**2.4 Brain metastases biology**

162 Advances in Cancer Therapy

**2. Epidemiology, pathology and prognosis of patients with brain metastasis**  Brain metastases occur in 20% to 40% of cancer patients. Posner and Chernik, from 1970 to 1976 at the Memorial Sloan-Kettering Cancer Center, autopsied 3219 cancer patients. They found that 24% of these patients had intracranial metastases and the 20% had leptomeningeal metastases (Posner and Chernik 1978). Other studies reported similar percentage even if the methodology is slightly different (Baker et al. 1942, Chason et al 1963, Nussbaum 1996). In decreasing frequency, lung cancer, unknown primary, breast, melanoma, renal, and colon cancers are the most common tumors to metastasize to the brain (Pectasides et al. 2006, Norden et al. 2005, Posner and Chernic 1978, Chason et al 1963,

Another important aspect is the precise localization of the metastases in the brain (Globus et al. 1942, Tom et al. 1946, Zimm et al. 1981, Tikhtman et al 1995). Approximately the 80% of metastases are localized in the cerebral hemisphere, 17% in the cerebellum and 3% in the brain stem (Tikhtman et al 1995). Delattre and contributors have also analyzed the localization of metastases in the specific regions of the cerebral hemisphere (Delattre et al. 1988). They found that brain metastases involved the frontal area for 21%, the parietal and the temporoparietal-occipital for 19%. These authors have also outlined that metastasis located preferentially at the junction of gray and white matter. Hwang et al. recently reexamined the importance of the vascular border zone and the gray and white matter junction in the distribution of brain metastases and agreed with Delattre and contributors (Hwang et al. 1988). They found in 302 metastatic brain lesions studied, that gray and white matter junction was the preferred site for 64 % of the brain metastases and the vascular border zones were the site of predilection for the 62%. These results support the notion that metastatic emboli tend to lodge in an area of sudden reduction of vascular caliber (gray/white matter junction) and in the most distal vascular field area (Border zone)

Most metastases appear round well-demarcated lesions that displace rather than invade the surrounding brain parenchyma. The lesions can vary in size ranging from microscopic to masses of 1-4 centimeters in diameter. The histopathologic features are similar to that of

Microinvasion is present although the majority of metastatic lesions appear well demarcated and sometimes with reactive astrocytosis surrounding the metastatic area (Shaffrey et al. 2004). A new vasculature can appear at peripheral zones and is part of the edema. In central areas, necrotic areas are present (Nathoo et al. 2005). Meningeal metastases diffuse into the

Untreated brain metastases have a dismal prognosis, generally no greater than 1-5 months. Gaspar et al. 2000, studying a RTOG (Radiation Therapy Oncology Group) data base, performed a Recursive – Partitioning Analysis [RPA] on 1200 patients. They identified the factors able to influence the prognosis of these patients (Gaspar et al. 2000, D'Ambrosio et al. 2007). Among the several prognostic factors, Karnofsky Performance Status (KPS) has been

subarachnoid space with accumulation around blood vessels (Shaffrey et al. 2004).

Nussbaum et al. 1996).

(Hwang et al. 1988).

tumor of origin.

**2.2 Shape of brain metastases** 

**2.3 Prognosis of brain metastases** 

**2.1 Localization of brain metastases** 

Researchers have gained insight into the mechanisms by which metastatic cells arise from certain primary tumors (i.e. breast, melanoma) and metastasize to brain. These findings have been obtained in a mouse model (Kang et al. 2004, Nicolson 1993, Beasley et al. 2011). These authors have outlined that metastases formation are not due to patterns of initial cell arrest, motility, or invasiveness, but rather to the ability of metastatic tumor cells to grow in that environment, this in agreement with the Paget hypothesis of seed and soil. In other words, the formation of metastasis requires the right cells with the compatible environment (Fidler 2002, 2010, Deichman 1998).

