Genes that Mediate Metastasis across the Blood-Brain Barrier

Abstract

Brain metastasis is an important cause of mortality in patients with cancer and represents the majority of all intracranial tumors. A key step during the metastatic journey of the cancer cell to the brain is the invasion through the blood–brain barrier (BBB). Nevertheless, the molecular mechanisms that govern this process remain unknown. The BBB has been blamed for limiting the access of therapeutic drugs to the brain, which provides a safe haven for the cancer cell in the brain and confers poor prognosis for the patient. Here, we explore the genes that control the transmigration of metastatic cancer cells across the BBB, offering new targets for the development of gene and cell therapies against brain metastases.

THE BLOOD-BRAIN BARRIER: A SELECTIVE GATEKEEPER

Brain metastases constitute the most common mass lesions in the brain. They require invasion of primary cancer cells from lung, breast, or skin, trafficking through the circulatory system, and colonization in the brain parenchyma (see Glossary) [1].

Some metastatic cancer cells possess the unique ability of crossing the blood-brain barrier (BBB), which is a selective gatekeeper composed of endothelial cells, pericytes, and the foot processes of astrocytes. The BBB separates the brain from systemic circulation and protects it from foreign infectious and/or toxic agents, including targeted therapies [2–4]. In fact, the BBB is estimated to block the transport of ~ 98% of molecules in systemic circulation [5].

By default, the network of closely faced endothelial cells and the continuous tight junctions renders the mechanism of metastasis through the BBB to be different from metastasis to other organs (Figure 1). Under physiological conditions, the BBB is so restrictive that not even ions can pass freely from one side to the other, making it challenging to study in experimental settings (Box 1). Nonetheless, >70% of brain metastases have shown leakage from blood vessels, indicating that the BBB is compromised in the setting of brain metastases [6]. Yet, the genetic drivers in cancer cells and in endothelial cells of the BBB that may be exploited to permit crossing the BBB remain unclear.

Figure 1: Anatomy and pathophysiology of the blood-brain barrier.

The metastatic cancer cell faces the BBB on its way to the brain. This barrier consists of multiple layers of different cell types that restrict the entrance of large macromolecules into the brain parenchyma. The innermost layer is formed by endothelial cells, which contain different transporters in their membranes and are connected through tight junctions. Pericytes attach to the abluminal site of the endothelial cells and are embedded in a sheet of extracellular matrix (ECM). The outermost layer consists of the end-feet of astrocytes. While the BBB restricts the entry of many molecules and cells, it is not an impenetrable barrier to transmigration of metastasizing cancer cells (1). Indeed, in this regard, it has been proposed that the brain can function as a sanctuary site for metastatic cells that effectively breach the BBB, where they are subsequently protected from the effects of chemotherapy and other agents that cannot penetrate the brain (2). Eventually, metastasizing cancer cells colonize their new microenvironment and grow into macrometastases (3).

Box 1. Techniques Used to Study Transmigration Across the Blood-Brain Barrier.

In vitro models:

• Monoculture models that consist of cerebral endothelial cells lined up on a semipermeable support under static culture conditions [105].

• Co-culture models that comprise endothelial cells, astrocytes, and/or pericytes to attempt to form paracellular tight junctions [105].

• Stem cell models that differentiate into cerebral endothelial cells that can line up into a BBB model. Neural stem cells can differentiate into neuronal cells and can help in generating a more complex neurovascular unit [106]. Mesenchymal stem cells can also be used to simulate pericytes as both cell types express similar cell surface markers [107, 108].

• The Transwell system as a side-by-side vertical diffusion model comprising a microporous semipermeable membrane of seeded cells inundated in culture medium that separates the vascular and parenchymal side compartments [107, 109, 110].

• Microfluidic-on-chips as a dynamic 3D model where endothelial cells are cultured in the lumen of fiber tubes inside a closed chamber. The neurovascular unit is seeded in the extraluminal compartment. A variable-speed pulsatile pump generates an intraluminal flow that is physiologically comparable to that observed in capillaries [105].

In vivo models:

• Intravenous injection of a compound, after which the analyte concentration is determined at different times in the brain, plasma, and cerebrospinal fluid [111].

• Brain perfusion technique allows a direct indication of the permeation of a substance across the BBB [112]. The compound of interest is dissolved and pumped into the vasculature that leads to the brain. At a specific time, perfusion is stopped, the animal is decapitated, and the amount of substance in the brain is determined. Radiolabeled compounds are frequently employed.

• Tomographic methods comprise PET and single photon emission-computed tomography (SPECT) that study brain uptake kinetics, cerebral blood flow, BBB integrity, and efflux mechanisms [113].

• Microdialysis sampling as a minimally invasive method that allows the monitoring of neurotransmitters in the brain. It can be used to explore the transport and metabolism of drugs and other substances at the BBB [114].

• Zebrafish offer an optically transparent model organism with an endothelial BBB that help assess the delivery of natural products, fluorescence dyes and drugs across the BBB [92, 115].

In situ models:

• Brain efflux methods complement in vivo models of the BBB by superimposing a vascular perfusion fluid replacement of the circulating blood in the animal [105].

Ex vivo models:

• Patient-derived circulating tumor cells or exosomes expanded ex vivo and then reintroduced into animal models to recapitulate metastatic transmigration across the BBB [60].

• Multi-organ microfluidic chip model consisting of two bionic organ units - a primary site and a brain with a functional BBB structure [43].

Given the poor prognosis and limited effective therapies [7–9], it is of crucial importance to understand the molecular mechanisms of brain-specific metastasis. The gene and/or protein expression profile of metastatic cells is a key determinant of site-specific metastasis formation. Definite expression formulations might dictate predilection to specific target organs. It is hypothesized that the brain, with its unique physiological barriers, dictates an expression profile from metastatic cells to interact with and/or penetrate the endothelial linings of the BBB.

Genetic alterations in tumor cells drive the metastatic capabilities of cancer cells and their ability to transmigrate through the BBB. This review aims to explore the genetics and biological mechanisms that mediate the transmigration process of the metastatic cancer cells through the BBB. Identifying genetic biomarkers might serve as novel diagnostic and predictive markers for metastasis progression and for development of new molecular targeted therapies against brain metastases.

