The treatment protocols for glioblastoma multiforme (GBM) involve a combination of surgery, radiotherapy and adjuvant chemotherapy. Despite this multimodal approach, the prognosis of patients with GBM remains poor and there is an urgent need to develop novel strategies to improve quality of life and survival in this population. In an effort to improve outcomes, intra-arterial drug delivery has been used in many recent clinical trials; however, their results have been conflicting. The blood–brain barrier (BBB) is the major obstacle preventing adequate concentrations of chemotherapy agents being reached in tumor tissue, regardless of the method of delivering the drugs. Therapeutic failures have often been attributed to an inability of drugs to cross the BBB. However, during the last decade, a better understanding of BBB physiology along with the development of new technologies has led to innovative methods to circumvent this barrier. This paper focuses on strategies and techniques used to bypass the BBB already tested in clinical trials in humans and also those in their preclinical stage. We also discuss future therapeutic scenarios, including endovascular treatment combined with BBB disruption techniques, for patients with GBM.
Statistics from Altmetric.com
Glioblastoma multiforme (GBM), the most common malignant primary brain tumor, has an infiltrating nature that precludes complete surgical resection. According to the Central Brain Tumor Registry of the USA, GBM accounts for 45.6% of all malignant brain tumors and 15.4% of all primary tumors, with an annual incidence of 3.19 cases per 100 000 persons in the USA.1 ,2 The median age of patients at diagnosis of GBM is 64 years and the male/female ratio is 1.6.1 The current standard of care consists of a multimodal approach: maximal surgical resection, followed by adjuvant temozolomide and radiotherapy.3 The median survival is limited to 16–19 months, with approximately 25–30% of patients alive at 2 years after diagnosis.2
While the development of new surgical strategies and techniques combined with the refinement of already existing ones has led to better results regarding the extent of tumor resection,4–6 adjuvant therapeutic protocols, including radiotherapy and chemotherapy (eg, temozolomide and bevacizumab), have evolved and new ones have been designed, such as intra-arterial delivery of chemotherapeutic drugs, further improving the survival of patients with GBM.7–9
Despite evidence suggesting that genomic and molecular characteristics may influence tumor responses to drugs, GBM inevitably recurs and spreads in all patients. Bevacizumab has been tested in two phase III studies in which it was combined with standard chemotherapy and radiotherapy in patients with newly diagnosed GBM; there were improvements in progression-free survival but not overall survival.8 ,9
Regardless of the method of delivering drugs, exposure of tumor tissue to chemotherapeutics is hindered mainly by the blood–brain barrier (BBB), which constitutes a biomechanical and neurophysiological obstacle to drug penetration into the tumor. During the last decade various strategies have been developed to increase the permeability of the BBB. This paper focuses on the most recent and promising ones and on future therapeutic scenarios, including endovascular treatment combined with BBB disruption techniques, for patients with GBM.
The blood–brain barrier (BBB)
There are three barriers between the systemic circulation and the central nervous system (CNS) microenvironment: (1) the choroid plexus–ventricular cerebrospinal fluid (CSF) barrier; (2) the arachnoid–subarachnoid CSF barrier; and (3) the BBB. Of these three barriers, the last plays the greatest role in controlling and maintaining CNS interstitial fluid homeostasis.10
From an anatomical point of view, the BBB is a large site of exchange between the blood and the neuropil, with an average surface area of 100 cm2 per gram of brain tissue in mammals.11 Starting from the capillary lumen, the barrier consists of capillary endothelium with adjoining tight junctions lining a continuous non-fenestrated basal lamina. Thanks to the crucial function of two (among other) transmembranous proteins called occludins and claudins, tight junctions seal the paracellular space restricting even the flow of ions. The basal lamina is covered by a layer of pericytes, which take the place of the smooth muscle cells that are present in larger vessels. Astrocytic end-foot processes completely overlay the basal lamina forming a sheathed neurovascular structure. The contribution of astrocytes to the properties of the BBB is critical as these cells provide an additional mechanical barrier and they also modulate the phenomenon of endothelium polarization. For this reason, any disease of the astrocytes will inexorably lead to dramatic repercussions on the function of the BBB.11–13