Metastatic process is a high selective non-random process consisting in a series of linked sequential steps. In a heterogeneous population of primitive tumor cells only some are able to survive and to lodge at distant sites (Shaffrey et al. 2004). The outcome of cancer metastases depends on multiple interactions between metastatic cells and homeostatic mechanisms. Each metastatic step is selective and if the various steps are not completed, the metastatic process may fail. This is probably why only 0.01% of the cells that reach the circulation form a metastatic colony (Shaffrey et al. 2004, Kang et al 2004).

For tutorial purpose we can divide the metastatic process to the brain in the following steps (Fidler 2010, Deichman 1998):

A) process of selection among the heterogeneous tumor cells of origin into an aggressive and able to mobilize subpopulation; B) penetration of this selected subpopulation into the host circulation; C) localization into the microvasculature of the brain; D) crossing of Blood Brain Barrier (BBB); D) migration and growth in the brain structure.

The process of selection (A) is the result of different pressures exerted by the tumor microenvironment and the genetic instability intrinsic to the tumor cells living in that environment. For different pressures we intend: tumor hypoxia, different metabolic advantageous tumor micro-area, immunologic pressure, presence or absence of an angiogenesis process. These different pressures can select among cells genetically unstable, cells able to survive in a different environment, to mobilize and to reach the blood stream (Deichman 1998). To gain access to the general circulation and to colonize to distant organs metastatic cells must invade tumor associated vasculature. The molecular mechanisms controlling the penetration of blood vessels are not completely understood (Beasley et al 2011, Fidler 2002, 2010, Deichman 1998). Once these cells have reached the blood vessels, various mechanisms are needed to survive both the immune system and the shear stress. To avoid the identification the immune cells, metastatic cells shielded under an agglomerate of platelets and red blood cells (Deichman 1998)]. Other important mechanisms are the resistances by metastatic cells to the apoptotic effects of reactive oxygen radicals produced

Brain Metastases: Biology and Comprehensive

Denkins et al. 2004).

as breast cancer and melanoma.

demonstrated by Palmieri et al. (2007).

Strategy from Radiotherapy to Metabolic Inhibitors and Hyperthermia 165

Whether the progressive growth of brain metastases depends on neovascularisation is also unclear. As outlined by Bucana: "immunohistochemical and morphometric analyses show that the density of blood vessels within experimental metastases in the brains of nude mice, or within brain metastases derived from human lung cancer, is lower than in the adjacent, tumor-free brain parenchyma. However, blood vessels associated with brain metastases are dilated and contain many dividing endothelial cells. Immunohistochemical analysis also reveals that tumor cells located less than 100 micrometer from a blood vessel are viable, whereas more distant tumor cells undergo apoptosis. The blood-brain barrier is intact in and around experimental brain metastases smaller than 0.25 mm in diameter, but is leaky in larger metastases "(Bucana et al. 1999). Regarding melanoma other authors have found that neurotrophins (NTs) can promote brain metastases. NTs enhance the production of ECM degradation enzymes such as heparanases. Heparanases do not only degrade ECM but also the basement membrane of BBB Nathoo et al 2005, Menter et al 1994, Marchetti et al 2003,

The potential of angiogenesis in breast metastases has been further been studied. VEGF has been reported to increase the penetration of metastatic MDA-MB-231 breast carcinoma and to play a role in brain metastases dormancy in the absence of inhibitory antiangiogenic factors (Kim et al. 2004, Santarelli et al 2007, Palmieri et al. 2007, Yano et al,. 2000, Kaplan et al. 2005, Chen et al. 2007). Yano and collaborators however do not agree regarding the importance of VEGF on brain metastases and in an experimental mouse model using six different human cancer cell lines has reported that VEGF expression was necessary but not

