CHAPTER 105 Barriers to Delivery of Therapeutics to Brain Tumors

Malignant gliomas are the most common type of primary brain tumor in adults. According to the Central Brain Tumor Registry of the United States, it is estimated that approximately 20,000 new cases develop each year.¹ Treatment of malignant glioma represents one of the most formidable challenges in oncology. Despite treatment with surgery, radiation therapy, and chemotherapy, the prognosis remains poor, particularly for patients with glioblastoma multiforme (GBM), who have a median survival of 12 to 15 months.² After recurrence, rapid tumor progression results in a median progression-free survival and overall survival of 9 and 25 weeks, respectively. The invasiveness of these tumors eliminates the possibility of curative surgical resection. Even at initial evaluation, infiltration by tumor cells extends at least 2 cm away from the radiographic contrast–enhancing mass.³^,⁴ One of the main difficulties associated with treating malignant gliomas is the presence of the blood-brain barrier (BBB) and the partially functional blood-tumor barrier, which serves to prevent effective delivery of potentially active chemotherapeutic compounds. There is vast interest in and consequently multiple research efforts focused on the design of new approaches that will improve drug delivery to brain tumor cells with limited systemic toxicity. This chapter summarizes the methods that have been developed for overcoming the ability of brain tumor cells to remain safe from therapy behind multiple physical and physiologic barriers.

Blood-Brain Barrier

The BBB consists of a layer of endothelial cells that line the blood vasculature throughout the brain.⁵ These endothelial cells are held together by tight junctions produced in response to signals from astrocytes. Certain molecules, particularly large water-soluble molecules, are prohibited from diffusing through the gaps between these cells. Small, lipophilic molecules can, however, diffuse across the membranes, but they too are limited in distribution by diffusion. The BBB limits central nervous system (CNS) delivery of many common chemotherapeutic agents, and its permeability to most molecules can be predicted on the basis of each agent’s octanol-water partition coefficient and the unidirectional transfer coefficient (K_in).⁶^–⁸ The octanol-water distribution coefficient measures solute lipophilicity, whereas K_in is a quantitative measure of the ability of a drug to pass from plasma into the brain. K_in is largely determined by lipid solubility because agents must first dissolve in the lipid membranes of the BBB to cross by lipid-mediated diffusion.⁹ K_in can be plotted against the partition coefficient by using 20 reference permeability markers that bind minimally to plasma proteins and cross the BBB by passive diffusion (Fig. 105-1).¹⁰ Many of the chemotherapeutic agents fall below the line predicted for BBB passive diffusion.

FIGURE 105-1 Relationship between blood-brain barrier (BBB) permeability and the octanol-water partition coefficient of chemotherapeutic agents. The solid line is the least-squares fit to the data for agents that are not actively taken up by the brain or pumped out by the BBB. FU, 5-fluorouracil.

(From Smith QR. A review of blood-brain barrier transport techniques. Methods Mol Med. 2003;89:193-208. With kind permission of Springer Science+Business Media).

Active or facilitated transport is used for substances with low partition coefficients. This type of transport is dependent on ion channels, specific transporters, energy-dependent pumps, and receptor-mediated endocytosis. Glucose, amino acids, and small intermediate metabolites are carried into the brain via facilitated transport, whereas larger molecules such as insulin and transferrin are carried across the endothelial layer via receptor-mediated endocytosis.¹¹

Certain factors contribute to poor uptake of chemotherapeutic agents across the BBB, including plasma protein binding, solute molecular weight, and active efflux transport. Most chemotherapeutic agents are bound to plasma proteins, which reduces the free fraction of the drug in plasma that is available to cross the BBB. The molecular weight of most of these agents exceeds 400 daltons, significantly larger than what would normally passively leak through the BBB.¹²^,¹³ Another component of the BBB that makes it difficult for chemotherapeutic agents to function is its expression of high levels of drug efflux pumps, such as P-glycoprotein.¹⁴ These pumps actively remove chemotherapeutic drugs from the brain. A number of newer chemotherapeutic agents, such as gemcitabine, docetaxel, pemetrexed, irinotecan, and topotecan, which show promising antitumor activity against systemic tumors, have limited delivery across the BBB because of active efflux transport and plasma protein binding.¹⁵ The thymidine kinase inhibitor imatinib binds heavily to plasma proteins and is a substrate for active efflux pumps, and this is also suspected to be the case for erlotinib and gefinitib.¹⁶ The organic anion transporters and glutathione-dependent multidrug resistance–associated proteins also contribute to the efflux of organic anions from the brain and cerebrospinal fluid (CSF).¹⁷ There has been research focusing on blocking these types of transporters to improve therapeutic drug delivery across the BBB.