ROUTES OF TRANSMIGRATION

Transmigration across the BBB is a complex process that is tightly regulated. The endothelial cells of the BBB are closely attached by tight junctions that restrict passive diffusion [10]. Pericytes, also, play a role in stabilizing the brain vasculature, regulating the expression of tight and adherent junctions, restricting transcytosis, synthesis of extracellular matrix components, and initiating neovascularization [11]. The basal membrane, constituted of endothelial cells and pericytes, provides another layer of insulation [11]. Astrocytic foot processes that are in close contact with the basal membrane connect the BBB to the neuronal brain parenchyma. This connection helps to control the diffusion restrictions across the BBB [12].

During the formation of brain metastases, transmigration through the BBB is a rate-limiting step. Two possible routes have been described: the paracellular pathway (between endothelial cells) and the transcellular pathway (across the endothelial cell) [3]. In vitro and in vivo studies have shown that cellular diapedesis can occur either by creating a paracellular gap or by establishing a transcellular pore (Figure 2).

Figure 2: Mechanisms of transmigration across the blood-brain barrier.

Metastatic cells utilize different approaches to cross the multiple cell layers that constitute the blood-brain barrier (BBB). Passage through the paracellular route (left) includes proteolysis of junctional adhesion molecules, such as JAM-B, using proteases and serum factors or intercalation between the layers of endothelial cells and disrupting the integrity of the BBB. The transcellular pathway (right) involves strong adhesion through integrins and the formation of a channel or a pore within the endothelial cell that allows the transmigration of the invading cancer cell into the brain.

The paracellular route across tight junctions allows passive diffusion of water, gases, and lipid-soluble molecules [10]. Physiologically, molecules greater than 400 Da are unable to cross through the BBB, as they are too large for passive diffusion. Charge and chemical (hydrophilic/lipophilic) properties also play a role in the transmigration process [13]. Upon contact with the endothelial lining of the BBB, metastatic cancer cells can disrupt the tight junctions by releasing proteases and/or miRNAs that compromise the integrity of the BBB and increase permeability [3, 14].

Transcellularly, solutes and nutrients can cross by passive diffusion or active transport. Lipid-soluble molecules can cross passively, whereas molecules like glucose and amino acids might necessitate carrier-mediated transporters and metabolic energy to cross the BBB [10].

In addition, the increased tumor heterogeneity of brain metastases and the overexpression/suppression of certain genetic variants can lead to the expression/downregulation of proteins that facilitate the paracellular or transcellular routes [15–19] (Table 1).

Table 1:

Genes and proteins that mediate migration of metastatic cancer cells through the blood-brain barrier

Gene/Protein	Expression in metastasis	Mechanism of action	Primary tumor	Refs
ADAM8	Overexpression	Upregulated MMP9	Breast cancer	[41]
Adenosine A2A receptor	Downregulation	Inhibition of the A2A receptor downregulates expression of CLDN5, occludin and ZO-1	NSCLC	[98]
AKR1B10	Overexpression	Affects the expression of MMP-2 and MMP-9 via MEK/ERK signaling	NSCLC	[43]
ALCAM	Overexpression	Increases cancer cell adhesion to endothelial cells	NSCLC	[70]
Angiopoietin-2	Overexpression	Changes ZO-1 and claudin-5 tight junction protein structures	Breast cancer	[22]
Annexin A1	Overexpression	Cancer cell expresses high levels of annexin A1 and releases it outside, facilitating entry into brain	SCLC	[83]
Cathepsin S	Overexpression	Induces the proteolysis of the junctional adhesion molecule, JAM-B	Breast cancer	[23]
CEMIP	Overexpression	Uptake of CEMIP+ exosomes by brain endothelial and microglial cells induce endothelial cell branching and inflammation leading to vascular remodeling and metastasis	Breast cancer	[58]
CLDN5	Downregulation	Untightens endothelial cellular adhesion and destabilize the tight junctions	Lung cancer	[95]
COX2	Overexpression	Facilitates the angiogenic formation of new blood vessels that help tumor cells connect to the circulation	Breast cancer	[71, 72]
CXCR4	Overexpression	Activates the PI3K/AKT signaling	Breast cancer	[80]
Desmin+	Overexpression	-	Breast cancer	[16]
FUT4	Overexpression	Codes for CD15 expression that increases adhesion of metastatic cancer cells to the BBB	Lung cancer	[86]
FUT7	Overexpression	Codes for CD15s expression that increases adhesion of metastatic cancer cells to the BBB	Lung cancer	[86]
GBP1	Overexpression	Induced by the T cells upon interaction with breast cancer cells vamps the metastatic potential	Breast cancer	[82]
HBEGF	Overexpression	Activated EGF receptors and their downstream signaling.	Breast cancer	[71, 73]
Heparanase	Overexpression	Degrading and compromising ECM structural integrity leading to the release of growth factors, chemokines, and angiogenic factors that are important for the passage of metastatic cancer cells. Promotes Akt signaling and stimulates PI3K- and p38-dependent endothelial cell migration and invasion	Melanoma	[50–52]
Laminin α2	Downregulation	-	Breast cancer	[16]
MFSD2a	Downregulation	Increased the propensity of brain metastasis by reducing DHA transport and altering lipid metabolism	Breast cancer Lung cancer Prostate cancer	[88]
miR-105	Overexpression	Targets the tight junctional protein ZO-1	Breast cancer	[55]
miR-143–3p	Overexpression	Inhibits VASH1 expression and triggers proteasomal mediated degradation of endothelial junctions	Lung cancer	[57]
miR-181c	Overexpression	Delocalize actin by downregulating phosphoinositide-dependent kinase-1 (PDPK1)	Breast cancer	[56]
miR-509	Downregulation	Increases RhoC and TNF-α that control the invasion of cancer cells and the trans-endothelial cell migration through MMP9.	Breast cancer	[94]
MLCK	Overexpression	Activated at the adhesion site, leading to regional myosin contraction and, eventually, intravasation of the tumor cell	Breast cancer	[81]
MMP1	Overexpression	ECM degrading capacity and facilitate the angiogenic formation of new blood vessels	Breast cancer	[38]
MMP2	Overexpression	Deregulation of the MAPK pathway	Breast cancer	[40]
PLEKHA5	Overexpression	Adaptor protein that mediates Phosphoinositide 3-kinases (PI3K) signaling	Melanoma	[78]
PLGF	Overexpression	Activation of Rho/ROCK followed by activation of ERK1/2 and Ser/Thr phosphorylation of occludin	SCLC Breast cancer Colorectal cancer Gastric cancer	[44–48]
Rab7	Downregulation	Increases endothelial recycling endocytic transcytosis	Breast cancer	[99]
Rho GTPase	Overexpression	Increasing actomyosin contractility	SCLC	[64]
S1P3	Overexpression	Release of IL-6 and CCL2, which causes relaxation of the tight junctions between endothelial cells	Breast cancer	[54]
Sdc1	Overexpression	Increased adhesion to the perivascular regions of the brain and led to eventual trans-endothelial migration	Breast cancer	[85]
SEMA4D	Overexpression	Binds to Plexin-B1 receptors in pericytes to promote pro-inflammatory responses that disrupts BBB endothelial tight junctions	Breast cancer	[61]
Seprase	Overexpression	Formation of proteinase-rich membrane domains by these enzymes is believed to contribute to the ECM degradation	Melanoma	[24]
Serine Protease	Overexpression	Apoptotic induction of the endothelial cells and disruption of tight junctions	Melanoma	[20]
Slug	Overexpression	Binds to the promoter region of MMP1 in breast cancer cells, leading to its overexpression	Breast cancer	[39]
ST6GALNAC5	Overexpression	Enhances adhesion of cancer cells to brain endothelial cells	Breast cancer	[71]
TNF	Overexpression	TNF receptor-1 pathway	Breast cancer	[21, 22]
uPA	Overexpression	Formation of proteinase-rich membrane domains by these enzymes is believed to contribute to the ECM degradation	Melanoma	[24]
VASH1	Downregulation	Triggers proteasomal mediated degradation of endothelial junctions	Lung cancer	[57]
Visfatin	Overexpression	Upregulation of the cytokine CCL2, which opens the tight junctions	SCLC	[53]
VLA-4	Overexpression	Induces cells to intercalate into the tight monolayer of endothelial cells causing its destabilization	Melanoma	[69]
αB-crystallin	Overexpression	Adhesion occurs via α3β1 integrin-dependent mechanism	Breast cancer	[84]
αv-integrin	Overexpression	Influence cancer cell attachment and angiogenesis	Breast cancer Lung cancer Renal cancer Melanoma	[67]