Recently, the idea of premetastatic niche is leading the way (Kaplan et al 2006). Some authors support that the arrival of bone marrow-derived hematopoietic progenitors cells in distant sites represent early changes in the local environment (premetastatic niche) that dictates the pattern of metastatic spread and explains tumor dormancy (Kaplan et al. 2005,2006). Santarelli et al. 2007, outline that the reactive monocytosis and activated microglia present in the premetastatic niche increase the local inflammatory response and can induce the growth of tumor cells transplanted to the brain. Furthermore these authors outline that it is plausible that the brain is able to generate the adequate environment in anticipation and preparation of the ensuing metastatic colonization (Santarelli et al. 2007). Other studies have evidenced new mechanistic insights regarding some kind of tumors such

**Her-2 receptor**. Overexpression of Her-2 receptor in breast carcinoma seems not only correlated to a poor prognosis but also to an increased colonization to the brain as

**Metabolic factors.** Regarding only breast carcinoma, Chen et al (2007) have evidenced by proteomic analysis, that brain- metastasizing cancer cells over expressed enzymes involved in aerobic glycolysis and tricarboxylic acid cycle (TCA) cycle. From this study the authors outline that breast cancer that colonize the brain are able to adapt to the energy metabolism

**Stat3.** In melanoma, the activation and over expression of Stat3 (Signal Transducer and Activator of Transcription 3), as reported by Tong-Xin is associated to brain metastases and

**Metastasis suppressor genes.** Recently seven Metastasis Suppressor Genes (MSGs) have been identified. These genes have no effect on the growth of primary tumors but have the

of the brain or develop metabolism able to survive in that specific environment.

might be considered a new potential target in this clinical situation (Tong et al 2006).

sufficient for the production of brain metastasis (Yano et al. 2000).

by macrophages and neutrophils and the production of prostaglandins of E type (Fidler 2002). Furthermore the adhesion of the platelets on the metastatic cells induces a hypercoagulable state that increases the metastatic potential. In fact this adhesion increases the resistance to both the immune system and the shear stress (Kehrli 1999).

The most common metastatization to brain occurs by hematogenous route. One route is via the general circulation. In fact in the resting state, the brain receives 15% to 20% of the body's blood flow, thus making it likely that circulating tumor cells will reach the brain. The second, via the vertebral venous system (Batson's plexus) which explains the absence of lung metastases found in certain patients with lung cancer. This spreading way is disputed and not confirmed by all authors (Deeken et al. 2007).

Once metastatic cells have survived the circulatory stream, they may adhere to the endothelium, extravasate into the organ and then begin to proliferate in the new parenchyma. The arrest in the microcirculatory system is regulated by several factors, among them the multiple vascular adhesion molecules and the size of circulating emboli. Among adhesion molecules two families seem to be implicated: the selectins and the products of Immunoglobulin (Ig) genes and the integrins (Deeken et al. 2007). Integrins are major adhesion and signaling receptors that mediate cell migration and invasion (Shaffrey et al. 2004). In the case of human non small lung cancer (NSLC) the block of adhesion molecules integrin α3β1 has been demonstrated to significantly decrease the brain metastasis (Nathoo et al 2005).

After reaching the brain capillaries, the metastatic cells must cross the BBB, degrade the brain Extracellular Matrix (ECM) and invade the brain parenchyma. This interaction is tight regulated by the paracrine and autocrine growth mechanisms present in the brain (Nicolson 1993). BBB is constituted by brain endothelial cells associated with pericytes and astrocyte foot processes. Brain endothelial cells have continuous tight junction, no fenestrations and is highly selective in its permeability (Perides et al. 2006). In experimental melanoma the fibrinolytic system facilitates tumor cell migration across the BBB as demonstrated by Perides and his group (Perides et al. 2006). For metastasis to the brain from breast tumor, the cooperation between metastatic cells and astrocyte is of importance (Weil et al. 2005).