Blood-Tumor Barrier

Although the endothelial tight junction component of the BBB can be compromised significantly in the enhancing part of a glioma and hence result in the hallmark of contrast enhancement on computed tomography or magnetic resonance imaging, other remaining physiologic barriers and new physical barriers can restrict delivery of drug to brain tumors. When compared with the normal, ordered vasculature of healthy tissues, blood vessels in tumors are often highly abnormal, such as distended capillaries with leaky walls and sluggish flow, and thereby lead to inconsistent drug delivery. “Leakage” from the tumor vasculature results in accumulation of interstitial fluid that subsequently increases intratumoral interstitial pressure, thus limiting the diffusion of drugs into brain tumors.¹⁸ It has been shown that tumors can consist of areas of normal interstitial pressure, in which the pressure is 1 to 2 mm Hg, interposed with peritumoral areas in which interstitial fluid pressure can be 50 mm Hg or greater.¹⁹ Recent studies have shown that targeting vascular endothelial growth factor with the drug bevacizumab may decrease interstitial pressure to allow greater entry of drug into the tumor.²⁰ Highly irregular blood flow results in localized hypoxia, which subsequently leads to resistance of tumor to certain anticancer agents and to radiotherapy as well. Increased bioreductive enzyme expression is an adaptive strategy for solid tumors to detoxify anticancer drugs, and the hypoxic environment further contributes to the increased bioreductive activity of tumors.²¹

Blood–Cerebrospinal Fluid Barrier

The specialized tight junctions of epithelial cells in the choroid plexus (the blood-CSF barrier) that separate CSF from blood are also responsible for barrier function.²² The tight junctions of the blood-CSF barrier, in comparison to the BBB, are between the epithelial cells of the choroid plexus rather than at the endothelial layer. By sealing neighboring epithelial choroidal cells continuously together, the tight junctions strongly restrict the paracellular movement of solutes and thus considerably limit the diffusion of polar drugs into the CNS via the choroid plexus. The fenestrated choroid plexus capillaries provide little resistance to the movement of small molecules. However, the blood-CSF barrier does not contribute significantly to entry of drug into the brain parenchyma because the surface area of the blood-CSF barrier is approximately 1000-fold less than the surface area of the BBB.²³ Similar to the BBB, however, the blood-CSF barrier also contains drug-metabolizing enzymes located within the choroid plexus and in some circumventricular organs, and these enzymes primarily serve a protective role against exogenously administered molecules.²⁴

Drug Modifications for Enhanced Drug Delivery to Brain Tumors

Lipophilic Analogues

As stated previously, the BBB allows the diffusion of small, nonionic, lipid-soluble molecules, whereas larger, more water-soluble ionic molecules do not readily cross (Table 105-1). There are endogenous transporters that allow some large and polar molecules to be transported into the brain. To improve drug delivery to brain tumors, certain strategies have been formulated that involve passive drug uptake into brain by means of lipophilic analogues. Carmustine is an alkylating agent used to treat brain tumors and other well-known malignancies. Multiple carmustine analogues have been studied in clinical trials and have demonstrated decreased alkylating activity and increased dose-limiting toxicity when compared with carmustine.²⁵ This is probably due to factors affecting drug-receptor interactions or increased binding to plasma proteins resulting in lower drug concentrations available for diffusion into the brain. Moreover, lipophilic analogues are less soluble in brain interstitial fluid, which limits their activity against tumor cells.²⁶