GENES MEDIATING BARRIER REMODELING AND PERMEABILITY ALTERATIONS

Brain metastatic cells are known to compromise BBB integrity. Using the paracellular route, they break junctional adhesion molecules while crossing the BBB [20] (Figure 3). Triggering specific signaling pathways can increase leakage and lead to pore formation. For example, in 4T1 models of triple-negative breast cancer (TNBC), the activation of tumor necrosis factor (TNF)-TNF receptor-1 pathway can result in pores of relatively large sizes (11–13nm) in the BBB, which ultimately leads to higher permeabilization [21, 22].

Figure 3: Anatomy of the tight junction.

Tight junctions between endothelial cells are formed by transmembrane proteins, such as junctional adhesion molecules (JAMs), cadherins, claudins, and occludins, and by cytosolic proteins, such as catenins and zonula occludens (ZO) proteins.

Transmigration through Proteolysis and Released Serum Factors

In breast cancer brain metastases, high expression of cathepsin S – a regulator of BBB transmigration produced by tumor cells and macrophages – is significantly associated with decreased survival in patients with breast cancer [23]. In preclinical studies with MDA-MB-231 breast cancer cells, depletion of macrophage and tumor cell reserves of cathepsin S significantly decreased brain metastases [23]. The release of cathepsin S by metastatic cells induces the proteolysis of the junctional adhesion molecule, JAM-B [2]. This in turn allows the metastatic cell to cross the BBB. Furthermore, cathepsin S inhibitors were developed and were successful in significantly reducing brain metastasis in vivo [23]. In malignant melanoma, the release of serine proteases by metastatic melanoma cell lines, A2058 and B16/F10, induced apoptosis of the endothelial cells and disrupted the continuity of tight junctions [20]. This allowed the transmigrating melanoma cells to overcome the BBB. Protease systems, including seprase and urokinase-type plasminogen activator (uPA), have been shown to facilitate transmigration of LOX melanoma cells through brain endothelial cells both in vivo and in vitro [24]. Several studies showed that these enzymes are significantly associated with brain metastasis [25–33]. The formation of proteinase-rich membrane domains by these enzymes is believed to contribute to extracellular matrix (ECM) degradation [24], which subsequently increases the invasive ability of cancer cells across the BBB. Metastatic melanoma cells, for example, have been shown to express several types of proteases including uPA [24], seprase [27], and matrix metalloproteinases (MMP) [34–36] that disintegrate tight junctions. Translational studies revealed that targeting these proteases with the protease inhibitor Pefabloc markedly reduced the metastatic potential of melanoma cells [20].