To determine why certain tumors produce site-specific metastases to the brain, Fidler and collaborators studied cellsfrom K-175 melanoma syngeneic to C3H/HeN mice and the B16 melanoma syngeneic to C57 Bl/6 mice (Fidler 1999, 2002, 2007). Regardless of the route of injection (internal carotid arteries-or directly into the cerebrum) K-175 produced melanocytic metastases in the brain parenchyma, whereas B16 cells produced lesions in the meninges and ventricles. These researches tried to understand which factors were responsible for the growth the melanoma cells in the specific areas (brain parenchyma, meninges). Some important aspects have been elucidated. For example B16 cells did not produce measurable gelatinase A activity whereas K- 175 cells did. The presence of gelatinase theoretically can facilitate cells extravasation and the growth into the parenchyma. However studies with hybrids of B16 and K-1735 able to produce gelatinase A, failed to grow into brain parenchyma. Fidler switched his research on the different growth factors present in the brain. He studied growth factors such as Epidermal Growth Factor (EGF), basic Fibroblast Growth Factor (bFGF), Platelet Derived Growth Factor (PDGF), however only Transforming Growth Factor-beta 2 (TGF-β2 )showed a greater concentration in the brain and was able to inhibit the growth of B16 and B16/K-1735 hybrid cells explaining the incapacity of these hybrids in producing intraparenchymal brain metastases (Fidler 1999).

by macrophages and neutrophils and the production of prostaglandins of E type (Fidler 2002). Furthermore the adhesion of the platelets on the metastatic cells induces a hypercoagulable state that increases the metastatic potential. In fact this adhesion increases

The most common metastatization to brain occurs by hematogenous route. One route is via the general circulation. In fact in the resting state, the brain receives 15% to 20% of the body's blood flow, thus making it likely that circulating tumor cells will reach the brain. The second, via the vertebral venous system (Batson's plexus) which explains the absence of lung metastases found in certain patients with lung cancer. This spreading way is disputed

Once metastatic cells have survived the circulatory stream, they may adhere to the endothelium, extravasate into the organ and then begin to proliferate in the new parenchyma. The arrest in the microcirculatory system is regulated by several factors, among them the multiple vascular adhesion molecules and the size of circulating emboli. Among adhesion molecules two families seem to be implicated: the selectins and the products of Immunoglobulin (Ig) genes and the integrins (Deeken et al. 2007). Integrins are major adhesion and signaling receptors that mediate cell migration and invasion (Shaffrey et al. 2004). In the case of human non small lung cancer (NSLC) the block of adhesion molecules integrin α3β1 has been demonstrated to significantly decrease the brain

After reaching the brain capillaries, the metastatic cells must cross the BBB, degrade the brain Extracellular Matrix (ECM) and invade the brain parenchyma. This interaction is tight regulated by the paracrine and autocrine growth mechanisms present in the brain (Nicolson 1993). BBB is constituted by brain endothelial cells associated with pericytes and astrocyte foot processes. Brain endothelial cells have continuous tight junction, no fenestrations and is highly selective in its permeability (Perides et al. 2006). In experimental melanoma the fibrinolytic system facilitates tumor cell migration across the BBB as demonstrated by Perides and his group (Perides et al. 2006). For metastasis to the brain from breast tumor, the

cooperation between metastatic cells and astrocyte is of importance (Weil et al. 2005).

To determine why certain tumors produce site-specific metastases to the brain, Fidler and collaborators studied cellsfrom K-175 melanoma syngeneic to C3H/HeN mice and the B16 melanoma syngeneic to C57 Bl/6 mice (Fidler 1999, 2002, 2007). Regardless of the route of injection (internal carotid arteries-or directly into the cerebrum) K-175 produced melanocytic metastases in the brain parenchyma, whereas B16 cells produced lesions in the meninges and ventricles. These researches tried to understand which factors were responsible for the growth the melanoma cells in the specific areas (brain parenchyma, meninges). Some important aspects have been elucidated. For example B16 cells did not produce measurable gelatinase A activity whereas K- 175 cells did. The presence of gelatinase theoretically can facilitate cells extravasation and the growth into the parenchyma. However studies with hybrids of B16 and K-1735 able to produce gelatinase A, failed to grow into brain parenchyma. Fidler switched his research on the different growth factors present in the brain. He studied growth factors such as Epidermal Growth Factor (EGF), basic Fibroblast Growth Factor (bFGF), Platelet Derived Growth Factor (PDGF), however only Transforming Growth Factor-beta 2 (TGF-β2 )showed a greater concentration in the brain and was able to inhibit the growth of B16 and B16/K-1735 hybrid cells explaining the incapacity of these hybrids in producing intraparenchymal brain metastases

the resistance to both the immune system and the shear stress (Kehrli 1999).

and not confirmed by all authors (Deeken et al. 2007).

metastasis (Nathoo et al 2005).

