in an anatomically correct site, followed by metastatic spread to a range of distant organs, thereby mimicking human breast cancer. The main advantage is that it can be used to study the role of the immune system, both in tumor progression in bone as well as in response to anticancer therapy. However, syngeneic models do have some important limitations. The murine tumor cells differ from human breast cancer cells in the genetic makeup and potentially in their growth characteristics and responses to therapy. In addition, the primary mammary tumors grow quickly once established, leaving a limited window of opportunity for studies of the metastases unless the primary tumor is removed. Most importantly, syngeneic models are not bone metastasis specific, and a large number of animals are required in order to generate sufficiently large groups with skeletal involvement. There is also considerable variability in the time it takes for metastasis to appear in different sites and in different animals, presenting particular challenges in terms of when to initiate treatment, the frequency and length of therapeutic interventions, as well as which outcome measures to use.
Mechanisms of bone metastasis formation
A version of the model using 4T1-12B-Luc cells has been developed for studies of the role of innate and acquired immunity in metastasis [ ]. Implantation of the parental 4T1-12B cell line resulted in only two of six animals developing bone metastases, and these were evident throughout the skeleton including skull, sternum, and ribs as well as long bones. The animals also displayed brain, intestine, and kidney metastasis. In contrast, implantation of a variant of the cell line (4T1-1V) caused skeletal metastases in six of seven animals. This study highlights the differential ability of closely related cell lines to spread to the bone microenvironment, and illustrates the some of the problems associated with studying bone metastases in a model where visceral metastases are prevalent. Populations of 4T1 cells with increased capacity to form bone metastasis were used by Rose and colleagues to identify novel genes involved in bone metastases formation [ ].
Effects of therapies
Effects of the antiresorptive agent ZOL on metastasis have been studied using orthotopic implantation of 4T1/luc cells in the mammary fat pad in female BALB/c mice [ ]. Animals received a single or four i.v. injections of ZOL before metastasis formation was evaluated. Only the highest dose of ZOL (5 μg/mouse) caused a marked reduction in bone, lung, and liver metastases, resulting in prolonged overall survival. ZOL induced increased levels of tumor cell apoptosis in the bone metastases, but not in the visceral metastases. These results support that in addition to causing tumor cell death in bone, ZOL affects breast cancer metastasis to visceral organs, potentially by inhibiting migration and invasion of breast cancer cells.
The role of vacuolar ATPase (V-ATPase) in CIBD has been investigated by administration of the V-ATPase inhibitor FR202126 following injection of 4T1 cells into the mammary fat pad of immunocompetent BALB/c mice [ ]. Administration of the V-ATPase inhibitor caused a reduction in the levels of osteolysis, suggesting that it might be a useful antiresorptive agent, potentially able to alleviate bone pain caused by stimulation of acid-sensitive receptors.
In vivo models of prostate cancer bone metastases ( Table 4.2 )
Bone metastasis is also a common clinical feature of advanced prostate cancer, associated with accelerated levels of bone turnover and considerable morbidity. Prostate cancer bone metastases are generally described as osteosclerotic, involving excess bone formation, but there is also a substantial increase in the osteoclastic bone resorption providing a lytic component. There are relatively few cell lines and in vivo model systems available for studies of bone metastases in prostate cancer, and none of these accurately represent the osteoblastic/osteosclerotic phenotype that characterizes human disease [ , ]. Although mixed lesions have been described, the models are mainly restricted to mimicking the osteolytic component of the disease, illustrating our limited understanding of the dysregulation of bone formation caused by prostate cancer cells [ ]. A recent comprehensive review of translational models of prostate cancer bone metastasis highlights the strengths and weaknesses of the current models, as well as the remaining challenges, including how to recapitulate elements of the immune system and tumor heterogeneity [ ].
Table 4.2
Current main animal models used in studies of prostate cancer bone metastases.
| Model type | Cells administered | Route of administration | No of cells injected | Duration | Disease characteristics and features of models |
|---|---|---|---|---|---|
| Xenograft immunodeficient animals | PC-3 PC-62 |
Intracardiac | 2 × 10 5 1 × 10 5 |
2–5 weeks | Lytic lesions in hind limbs and craniofacial regions. Elevated bone turnover |
| LNCaP PC-3 LuCaP35 LuCaP23.1 |
Intratibial | 1–2 × 10 5 | Up to 30 weeks Up to 5 weeks Up to 7 weeks Up to 25 weeks |
Mixed osteolytic and osteoblastic lesions Osteolytic lesions Osteolytic lesions Osteoblastic lesions |
|
| PCa1-met | Orthotopic prostate | Tissue implanted | 2–5 weeks | Multifocal disease lymph nodes, lung, liver, kidney, spleen, bone | |
| Syngeneic TRAMP model immunocompetent | Spontaneous Transgenic Andenocarcinoma of mouse prostate |
N.A | N.A | 12 weeks onward | Multiorgan disease including lymph nodes, lungs, kidney, adrenal gland, bone |
| Human-to-human models | PC-3 DU145 LNCaP |
Direct injection into implanted human fetal bone fragments | 1 × 10 4 1 × 10 4 5 × 10 4 |
Tumor cells injected 4 weeks following implantation of human fetal bone Animals culled 6 weeks later |
Mets to human bone osteolytic lesions Mets to human bone osteolytic lesions Osteoblastic and osteolytic lesions |
| LNCaP PC-3 |
Intravenous Adult human bone implanted |
1 × 10 7 5 × 10 6 |
Tumor cells injected 3–4 weeks following implantation of human adult bone Animals culled 8 weeks later |
Metastases to human bone in 35% of cases—osteoblastic lesions Metastases to human bone in 65% of cases—osteolytic lesions |
You're Reading a Preview
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here