MMP1 and MMP2 possess ECM degrading capacity and facilitate the angiogenic formation of new blood vessels to help tumor cells connect to the circulation [37, 38]. Targeted knockdown of MMP1 in brain metastatic MDA-MB-231 cells reduces primary tumor growth and significantly inhibits brain metastasis [38]. Furthermore, MMP1 upregulation in resistant metastatic breast cancer cells is controlled by the Slug protein [39]. Slug expression resulted in MMP1 upregulation in MCF-7/ADR cells and was significantly associated with reduced drug sensitivity [39]. However, Slug knockdown reduced MMP1 expression and enhanced the drug sensitivity of breast cancer cells [39]. Slug was shown to bind to the promoter region of MMP1 in breast cancer cells, leading to its overexpression. MMP2, also, plays a role in the invasion and metastasis of breast cancer cells to the brain. Extracellular signal-regulated kinases (ERK)1/2 and mitogen-activated protein kinase (MAPK) have been correlated to MMP2 expression in ENU1564 cells. It was found that MMP2 plays a role in the development of breast cancer brain metastases through the deregulation of the MAPK pathway [40]. Moreover, MMP9 seems central in other works relating to brain metastasis. The metallopeptidase domain 8 (ADAM8) was found to promote early metastatic processes in breast cancer through trans-endothelial migration across the BBB [41]. ADAM8 overexpression upregulated MMP9 in MDA-MB-231 cells and led to the shedding of P-selectin glycoprotein ligand-1 (PSGL-1), a substrate of ADAM8 that promotes adhesion of leukocytes to endothelial cells [41]. Thus, the major effect caused by shedding of PSGL-1 is an increase in mobility and invasiveness into the BBB [42]. As such, MMP9 and ADAM8 are plausible targets for brain metastasis formation. AKR1B10 also affects the expression of MMP-2 and MMP-9 via MEK/ERK signaling [43]; protein expression of AKR1B10 is significantly elevated in brain metastases from non-small cell lung cancer (NSCLC) [43]. Silencing AKR1B10 in brain metastatic PC-9 cells suppressed their transmigration through the BBB [43], suggesting that AKR1B10 could be a prospective therapeutic target for NSCLC brain metastases.

In small cell lung cancer (SCLC), placental growth factor (PLGF) was significantly elevated in the serum of patients with brain metastases and in NCI-H250 cells. PLGF, which belongs to the vascular endothelial growth factor (VEGF) subfamily, binds to its receptor, VEGF receptor-1 (VEGFR-1), on brain endothelial cells [44]. Subsequently, the intracellular Rho/ROCK signaling is activated and followed by ERK1/2 activation [44]. ERK1/2 activation causes Ser/Thr phosphorylation of occludin, a tight junction protein. This results in junctional disassembly and allows for the trans-endothelial migration of SCLC cells [44]. In breast cancer, several studies showed that elevated levels of PLGF were significantly associated with metastasis and mortality [45, 46]. Similar results were also reported in metastatic colorectal and gastric cancers [47, 48], which suggest that PLGF could serve as a possible target for therapeutic regimens.

Neurotrophins are proteins that promote the survival, development, and function of neurons [49]; nevertheless, they can act as mediators of BBB passage for cancer cells. Melanoma cells express receptors that have a high affinity to brain-derived neurotrophins. The binding of neurotrophins to their receptors induces the release of heparanase [50]. Heparanase, an endoglycosidase, is preferentially expressed in brain metastases and its overexpression in tumor cells confers an invasive phenotype in experimental animals [51]. Mechanisms by which heparanase facilitates cancer progression likely involves degrading and compromising ECM structural integrity [52]. This leads to the release of growth factors, chemokines, and angiogenic factors from the ECM that are important for the passage of metastatic cancer cells. In melanoma brain metastases, heparanase is highly expressed [52]. Pre-treating melanoma cells with heparanase leads to a significant increase in the number of cells that invade the brain [52]. These observations and the unexpected identification of a single functional heparanase suggest that the enzyme is a promising target for anti-cancer drug development.

Visfatin, a new pro-inflammatory adipocytokine, is predominantly released by macrophages in visceral fat tissue and is involved in various malignant tumors. In SCLC, visfatin was found to induce the upregulation of the cytokine, CCL2, in NCI-H446 cells [53], which opens tight junctions and promotes metastatic migration. In breast cancer brain metastases, S1P3 is overexpressed as a consequence of the neuroinflammatory reaction induced by the BBB-permeable cancer cells. Astrocytes that overexpress S1P3 release interleukin (IL)-6 and CCL2, which cause relaxation of the tight junctions and increases the BBB permeability [54]. Inhibition of S1P3 leads to the tightening of the BBB both in in vitro and in vivo settings using MDA-MB-231-BR, JIMT-1-BR, and SUM190-BR cells [54], hinting that BBB permeability can be altered via S1P3 regulation.

MiRNAs secreted by cancer cells can mediate crosstalk between metastatic cells and the host microenvironment. In MDA-MB-231 metastatic breast cancer, secreted miR-105 targets the tight junctional protein zonula occludens (ZO)-1, allowing the cancer cells to transmigrate into the brain [55]. Non-metastatic MCF-10A cancer cells exhibit metastatic properties in distant organs when miR-105 is overexpressed, whereas its inhibition removes these abilities [55]. It has been shown that miR-105 can be identified during the pre-metastatic stage, and its levels in the blood and tumor can reflect ZO-1 expression and metastatic progression in early-stage breast cancer [55, 56]. miR-181c, also, has been shown to delocalize actin by downregulating its target gene, phosphoinositide-dependent kinase-1 (PDPK1) [56], which leads to alterations in actin dynamics. In lung cancer brain metastases, miR-143–3p was upregulated when compared to primary cancer tissues [57]. Further investigation revealed that it can increase the invasion capability of an in vitro BBB model by targeting the binding sites of vasohibin-1 (VASH1) to inhibit its expression, triggering proteasomal mediated degradation of endothelial junctions [57]. This implicates a causal role for the miR-143–3p/VASH1 axis in lung cancer brain metastasis. Inhibition of such miRNAs and/or the suppression of their extracellular carriers decrease invasiveness of MDA-MB-231 breast cancer cells through BBB [56].

Tumor-secreted exosomes can also contribute to the transmigration across the BBB. Proteomic analysis identified the protein CEMIP to be increased in exosomes from brain-specific metastases [58]. CEMIP depletion in tumor cells impaired brain metastasis, disrupting invasion and cancer cell attachment to the brain vasculature [58]. The uptake of CEMIP⁺ exosomes by brain endothelial and microglial cells induce endothelial cell branching and inflammation in the perivascular niche by upregulating the pro-inflammatory cytokines encoded by Ptgs2, TNF and Ccl/Cxcl that are known to promote brain vascular remodeling and metastasis [56, 59]. In patient-derived exosomes and xenografts, CEMIP was elevated. It further predicted brain metastasis progression and overall survival [58].