(Fidler 1999).

Whether the progressive growth of brain metastases depends on neovascularisation is also unclear. As outlined by Bucana: "immunohistochemical and morphometric analyses show that the density of blood vessels within experimental metastases in the brains of nude mice, or within brain metastases derived from human lung cancer, is lower than in the adjacent, tumor-free brain parenchyma. However, blood vessels associated with brain metastases are dilated and contain many dividing endothelial cells. Immunohistochemical analysis also reveals that tumor cells located less than 100 micrometer from a blood vessel are viable, whereas more distant tumor cells undergo apoptosis. The blood-brain barrier is intact in and around experimental brain metastases smaller than 0.25 mm in diameter, but is leaky in larger metastases "(Bucana et al. 1999). Regarding melanoma other authors have found that neurotrophins (NTs) can promote brain metastases. NTs enhance the production of ECM degradation enzymes such as heparanases. Heparanases do not only degrade ECM but also the basement membrane of BBB Nathoo et al 2005, Menter et al 1994, Marchetti et al 2003, Denkins et al. 2004).

The potential of angiogenesis in breast metastases has been further been studied. VEGF has been reported to increase the penetration of metastatic MDA-MB-231 breast carcinoma and to play a role in brain metastases dormancy in the absence of inhibitory antiangiogenic factors (Kim et al. 2004, Santarelli et al 2007, Palmieri et al. 2007, Yano et al,. 2000, Kaplan et al. 2005, Chen et al. 2007). Yano and collaborators however do not agree regarding the importance of VEGF on brain metastases and in an experimental mouse model using six different human cancer cell lines has reported that VEGF expression was necessary but not sufficient for the production of brain metastasis (Yano et al. 2000).

Recently, the idea of premetastatic niche is leading the way (Kaplan et al 2006). Some authors support that the arrival of bone marrow-derived hematopoietic progenitors cells in distant sites represent early changes in the local environment (premetastatic niche) that dictates the pattern of metastatic spread and explains tumor dormancy (Kaplan et al. 2005,2006). Santarelli et al. 2007, outline that the reactive monocytosis and activated microglia present in the premetastatic niche increase the local inflammatory response and can induce the growth of tumor cells transplanted to the brain. Furthermore these authors outline that it is plausible that the brain is able to generate the adequate environment in anticipation and preparation of the ensuing metastatic colonization (Santarelli et al. 2007). Other studies have evidenced new mechanistic insights regarding some kind of tumors such as breast cancer and melanoma.

**Her-2 receptor**. Overexpression of Her-2 receptor in breast carcinoma seems not only correlated to a poor prognosis but also to an increased colonization to the brain as demonstrated by Palmieri et al. (2007).

**Metabolic factors.** Regarding only breast carcinoma, Chen et al (2007) have evidenced by proteomic analysis, that brain- metastasizing cancer cells over expressed enzymes involved in aerobic glycolysis and tricarboxylic acid cycle (TCA) cycle. From this study the authors outline that breast cancer that colonize the brain are able to adapt to the energy metabolism of the brain or develop metabolism able to survive in that specific environment.

**Stat3.** In melanoma, the activation and over expression of Stat3 (Signal Transducer and Activator of Transcription 3), as reported by Tong-Xin is associated to brain metastases and might be considered a new potential target in this clinical situation (Tong et al 2006).