TABLE 105-1 Mechanisms of Drug Delivery to Brain Tumors

	ADVANTAGES	DISADVANTAGES
Drug Modifications
Lipophilic analogues	Allows passage of large, polar molecules	Limited activity against tumor cells
Lipophilic prodrugs	Better penetration than with parent drug	Require chemical transformation to achieve active form
ADEPT/GDEPT	Enhances prodrug activity More specific targeting	Antigen/antibody counteraction Difficulty achieving selective gene delivery to a sufficient number of tumor cells
Carrier-mediated drug transport	Facilitates endogenous transport	Requires a specific carrier protein
Receptor/vector-mediated drug targeting	Improves brain uptake by coupling drugs to vectors	Imposed limitation by the immune system Causes transient inflammation on injection
Barrier Disruption Strategies
Osmotic BBBD	Increases drug concentrations delivered to tumor cells	Potential for neurotoxicity
Biochemical BBBD	More precise time window for delivery of drug to the brain	Potential for neurotoxicity
Ultrasound-mediated BBBD	Localized and reversible image-guided disruption of the BBB	Requires craniotomy
Direct Delivery Methods
Catheter with pump systems	Direct delivery of drugs	Infection Catheter obstruction Inadequate drug distribution
CED	Achieves both local and regional drug delivery Bypasses the BBB Allows the use of tumor-targeted drugs that may have systemic toxicity	Requires implanting catheters into the brain Catheters to date are not designed for CED and have reliability problems Current approaches have not allowed imaging of distribution
Implanted polymers	Continuous drug delivery	Local neurotoxicity Poor wound healing Limited drug penetration
Other Approaches for Delivery
Intraventricular/intrathecal	Effective for leptomeningeal spread	Poor delivery to brain parenchyma
Intravascular	Improved efficacy when used with targeting strategies	Limited efficacy as stand-alone therapy
Liposomal drug encapsulation	Can be modified to better target tumor cells	Highly unstable
Nanoparticulate systems	Natural polymers Controlled drug delivery	Transport of drugs across the BBB depends on outer coating
Intranasal applications	Noninvasive Rapid delivery Avoids hepatic drug metabolism	Further investigation is still required
Magnetic microspheres	Enhanced tumor targeting by retaining cationic particles	Requires external magnetic field
Boron neutron capture therapy	Low systemic toxicity Selective localization in tumors	Requires further design to enhance tumor cell killing

ADEPT, antibody-directed enzyme prodrug therapy; BBB, blood-brain barrier; BBBD, blood-brain barrier disruption; CED, convection-enhanced delivery; GDEPT, gene-directed enzyme prodrug therapy.

Lipophilic Prodrugs

As an alternative to analogues, prodrugs require chemical or biochemical transformation to achieve the active form within the body.²⁷^,²⁸ Prodrugs are designed to overcome the pharmaceutical or pharmokinetic limitations, or both, of the parent molecule to better penetrate the BBB. Lipophilic ester prodrugs of the anticancer agent chlorambucil have been developed to increase efficacy in the treatment of brain tumors.²⁹ After equimolar doses of chlorambucil and chlorambucil–tertiary butyl ester, brain delivery of the ester was 35-fold greater than that of chlorambucil. However, despite an enhanced CNS delivery ratio, none of the prodrugs demonstrated anticancer activity superior to the equimolar administration of chlorambucil against a brain-sequestered carcinosarcoma in the rat.³⁰

Antibody- and Gene-Directed Enzyme Prodrug Therapy

Further strategies have been used to enhance prodrug activity at the site of action with minimal effect on the rest of the body. If the prodrug can be selectively cleaved to the active drug by specific enzymes near the site of action, function would be enhanced. Prodrug-activating enzymes were initially targeted to tumor cells by using antibodies and more specifically genes. In antibody-directed enzyme prodrug therapy, enzymes that activate prodrugs are directed to human tumor xenografts by conjugating them to tumor-selective monoclonal antibodies.²³ Once the enzyme is conjugated with an antitumor antibody, it is delivered to the tumor by intravenous (IV) infusion. When the conjugate is cleared from blood, a prodrug is then delivered. The prodrug is now able to be activated by the tumor-associated enzyme. Martin and associates were able to use this strategy in phase I clinical trial in patients with nonresectable metastatic colon carcinoma.³¹ Unfortunately, there have been limitations to this strategy in that some tumor-specific antigens inactivate the antibody-enzyme complexes delivered. These limitations are what led to the development of gene-directed enzyme prodrug therapy (GDEPT).

GDEPT was the first type of gene therapy used for treating human brain tumors. In GDEPT, an inactive prodrug can be activated to release a cytotoxic drug by an enzyme that has been delivered to the tumor for expression. Specific enzymes used include thymidine kinase (TK), nitroreductase, and cytochrome P-450. The most studied approach uses the TK gene inserted into herpes simplex virus, followed by combination treatment with the prodrug ganciclovir. In one trial using this strategy in patients with malignant glioma, 12 of 15 patients showed moderate response to the treatment, but no complete responses were observed.³² As with other approaches to gene therapy, this method is limited by the difficulty of achieving selective gene delivery to a sufficient number of tumor cells.