Circulating tumor cells (CTCs) expanded ex vivo from patients with breast cancer can generate brain metastases in mice. Genome-wide sequencing analyses identified SEMA4D as a regulator of tumor cell transmigration through the BBB [60]. Semaphorins are a large family of secreted or transmembrane proteins that have versatile roles in axonal guidance and the immune system [61]. Most of the effects of SEMA4D are mediated by its interaction with Plexin-B1 receptors in pericytes, which promote pro-inflammatory responses [61]. Studies in CNS diseases revealed that SEMA4D disrupts endothelial tight junctions of the BBB [62, 63], driving transmigration through the paracellular route. Overexpression of the SEMA4D in the primary tumor site was associated with brain-specific metastases. In addition, CTC lines BRx50 and BRx42, which are brain metastatic, exhibited higher SEMA4D expression levels compared to CTC lines with no brain tropism in mice [60].

Physical and Anatomical Alterations of the Tight Junctions

Physical alterations in the BBB increase its permeability. In breast cancer brain metastases, the pericyte protein, desmin, is overexpressed which leads to remodeling of the BBB and increased metastasis of human MDA-MB-231-BR-Her2 and murine 4T1-BR5 cancer cells, suggesting a molecular target that modifies permeability [15]. Elevated desmin levels were also found in human craniotomy specimens of breast cancer brain metastases [16]. The decreases in astrocytic basement membrane component laminin α2 and the CD13 pericyte subpopulation are also a feature of increased BBB permeability in patients with breast cancer brain metastases [16].

Rho GTPases are a family of small signaling G proteins and a subfamily of the Ras superfamily. The members of the Rho GTPase family have been shown to regulate many aspects of intracellular actin dynamics. In SCLC, they are activated during trans-endothelial migration, thereby facilitating the breakdown of intercellular junctions through increasing actomyosin contractility [64]. The inhibition of their effect can prevent the trans-endothelial migration of disseminated SCLC cells across the BBB [65, 66].

GENETIC MEDIATORS OF ADHESION AND EXTRAVASATION

The transcellular route of brain metastasis involves genetic mediators of adhesion that drive the expression of integrins, endothelial receptors, and cellular proteins. These proteins and cellular components enhance the attachment of the metastatic cancer cell to the BBB and promote its migration, passively or actively, through the endothelial cell layer.

Adhesion is often facilitated by integrins that influence cancer cell attachment and angiogenesis and promote invasion and migration of tumor cells. In brain metastases of breast, lung and renal carcinomas and in malignant melanoma, the strong αv-integrin expression – especially of αvβ3 and αvβ8 integrins – in patient-derived samples hints at its involvement in the formation of brain metastasis [67, 68]. As such, αv-integrins may be diagnostic biomarkers and therapeutic targets for patients with metastatic cancers (see Clinician’s Corner). Very Late Antigen-4 (VLA-4) is another integrin that is expressed on the cell surfaces of stem cells, progenitor cells, and other inflammatory cells. It plays an important role in mobilizing leukocytes to the site of inflammation through cellular adhesion. In patient-derived samples, it was found that more than 90% of all human melanoma brain metastases are positive for VLA-4 expression [69]. VLA-4 was found to enhance the adhesion of melanoma hMel-CD49d cancer cells to the endothelial cells and allowed the incorporation of cancer cells in between the endothelial monolayer, causing disruption of the BBB. Inhibiting VLA-4 activity annulled the adhesion of brain metastatic cells with the BBB and prevented extravasation of cancer cells [69]. It was concluded that VLA-4 expression increases the risk of metastasis across the BBB, and VLA-4 was proposed as a therapeutic target for the inhibition of metastatic formation in the brain.

Clinician’s Corner.

• Brain metastasis portends poor prognosis as there are limited effective therapies that can cross the blood-brain barrier (BBB).

• Understanding the transmigration routes of cancer cells across the BBB is essential for the development of targeted therapies and preventive measures that can target metastatic cancer cells before entering the brain.

• Genes and proteins regulating transmigration present an opportunity for developing diagnostic assays that can help in the early detection and prevention efforts against brain metastasis.

• Gene therapy that relies on overexpressing genes that repress the transmigration of metastatic cancer cells or downregulating genes that drive transmigration across the BBB is a therapeutic approach that can improve survival rates of patients with brain metastasis

The role of ALCAM in brain metastasis formation has been investigated in NSCLC. CRISPR/Cas9 gene ablation of ALCAM in the NCI-H460 NSCLC cell line showed a significantly decreased cell adhesion capacity to human brain endothelial cells [70]. Intracranial and intracardiac cell injection of NCI-H460 cells in NMRI-Foxn1nu/nu mice revealed that significantly lower tumor cell dissemination in mice injected with ALCAM-Knockout cells in both mouse models and that the number and size of brain metastases were significantly diminished in ALCAM depleted tumors [70]. This suggests that increased ALCAM expression promotes the formation of brain metastases in NSCLC by increasing tumor cell dissemination and interaction with the brain endothelial cells.

In human-derived samples of breast cancer brain metastasis, cyclooxygenase (COX2), heparin-binding EGF (HBEGF), and α2,6-sialyltransferase (ST6GALNAC5) were identified as mediators of cancer cell passage through the BBB [71]. COX2 is known to promote specific extravasation of tumor cells into lung capillaries to seed pulmonary metastasis. When COX2 is expressed in human breast cancer cells, along with epiregulin (EREG) and MMPs, it facilitates angiogenesis [72]. EREG, however, has been shown to increase tumorogenicity and invasion of breast cancer cells to the lungs, but not brain. Still, COX2 knockdown inhibited the transmigration of breast cancer cells across the BBB [71]. This raises the question of whether the passage of cancer cells through different organs can prime them to express genes that mediate their passage through the BBB. HBEGF, which plays an important role in various biological processes such as heart development and maintenance has also been shown to play a role in cellular adhesion to the BBB. Its mode of action involves the activation of EGF receptors (EGFR) and their downstream signaling. In cancer, HBEGF expression has been associated with malignant tumorigenicity, metastasis, and the development of resistance to therapy [73]. In in vitro and in vivo settings, inhibition of HBEGF has been shown to halt tumor growth and progression [71, 73]. Knockdown of HBEGF along with COX2 and EREG has been associated with inhibition of breast cancer cell permeability across the BBB [71].