**Metastasis suppressor genes.** Recently seven Metastasis Suppressor Genes (MSGs) have been identified. These genes have no effect on the growth of primary tumors but have the

Brain Metastases: Biology and Comprehensive

< 70 and incomplete removal were significant factors.

with radiosurgery will be discussed later.

**Gamma Knife**, **Linac radiosurgery**.

**3.2 Radiation therapy** 

based radiosurgery.

Strategy from Radiotherapy to Metabolic Inhibitors and Hyperthermia 167

resected. In a third Group [C], 26 patients were resected for single metastasis and this group was used as control. From the comparison of the various subgroups the following results emerged: Group C and Group B obtained the same survival time and a multivariate analysis demonstrated that the only variables significantly affecting the survival were the groups of patients and the extent of the primary tumor (Bindal et al. 1993). The recurrence was similar in group B and C suggesting that an aggressive surgical approach may be useful. This approach has been confirmed by another study on 138 patients (Iwadate et al 2000). This study included other two variables: age and Karnofsky index. Age >60 years and Karnofsky

Radiotherapy is the mainstay therapy of brain metastases. Currently there are three major categories of radio therapeutic treatments of brain metastasis: WBRT, radiosurgery, stereotactic radiotherapy. There are two options for radiosurgery: Gamma knife, Linac –

*WBRT* has been demonstrated by Patchell 1990 and Vecht 1993 to increase life survival in association with surgery. The WBRT is a palliative procedure which aim is to achieve life prolongation, local control and improvement in quality of life. WBRT has been also used prophylactically, aimed to treat malignancies having high brain metastasizing affinity, such as small cell lung carcinoma (SCLC), leukemia and lymphoma (Meert et al. 2001, Brown et al. 2005). The most frequently applied doses range from 20 Gy to 30 Gy in 5-10 sessions. Doses from 30 to 36 Gy are used as prophylaxis in the case of SLC and from 12-18 for hematologic diseases (Alexander et al. 1995, Brown et al. 2005). The combination of WBRT

**Radiosurgery** is now possible because of the availability of CT and MRI and computer planning makes possible the delivery of high dose of radiation to a precise target tumor area. This delivering of precise high dose of radiation energy to a tumor is called radiosurgery. It can be achieved combining 3 elements: 1) stereotactic localization of metastatic lesion; 2) precise collimation of the radiation energy and 3) administration of the total dose coming from different points in space and intersected in a single target volume. The peculiarity of radiosurgery is the fall of dose at the target edges, this permit to concentrate the dose to the target tumor area sparing everything possible the healthy tissue surrounding the tumor (Lunsford et al. 1990). Two radiosurgical treatment facilities exist:

Historically, LeKsell in Sweden was the first to apply radiosurgery. Initially low energy xrays (280 kV) were used and concentrated stereotactically to the intracranial target. The technique was first accepted with skepticism; however after the initial studies by Lunsford et al. (1990) (University of Pittsburgh), radiosurgery has gained a considerable acceptance. Lately in 1967, Leksell, in collaboration with Larsson, developed according to the same

The **Gamma Knife** contains 201 cobalt sources of gamma rays arrayed in a hemisphere within a shielded structure. A primary collimator, forces all the emitted sources to a common focal area, then a secondary collimator adapts this primary focal beam to sizes from 4 to 18 mm, through computer software, to target to the corresponding size of brain metastasis. In this case the limiting size of this device are brain metastases with a major

principle of radiosurgery the first cobalt - 60 gamma unit (Gamma Knife).

ability to suppress metastases *in vivo*. Proteins that regulate different functions such as adhesion, migration, growth and differentiation are coded by these MSGs. These genes have been described for breast carcinoma (Seraj et al. 2000), melanoma (Leone et al. 1991) and prostate cancer (Dong et al. 1995).

Notwithstanding all these progresses the entire process of brain colonization remains actually poorly understood and better human and animal models are to be tried. It is our hypothesis that the peculiar metabolism of the normal brain with its high glucose uptake may explain the large incidence of metastases.