The mechanisms of transport across the BBB can be schematically summarized as passive diffusion, passive diffusion via tight junctions, endocytosis, carrier-mediated transport, and carrier-mediated efflux (figure 1). Molecules bigger than 500 Daltons (Da) such as insulin (6 kDa) and albumin (6.6 kDa) require carrier-mediated transport to cross the BBB; conversely, molecules such as ethanol (46.1 Da), nicotine (162.2 Da), and caffeine (194.2 Da) are transported passively, and the efficiency of this transport is determined by the lipophilic characteristics and molecular sizes of the compounds.13
Malignant brain tumors, including GBM, alter the normal BBB by secreting humoral factors such as bradykinin, histamine, serotonin, and platelet-activating factor,14 thereby deregulating vascular permeability, producing hypoxia-induced changes, and modifying carrier-mediated efflux.13 The alterations in vascular permeability are related to diffuse upregulation and redistribution of aquaporin 4 on astrocyte membranes, decreased claudin and occludin expression within the tight junction complex, basal membrane disruption, and leaky neovascularization.13 These modifications impair the control of BBB permeability, resulting in a decreased time-concentration exposure of chemotherapy agents delivered across the BBB.15 Additionally, hypoxia-inducible factor 1,16 induced by hypoxia, and an acidic microenvironment associated with induced production of vascular endothelial growth factor, contribute directly and indirectly to hypoxia and finally to BBB permeability.17 Upregulation of carrier-mediated efflux proteins may result in poor penetration of the drug into the tumor parenchyma; in particular, these proteins may transport chemotherapy agents that have reached the parenchyma or endothelial cells back into the vascular lumen through the so-called reflux phenomenon.11
These last two alterations (hypoxia and the reflux phenomenon) are targeted in GBM therapy through the use of an anti-angiogenic antibody (bevacizumab) that binds to vascular endothelial growth factor A, preventing interaction of the factor with its receptor expressed by endothelial cells which would lead to angiogenesis,18 and the administration of chemotherapeutics including paclitaxel, docetaxel, imatinib, topotecan, and temozolomide.13
In addition to the transport systems of the BBB, intracellular and extracellular catabolic enzyme processes also contribute to the barrier within both the endothelium and astrocytes. One enzymatic process of particular importance is the O-6-methylguanine-DNA methyltransferase reaction whose increased activity generates resistance to temozolomide.19
Independently of whether an intravenous or intra-arterial route is used to infuse chemotherapy, the drugs must circumvent the anatomical barrier, avoid efflux transport, and evade enzymatic conversion to an inactive metabolite. We believe that manipulating the BBB can play a determinant role in the adjuvant treatment of malignant primary brain tumors by maximizing the intratumoral concentrations of chemotherapy agents.
Well-known strategies to circumvent the BBB
Table 1 schematically reports strategies to circumvent the BBB, with a focus on published clinical trials and references to the relevant literature collected through a search of PubMed and Scopus using the terms ‘brain tumors’ and ‘blood-brain’ with subsequent further research conducted using each relevant paper's reference list.
One strategy is the induction of an inflammatory response in the endothelial lining of the BBB produced by mediators of the inflammatory response itself such as leukotriene C4, interleukin 2, tumor necrosis factor α, and interferon γ, and vasoactive compounds including bradykinin, kinin and their analogs (such as RMP-7). These have been used to disrupt the BBB, theoretically allowing increased uptake of chemotherapeutic agents into the CNS. RMP-7, a bradykinin analog, has the particularly interesting property of eliciting such a response selectively in the blood–tumor barrier rather than the healthy BBB, reducing neurotoxicity related to diffusion of toxic agents in healthy brain tissue.20 ,21
Hyperosmotic disruption of the BBB is based on endothelial dehydration and subsequent interference with tight junctions, leading to increased permeability. Combinations of superselective intra-arterial cerebral infusion (SIACI; achieved by catheterization of the intracranial arteries directly supplying tumors) or intra-arterial infusion (achieved by catheterization of the internal carotid artery or vertebral artery) of mannitol with intra-arterial bevacizumab or intravenous etoposide and intra-arterial melphalan and carboplatin have been described in three different clinical trials.18 ,22–24 Mannitol produces non-selective disruption of the BBB (unlike the abovementioned RMP-7), explaining the need for endovascular procedures to obtain adequate control over the diffusion of chemotherapeutic drugs.