ST6GALNAC5 is a sialyltransferase that specifically mediates brain metastasis. Normally, it is involved in the synthesis of α-series gangliosides in the nervous tissues. The expression of ST6GALNAC5 is restricted to the brain and may have implications in neurodegenerative diseases [74]. Nevertheless, it has been shown that the expression of ST6GALNAC5 in CN34-BrM2 and MDA231-BrM2 cell lines of breast cancer enhances their adhesion to brain endothelial cells and their passage through the BBB [71]. Conversely, in an in vitro model, the expression of ST6GALNAC5 decreased the interaction of MDA-MB-231 breast cancer cells and the BBB [75]. Further exploration revealed that ST6GALNAC5 increased the expression of GD1α ganglioside at the cell surface of invading breast cancer cells, limiting interaction with the BBB [75]. Furthermore, studies with patient-derived xenografts in TNBC showed that ST6GALNAC5 expression was not related to metastasis location [76]. Nonetheless, in colorectal cancer, ST6GALNAC5 expression was significantly associated with brain metastatic susceptibility in patient-derived samples [77]. Further studies are needed to clarify the role of ST6GALNAC5 as it is possible that the expression of this gene is situational and depends on the stage of the brain metastatic process.

In A375Br melanoma cells, the expression of PLEKHA5 was associated with early development of brain metastases [78]. Inhibition of PLEKHA5 caused a decrease of passage across the BBB. PLEKHA proteins play a role in brain development and are thought to function mostly as adaptor proteins because they lack catalytic subunits [79]. PLEKHA5 clearly has phosphoinositide-binding specificity implicating a role of phosphoinositide 3-kinases (PI3K) in its membrane recruitment. Moreover, phosphoinositide-binding specificities suggest that PLEKHA5 can be a key mediator of cellular signaling, cytoskeleton rearrangement, membrane protein targeting, and vesicular trafficking [78]. As such, the established role of PLEKHA5 in cell migration might facilitate BBB transmigration and promote it as a therapeutic target for brain metastasis.

Expression of chemokine receptors like CXCR4 and CCR7 is significantly associated with metastasis and mortality in breast cancer. In fact, the expression of CXCR4 through NF-κB promoted breast cancer cell migration and metastasis. The interaction of CXCR4 with its ligand, SDF-1α, in MDA-MB-231 and DU4475 breast cancer cells induced activation of the PI3K/AKT signaling pathway and the phosphorylation of focal adhesion kinase (FAK) and forkhead in rhabdomyosarcoma-like 1 (FKHRL1) [80]. This indicates that the activation of the PI3K/AKT signaling pathway by CXCR4/SDF1 is implicated in breast cancer cell migration through the BBB and targeting it may offer therapeutic potential.

In addition to its remodeling capacity (discussed before), heparanase mediates complex interactions between tumor cells and the host tissue environment. Cell surface expression of heparanase elicits a firm cell adhesion, reflecting an involvement in cell–ECM interaction. Heparanase promotes Akt signaling and stimulates PI3K- and p38-dependent endothelial cell migration and invasion [51]. Silencing of the heparanase gene inhibits tumor metastasis and angiogenesis in experimental models.

The visualization of how myosin light chain kinase (MLCK) is modulated during tumor cell intravasation helped elucidate the signaling interaction between metastatic breast cancer cells and endothelial cells lining the BBB [81]. Endothelial MLCK was found to be activated at the adhesion site of MDA-MB-231 cancer cells, leading to regional myosin contraction and, eventually, intravasation of the tumor cell [81]. Blocking endothelial myosin contraction cut the transcellular route of invasion [81], implicating an important role for MLCK in tumor intravasation and highlighting yet another therapeutic target.

The contribution of immune cells to the metastatic process has always been questioned. The co-culturing of T lymphocytes with breast cancer cells increased the metastatic potential of the cancer cells across the BBB. Guanylate-Binding Protein 1 (GBP1) was found to be induced by T cells upon interaction with metastatic MDA-MB-231-BM breast cancer cells [82]. In humans, GBP1 was upregulated in patients with breast cancer brain metastases. Silencing the gene in vitro reduced the invasive ability of breast cancer cells through the BBB [82]. This provides another potential target for intervention to prevent brain metastasis formation.

The brain microvasculature has been implicated as an important player for tumor cell attachment and growth. In SCLC, the endothelial cells in the brain induced NCI-H446 cells to secrete annexin A1 [83]. The release of annexin A1 was found to promote SCLC brain metastasis formation by enhancing cellular adhesion on the BBB [83]. Knockdown of annexin A1 in SCLC cells prevented transcellular migration and metastasis, highlighting the role of annexin A1 as a BBB-scaffolding protein in brain metastasis.

αB-crystallin is a small heat shock protein that in humans is encoded by the CRYAB gene. It prevents protein aggregation, inhibits apoptosis, and contributes to the intracellular architecture. In women with TNBC, αB-crystallin was found to be a regulator of cancer metastasis to the brain. The expression of αB-crystallin enhanced the adhesion of GILM2 and MDA-MB-231 breast cancer cells to the endothelial cells of the BBB and facilitated trans-endothelial migration [84]. Silencing of the CRYAB gene caused binding inhibition with endothelial cells. Further exploration revealed that adhesion occurs via α3β1 integrin-dependent mechanism [84]. As such, CRYAB and αB-crystallin represent an important therapeutic target in TNBC.

Syndecans (Sdcs) are cell surface proteins that function as co-receptors for growth factors, matrix proteins, cytokines, and chemokines. In their role, Sdcs are capable of mediating cell–cell and cell–matrix adhesion. In breast cancer brain metastasis, overexpression of syndecan-1 (Sdc1) was found to increase adhesion to the perivascular regions of the brain and lead to eventual trans-endothelial migration [85]. Silencing expression of Sdc1 in MDA-MB-231 and 4T1 breast cancer cells significantly reduced metastasis to the brain [85]. In addition, loss of Sdc1 also led to changes in breast cancer cell-secreted cytokines/chemokines, which may modulate the BBB.