Chemical Drug Delivery System and Carrier-Mediated Drug Transport

By using specific properties of the BBB, this system locks drugs in the brain on arrival and prevents them from recrossing the BBB.³³ Linking an active drug molecule to a bioremovable lipophilic targeting moiety (targeter) creates a complex that can then be oxidized to form a charge that has the effect of capturing the ionized drug–targeter inside the brain. Subsequently, slow release of the drug from the targeter results in sustained and brain-specific release of the free active drug. This system has been studied consistently with the drug lomustine.³⁴

The carrier-mediated drug delivery approach uses the facilitative endogenous transport systems that are present in brain endothelial cells. Specific transport systems that exist for the brain include glucose, amino acids, choline, vitamins, low-density lipoprotein, and nucleosides.³⁵^–⁴⁰ Glucose and the large neutral amino acids have high transport capacity, which makes them the most used agents for delivery of drug to the brain.⁴¹ Several amino acid mimetic agents such as L-dopa, α-methyldopa, and baclofen are drugs in clinical use that are readily taken up into the brain by these transport systems.

Receptor/Vector-Mediated Drug Targeting

This approach makes use of the endogenous BBB transport system by aiming at improving brain uptake by coupling nontransportable therapeutic molecules to a drug transport vector.²³ These vectors may include endogenous peptides such as insulin or transferrin, modified proteins, or antireceptor-specific monoclonal antibodies. A significant increase in radioactivity in rat brain tumor versus normal brain was achieved after IV administration of the radiolabeled epidermal growth factor peptide conjugated to the antitransferrin monoclonal antibody transport vector.⁴²

Barrier Disruption Strategies for Enhancing Drug Delivery

Although advances in chemical modification of drugs to make them lipophilic and the use of carriers to circumvent the BBB have been explored, transient BBB disruption (BBBD) became a favorable option in the recent past. Techniques that transiently disrupt the BBB were investigated, including techniques that create a paracellular route of transport through the endothelium by opening tight junctions and those that create a transcellular pathway through the endothelium. Several endogenous compounds, including neurotransmitters, hormones, and inflammatory mediators, can readily open the tight junctions of the BBB.⁴³ BBBD has also been seen in physiologic states such as hypertension, hypoxia, and ischemia. Hypertonic substances, including mannitol and biologically active agents such as bradykinin and angiotensin peptides, have also been shown to disrupt the BBB.

Osmotic Blood-Brain Barrier Disruption

Transient osmotic disruption of the BBB and blood-CSF and blood-tumor barriers can be achieved within a localized vascular distribution by intra-arterial (IA) infusion of a hyperosmotic agent, usually mannitol.⁴⁴ By injecting hypertonic solution, rapid diffusion of fluid out of the cells takes place and causes shrinkage of endothelial cells and subsequent opening of the tight junctions of the BBB for several hours (Fig. 105-2).⁴⁵ This infusion of hypertonic solution is usually followed by the IA administration of chemotherapy. Animal studies suggest that this method is able to increase concentrations of chemotherapeutic agents in the brain parenchyma up to 90-fold.⁴⁶ Delivery of methotrexate to the CNS is enhanced 4- to 7-fold when administered IA after osmotic BBBD as compared with IA administration without BBBD.⁴⁷ The ability of BBBD to improve the therapeutic efficacy of chemotherapy has been variable in humans. In chemoresponsive tumors, BBBD-associated chemotherapy compared favorably with conventional chemotherapy.⁴⁸^–⁵⁰ In primary CNS lymphoma phase II studies, a significant difference in survival was found when comparing patients treated with BBBD chemotherapy with or without prior whole-brain radiotherapy.⁵¹ A concern with the use of BBBD is the potential for neurotoxicity from the high concentrations of chemotherapy delivered to the normal brain. Chemotherapeutic agents such as doxorubicin, cisplatin, and taxanes cause neurotoxicity with BBBD, even though they are well tolerated systemically.⁵² The efficacy of BBBD-mediated chemotherapy for malignant gliomas is still being explored.