Convection-enhanced delivery involves the use of an intracerebral catheter placed in the tumor connected to a subcutaneous port access system to deliver chemotherapy in situ, thereby minimizing side effects related to the systemic diffusion of toxic agents. A spectrum of molecules has been tested so far, including the transforming growth factor (TGF) β2 inhibitor trabedersen25 (tumor growth and progression inhibition), oligodeoxynucleotides26 ,27 (modulation of the immune response against tumor tissue), and a recombinant chimeric cytotoxin composed of human interleukin 13 fused to a truncated mutated form of Pseudomonas aeruginosa exotoxin A (PE38QQR) also known as cintredekin besudotox25–27 (tumor cell uptake of bacterial mutated exotoxin thanks to interleukin-13 receptor expressed on GBM cell membranes).
Intranasal delivery of perillyl alcohol (POH) after surgical resection in patients with recurrent malignant glioma has shown promising results with 19% of patients remaining in clinical remission after 4 years of follow-up (retrospective study).28 ,29 POH bypasses the BBB by traveling along the trigeminal and olfactory pathways; however, its chemotherapeutic properties, which probably involve TGF-β and RAS intracellular cascade responses, have yet to be fully defined.
Efflux pump inhibitors
Efflux pump inhibitors such as the epithelial growth factor receptor kinase inhibitor erlotimib sabotage temozolomide-induced production of epidermal growth factor resulting in the transcription of the multidrug resistance gene (MDR-1) and expression of its protein P-glycoprotein, a 12-transmembrane ATP-dependent drug efflux pump. By preventing the expression of the latter, intracellular accumulation of chemotherapeutic agents such as temozolomide can be increased. Unfortunately at present, despite promising in vitro results, the efficacy of this strategy in vivo has been marginal.30
Vector-mediated delivery entails the use of a molecular (paclitaxel–angiopep-2 peptide–drug conjugate or GRN100531) or liposomal ‘Trojan horse’ (liposomal-encapsulated doxorubicin; Caelyx, Scheringh-Plough, Munich, Germany)32 to cross the BBB. Paclitaxel is a broad-spectrum antitumor microtubule-stabilizing agent. Angiopep-2 is a 19-amino acid sequence exhibiting a marked capacity for transcytosis across the BBB via a low-density lipoprotein receptor-related protein 1-mediated mechanism. The latter is normally expressed on the surface of the BBB and neoplastic cells including high-grade glioma.
Viral-mediated delivery, on the other hand, uses replication-incapacitated retrovirus produced by so-called vector-producing cells (eg, murine-derived fibroblasts). Once inside the malignant cell, the retrovirus introduces a herpes simplex thymidine kinase gene into the cell genome which, once expressed, makes the cell vulnerable to ganciclovir treatment. The major drawback of this technique is the poor control over retroviral tropism to tumor cells, thus necessitating in situ administration of the vector-producing cells (intraoperative field administration of the vector-producing cells in the resection cavity).33
Novel strategies to disrupt and/or circumvent the BBB
Novel strategies to disrupt or circumvent the BBB include the use of nanoparticle complexes containing chemotherapeutic agents, stem cell-mediated delivery, and thermomechanical disruption of the BBB.37 ,38 Table 2 summarizes the rationale of each technique, providing literature references.
Focused ultrasound (FUS) deserves particular mention. The use of ultrasonography in neurology and neurological surgery dates back to the early 1940s when the effects of high intensity FUS on biological tissues, with a focus on the CNS, started to be investigated.39 Ultrasound has been used for both diagnostic purposes (intraoperative localization of brain tumors after craniotomy) and as a therapeutic method (FUS to the substantia nigra and ansa lenticularis to treat patients with Parkinson's disease after a craniotomic approach).40 With the introduction of ultrasound phased arrays, allowing correction for cranial distortion, MRI thermometry and MRI-guided high-intensity FUS systems,41 craniotomy became unnecessary and deep beam targeting and far more reliable control have led to a resurgence in this field. The neurological applications of MRI-guided high-intensity FUS have now reached the clinical stage in the fields of movement disorders, neuropathic pain, embolic stroke, and primary and metastatic brain tumor ablation, whereas they are still in the preclinical stage for BBB disruption, epilepsy, and intracranial hemorrhage.42
Ultrasound waves induce tissue damage through thermoablation (heat caused by ultrasonic energy absorption) and cavitation (oscillating gas-filled cavities produced by components of the acoustic wave); these properties have been used to induce coagulative necrosis in primary and metastatic brain tumors.43 ,44
On the other hand, through a so-called ‘sonication’ process (closed-skull FUS exposure with concomitant administration of a microbubble saline solution), low power FUS provides safe and transient focal disruption of the BBB, allowing greater accumulation of MRI-detectable contrast agents45 and, more importantly, chemotherapeutic agents. Samiotaki et al46 described increased uptake of a neurotrophic, neuroprotective factor called neurturin in the caudate putamen and substantia nigra in murine brains. Using the same strategy, other investigators demonstrated safe delivery of a large spectrum of agents through the BBB after exposure to MRI-guided FUS and microbubbles in non-human models such as brain-penetrating nanoparticles,47 1,2-bis(2-chloroethyl)-I-nitrosurea, iron oxide magnetic particles conjugated to epirubicin,48 and even gene-based therapy agents.49 A study to evaluate the safety and feasibility of BBB disruption using transcranial MRI-guided FUS with intravenous ultrasound contrast agents in the treatment of brain tumors with doxorubicin is currently recruiting patients at the Sunnybrook Health Sciences Centre in Toronto, Canada, and a similar study is also being organized at our center.