Fucosylated CD15 and sialyated CD15s are carbohydrate epitopes on cancer cells that aid in adhesion. Overexpression of CD15 or CD15s epitopes in SEBTA-001, NCI-H1299, and COR-L105 lung cancer cell lines led to a significant increase in adhesion of cancer cells to the E-selectin on cerebral endothelial cells [86, 87]. This resulted in the disruption of cerebral endothelial cell monolayers. Knockdown of fucosyltransferases 4 (FUT4) and 7 (FUT7) that code for CD15 and CD15s expression, respectively, in metastatic cancer cells prevented disruption of an in vitro BBB model [86]. In addition, non-metastatic COR-L105 lung cancer cells turn metastatic when forced to overexpress either FUT4 or FUT7 [86]. This hints at the clinical significance of using cancer epitopes as biomarkers for metastasis.

GENE SUPPRESSION OF TRANSMIGRATION

Genes that suppress metastatic transmigration across the BBB are attractive targets for prognostic and therapeutic biomarkers. Comprehension of the biological and genetic mechanisms that govern transmigration suppression can provide insight into the processes that promote metastatic spread.

MFSD2a is a nutritionally regulated gene with important roles in mammalian tissue and organ growth, lipid metabolism and cognitive and motor functions. In the brain, MFSD2a selectively transports docosahexaenoic acid (DHA), an omega-3 fatty acid, across the BBB. Genetic deletion of MFSD2a protein in mice lead to impaired DHA transport and reduced levels of vital lipid metabolites. Using patient-derived xenograft models, the inhibition of MFSD2a increased the propensity of brain metastasis by reducing DHA transport and altering lipid metabolism in the metastatic microenvironment [88]. Deficiencies in DHA transport and metabolism are linked to cognitive deficits and ataxia [89, 90]. This suggests that restoring the DHA levels in the brain tumor microenvironment may be a novel therapeutic strategy to block metastatic cell growth and survival. Furthermore, MFSD2a is a key regulator of BBB function. Genetic ablation of MFSD2a results in a leaky BBB from embryonic stages through to adulthood, but the normal patterning of vascular networks is maintained [91]. Electron microscopy examination reveals a dramatic increase in CNS-endothelial-cell vesicular transcytosis in MFSD2a knockout mice, without obvious tight-junction defects [91]. Zebrafish models further confirmed that MFSD2a inhibits transcytosis, as MFSD2a mutants displayed elevated BBB permeability due to increased transcytosis [92]. Moreover, MFSD2a endothelial expression is regulated by pericytes that facilitate BBB integrity, whereby lipids transported by MFSD2a establish a unique environment that inhibits caveolae vesicle formation in CNS endothelial cells to suppress transcytosis and preserve BBB integrity [93].

MicroRNA profiling analyses revealed that miR-509 is significantly decreased in metastatic MDA-MB-231BrM and CN34BrM breast cancer cells that can cross the BBB [94]. Upon further exploration, miR-509 was capable of modulating the two genes, RhoC and TNF-α that control the invasion of cancer cells and the transendothelial cell migration across the BBB, respectively. Levels of MMP9 induced by TNF-α and RhoC were significantly associated with brain metastasis-free survival in breast cancer patients [94].

Claudin-5 (CLDN5) also plays a role in regulating the permeability of the BBB in lung cancer brain metastasis. Silencing of CLDN5 increased the paracellular permeability of A549 cancer cells and increased the invasion of lung adenocarcinoma [95]. The role that CLDN5 plays in tightening endothelial cellular adhesion and enhancing the function of the tight junctions reduces the formation of lung cancer brain metastasis. In TNBC, angiopoietin-2 was reported to breach CLDN5 junction protein structures and increase metastatic invasion of MDA-MB-BrM2 and 4T1 cells [22]. As such, therapeutic strategies that aim at protecting CLDN5 and targeting angiopoietin-2 should be favored.

Adenosine A2A receptors are widely distributed in the brain [96]. The activation of adenosine A2A receptors inhibit the interaction of SDF-1 and CXCR4, leading to a decrease in the brain metastatic capacity [97]. Adenosine A2A receptor activation inhibited the proliferation and viability of lung cancer PC-9 cells and thus suppressed brain metastasis in vitro [98]. Treatment using A2A receptor agonist and CXCR4 antagonist protected nude mice from brain metastasis in vivo [98]. Further investigation revealed that the activation of A2A receptor increased the expression levels of CLDN5, occludin and ZO-1 dramatically [98]. Conversely, when A2A receptor was blocked by antagonist, these junction proteins were downregulated [98]. This indicates that the activation of the A2A receptor protects the integrity and function of BBB against metastatic transmigration.

Tumor-derived extracellular vesicles can breach the intact BBB. By pretreating nude mice with extracellular vesicles derived from parental MDA-MB-231 tumor cells and then administering an intracardiac injection of tumor cells, the size and incidence of brain metastases increased significantly when compared to non-pretreated mice [99]. Gene knockout studies confirmed that suppressing the brain endothelial expression of Ras-related protein 7 (Rab7) increases the efficiency of the cancer cell transport [99]. Furthermore, high spatiotemporal resolution microscopy demonstrated that the endothelial recycling endocytic pathway is involved in this transcellular transport across the BBB [99].

Pericytes seem to possess protective qualities against metastatic cells. By studying the interaction between pericytes and metastatic A549 and KNS-62 lung cancer cells, it was observed that the expression of tight junction protein was preserved upon co-culture with pericytes [100]. In addition, resistance across the trans-endothelial pathway was left intact. The conditioned medium of pericytes with metastatic cancer cells suppressed the proliferation of the cancer cells significantly. This ultimately revealed that pericytes are major regulators of the resistance of the BBB to lung cancer metastasis to the brain [100]. Further research identified the sub-populations of pericytes that are involved in modulating metastasis in the breast cancer setting [16]. An increase of desmin+ pericytes and a decrease of CD13⁺ pericytes increased permeability across brain endothelial barriers, which led to the passage of large protein components, osmosis and brain edema [16]. As such, future preventive therapies targeting desmin⁺ pericytes can be promising in the face of scarce effective therapeutic options.

CONCLUDING REMARKS

The clinical threshold of danger in cancer is metastatic spread. Brain metastasis, per se, portends poor outcomes and low quality of life. Favorable prognosis accounts for diagnosis made prior to the growth of secondary cancerous sites. As such, understanding the various routes of transmigration across the BBB is essential for the development of new targeted therapies and for preventive measures that can target metastatic cancer cells before reaching the brain. Specific mediators of transmigration present an opportunity for novel strategies in facing the challenge raised by brain metastasis. Furthermore, BBB destruction is not limited to the formation of brain metastases. It is also involved in other diseases like stroke, trauma, Alzheimer’s, and primary brain tumors. As such, novel therapeutics focusing on the BBB can have the potential to contribute positively to other diseases as well. Nevertheless, there is a need to acknowledge the limitation of extrapolating the BBB migration strategy from current literature. Although in some cases, the mechanism of migration is clearly identified for some genes, the mechanism is not clear yet for many others. Future research must address this issue by attempting to decipher the exact mechanism of transmigration promoted by different genetic drivers.

It is also becoming clear that metastatic cells heterogeneously affect the integrity of the BBB [1, 3]. Specific subpopulations of pericytes are associated with differential permeability, and correspondingly, only drugs designed to fully penetrate the BBB will be therapeutically efficacious [101]. Moreover, questions have been raised on the role of chemotherapeutic agents in promoting brain metastasis. It has been shown that patients with HER2-positive breast cancer who received taxanes had worse prognosis and higher chance of developing brain metastasis [102, 103]. This raises concern as to whether other chemotherapeutic agents are also involved in promoting metastatic behavior and penetration of the BBB. Furthermore, the role of ionizing radiation in promoting BBB invasion has come into question in light of proteomic analyses showing that metabolic reprograming induced by radiation helps metastatic cancer cells accommodate to the brain [104].

Altogether, these accumulating studies infer that different components of the brain-tumor microenvironment can fundamentally regulate the properties of the BBB, demanding an integrated investigation of these genetic, proteomic and cellular mediators in the future to improve the efficacy of brain-targeted therapies. Moreover, diversifying cell lines and developing better experimental models is necessary to fully elucidate the mechanisms of transmigration of metastatic cells across the BBB. In addition, the development of diagnostic assays that can screen for the expression of these genes/proteins will help in the early detection and prevention efforts in the fight against cancer (see Outstanding Questions).

Outstanding Questions.

• What other genes are regulating invasion across the blood-brain barrier (BBB)? Taking into consideration tumor heterogeneity, are these genetic mutations common across all invading tumor cells to the brain?

• Knowing that the brain is usually the second most common site of metastasis, after lungs and bone, does the passage of cancer cells through tight tissue cause cellular stress and subsequent genetic alterations that can increase the cancer cell’s potential to invade the BBB?

• Knowing that patients with recurring cancer who had received chemotherapy and radiotherapy will have a higher chance of brain metastasis, are such therapies inducing genetic mutations that drive metastatic invasion across the BBB?

• If gene therapy is to be adopted to target metastatic cancer cells, what will be the length and extent of such therapeutics? Will it be convenient to translate such therapies to the clinic? Would such therapies be cost effective? Will gene therapy in combination with other therapeutic regimens offer better efficacy?

Highlights.

• The blood-brain barrier (BBB) is restrictive to the extent that not even ions can pass freely from one side to another.

• Most drugs are incapable of permeating through the BBB as transport is tightly regulated by efflux transporters.

• Brain metastatic cells are capable of crossing the BBB through the paracellular route or the transcellular pathway.

• Genetic alterations in cancer cells drive their metastatic capabilities and their ability to transmigrate through the BBB.

Glossary

ADAM8

A gene that encodes a membrane-anchored protein implicated in cell-cell and cell-matrix interactions

Angiopoietin-2

A biomarker in cancer; involved in assembling and disassembling the endothelial lining of blood vessels

Annexin A1

A calcium-dependent phospholipid-binding protein located on the cytosolic face of the plasma membrane

Brain parenchyma

The functional tissue in the brain that is composed of neurons and glial cells

Cellular diapedesis

The movement of cells out of the circulatory system and towards the site of tissue damage, infection, or tumor

Efflux pumps

Transport proteins involved in the extrusion of substrates (including from within cells into the external environment

ERK1/2

Proteins that regulate meiosis, mitosis, and post-mitotic functions in differentiated cells. Disruption of the ERK pathway is common in cancers

FAK

A protein kinase involved in cellular adhesion and spreading processes

FKHRL1

Transcription factor that plays important roles in regulating the expression of genes involved in cell growth, proliferation, differentiation, and longevity

Gangliosides

Components of the cell plasma membrane that modulate cell signal transduction events. They can serve as specific determinants in cellular recognition and cell-to-cell communication

GBP1

A protein that is induced by interferons and characterized by its ability to specifically bind guanine nucleotides

MAPK

A protein kinase that regulates cellular proliferation, gene expression, differentiation, mitosis, cell survival, and apoptosis

miRNAs

A small non-coding RNA molecule that functions in RNA silencing and post-transcriptional regulation of gene expression

NF-κB

A protein complex that controls transcription of DNA, cytokine production, and cell survival

PI3K

An enzyme involved in cellular functions such as cell growth, proliferation, differentiation, motility, survival and intracellular trafficking

Proteases

An enzyme which breaks down proteins and peptides

Seprase

An integral membrane serine peptidase that promotes cell invasiveness towards the ECM and support tumor growth and proliferation

Tight junctions

A cell-cell junction found barriers that are impermeable to the majority of soluble molecules between the two sides of the barrier. Major components include claudin and occludin proteins

Tumor heterogeneity

The observation that different tumor cells within the same tumor can show distinct morphological and phenotypic profiles

uPA

An enzyme that is made in the kidney and found in the urine. It is used to dissolve blood clots or to prevent them from forming

VEGF

A potent angiogenic factor

ZO-1

A scaffold protein that cross-links and anchors Tight Junction strand proteins