The main concern regarding chemotherapy for brain tumors remains the selection of drugs that can be transported with a reliable and controllable mechanism into the tumor, bypassing the biomechanical and neurophysiological defenses that characterize a functional BBB and, at the same time, minimizing systemic and neurological drug toxicity. Despite enormous efforts, achieving effective drug concentrations in the brain remains a significant challenge in CNS drug development. Failures are often due to a poor understanding of the complexity of CNS pharmacokinetics,57 resulting in insufficient intratumoral drug concentrations and ultimately in non-significant modifications of the prognosis of patients with brain tumors. In order to maximize concentrations of chemotherapeutic drugs within brain tumors, selective and effective modulation of the BBB may be a viable strategy.
Combining selective alteration of the BBB with intra-arterial infusion of drugs could further improve intratumoral drug accumulation. Drugs administered by SIACI have a high intratumoral drug concentration in four specific situations: (1) when there is low regional blood flow; (2) when there is high regional drug extraction; (3) when the drug is rapidly cleared from the systemic circulation; and (4) more consistent tissue drug concentrations are achieved by pulsed or bolus dosing.58 Superselective catheterization of intracranial arterial vessels directly supplying the tumor mass would enable intratumoral administration of high doses of chemotherapeutic drugs, greatly reducing systemic and CNS adverse effects. Cerebral blood flow could be modulated, perhaps by using a balloon, thereby increasing regional drug extraction and achieving local concentrations that are higher than those that can be reached with intravenous infusion. Using a combination of BBB breakdown and SIACI could possibly improve survival significantly in patients with GBM.
There is considerable evidence to support this hypothesis. Fortin et al57 recently reported that SIACI produced a significant increase in peak plasma concentration that translated into a significant 3–5.5-fold increase in intratumor chemotherapy concentrations, as most GBM are highly vascularized. Furthermore, comparing the two routes of administration (intra-arterial and intravenous), it was found that the concentration of an anticancer drug increased by 20-fold when infused via an intra-arterial route compared with the intravenous route (9 ng vs 0.5 ng platinum/g tissue, respectively).57 The really exciting finding, however, was that combining the intra-arterial infusion with permeabilization of the BBB led to an 18-fold higher drug concentration (160 ng platinum/g tissue) than that achieved with intra-arterial infusion alone, and a 320-fold higher concentration than when intravenous infusion was used to deliver the drug.57 These observations suggest that BBB disruption might be a valuable adjunct to increase drug delivery and might further boost tumor tissue uptake of chemotherapy agents administered by SIACI.
Despite these very encouraging observations, further research is still required, particularly since improvements in survival are not yet clear.18 ,57 The ability to modify the BBB probably needs to be improved and MRI-FUS is likely to play a major role in this goal, given its potential to selectively alter target tissue and the immediate quantitative control with MRI.
Finally, an increasing understanding of the genetic and molecular differences among GBM justifies the efforts made to develop chemotherapy regimens tailored to individual patients.
SIACI may play a significant role in delivering chemotherapy, viral vectors, gene therapy agents, and emerging genomic drugs. However, the real key to improved efficacy will probably be the combination of SIACI and selective BBB disruption. Both these strategies, together with careful selection of patients and drugs, will play important roles in the management of GBM in the future. We agree with Peruzzi et al59 that in the future we will need to bring the endovascular neurosurgeon into the neuro-oncology treatment team.
Contributors SP conceived and designed the work. AC, FD, FC, and RD drafted and revised the manuscript and gave final approval of the version to be published. All authors read and approved the submitted manuscript.
Competing interests SP is educational consultant for Penumbra inc
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.