21st Century Science Initiative Grant: Researching Brain Cancer

Despite substantial progress in the treatment of cancer, it remains the second leading cause of death in the United States and is expected to become the leading cause of death within the next decade. The prevalence and gravity of the disorder, its refractoriness to current approaches, and the toxicity and disabilities caused by current therapies suggest that fundamentally new approaches must be developed. This is a proposal to develop such a new approach based upon the convergence of two technologies, nanotechnology/materials science and stem cell biology, and their application to the treatment of brain tumors. Brain tumors are among the most devastating cancers. Primary brain tumors affect approximately 20,000 patients each year in the United States with about 12,500 deaths annually. Despite significant advances in surgical techniques, delivery of radiation, and chemotherapy, glioblastoma multiforme (GBM) remains a dire diagnosis, with a median survival of just over 14 months. Nearly 90% of patients suffer from tumor recurrence adjacent to the site of initial tumor resection despite radiographic evidence of disease remission following initial therapy. In part, tumor recurrence occurs because these tumors are not encapsulated and are thus not surgically curable; at the time of presentation, tumor cells have already migrated beyond the tumor mass and infiltrated surrounding brain.

The cancer stem cell (CSC) theory posits that treatment failures reflect persistence of brain tumor stem cells (BTSCs) following therapy (1,2). CSCs are characterized by the property of self-renewal (ability to produce more CSCs), and the capacity to generate all the various cell types that comprise a tumor. In addition, CSCs are thought to share many properties with normal stem cells—relative quiescence, activation of DNA repair mechanisms, and expression of drug transporters—that allow them to survive standard chemotherapies. Thus, generation of effective therapies for GBM will require development of agents that inhibit BTSC proliferation and that can also overcome the physical limitations of drug delivery into the brain. One approach to developing new therapies against CSCs is to activate biological pathways that drive these stem-like cells to differentiate and stop dividing. Many signals important in normal neural stem cell (NSC) biology play a comparable role in BTSC behavior (1,2), suggesting that signals that cause NSCs to differentiate may have a similar role in BTSCs.

However the treatment of primary brain tumors is complicated by many factors including dispersion of malignant cells throughout brain parenchyma, resistance to conventional treatments, susceptibility of adjacent normal brain to adverse effects of therapy, and the limited capacity of brain tissue for repair. In addition, bulk flow of extracellular fluid from the tumor site makes it difficult to sustain concentrations of antitumor agents even when they are delivered locally or with viral vectors. Added complications include leaky capillaries and a disrupted blood-brain barrier, difficulty distinguishing between normal brain and tumor, and many negative sequellae of therapy.

How can these complicated issues be addressed at the same time that a brain tumor might be effectively treated? A novel approach may be to reconstruct the extracellular matrix (ECM – the complex material that surrounds cells) so that it provides signals to BTSC and other proliferative cells to stop dividing and differentiate. At first glance reconstruction of the ECM without disruption of organ structure and integrity and without damaging normal cells in the brain might seem technically impossible. However we have recently developed peptide amphiphiles (PAs) that can be injected as liquids into neural tissues without damaging them and that self-assemble to form extracellular nanofibers with dimensions approximately the same as intermediate filaments (6). These PAs can be designed to present any desired bioactive sequence to surrounding cells at near Van der Waals densities (maximum possible molecular density)(3-4). They can also be made to self-assemble into sacs that conform to the dimensions of any cavity (7). The goal of the proposed studies is to validate this potential new approach to brain tumor therapy that involves construction of an artificial extracellular matrix that differentiates tumor cells.

Hypothesis: Self-assembling peptide amphiphiles (PAs) with sequences designed to differentiate BTSCs can limit or prevent recurrence of glioblastomas. The capacity to self-assemble and signal to surrounding cells for prolonged periods of time can overcome the physical limitations of drug delivery into the brain parenchyma, the dispersion of cancer cells within brain parenchyma, and the refractoriness of BTSCs to standard therapies.

We have developed a class of bioactive PAs that self-assemble into networks of nanofibers that emulate the architecture of ECMs (3-5) (Fig. 1A). Self-assembly is triggered when PA molecules contact ions in extracellular spaces in tissues leading to networks of nanofibers (Fig. 1C) that can display on their surfaces ultra high densities of biological signals. Examples of bioactive epitopes include cell adhesion sequences, neurite-extending epitopes, and growth factor binding sequences. The PAs can be injected as a liquid into brain or spinal cord where they self-assemble and signal to surrounding cells without evidence of toxicity (6).

Another novel system (7) involves a PA-polymer self-sealing sac that can be formed instantly by injecting polymer directly into PA solution (Fig. 2A) (7). This generates a robust sac that is strong enough to be suturable (Fig. 2B) and that can also self-seal when a hole is made in it (for example when punctured by a syringe to inject something inside it)(Figs. 2C). The sac structure has three distinct regions of potential sites for bioactivity: inside the sac, within the membrane, and on the outer surface of the sac (Fig. 2E). The sac is permeable to large proteins and stable in cell culture media. It is therefore possible to encapsulate cells in the sac and to keep them viable for long time periods (Fig. 2F,G). Its unique structural and physical characteristics, i.e. the ability of sacs to conform to a cavity of arbitrary shape, offer significant potential for delivery of cells, proteins or drugs in brain tumor cavities.

We first used neural stem cells (NSCs) to define effects on cell proliferation and differentiation of PAs with different bioactive sequences. IKVAV is a neuroactive sequence known to interact with integrin receptors and to promote process outgrowth from neurons. Nanofibers containing the IKVAV epitope suppressed astrocytic differentiation of cultured NSCs and promoted exit from cell cycle with neuronal differentiation (5). Subsequent studies in an animal model of spinal cord injury demonstrated that the IKVAV-PA exerts similar effects on glial precursor cells in vivo (6). IKVAV-PA also promoted rapid exit from cell cycle of cultured mouse or human embryonic stem cells (ESCs). Transplantation of undifferentiated ESCs leads to formation of teratomas, and injection of ESCs into cortex without PA led to tumor formation in 100% of the animals with death by 3 months of 70% of the mice. However after co-injection of PA along with the cells, there were no deaths after 3 months and tumor formation in only 10% of the animals. Similar beneficial effects of the IKVAVPA were found after subcutaneous injection of ESCs and after intracerebral injection of a rodent glioma cell line. Despite these exciting findings there were several intellectual concerns about pursuing the IKVAV-PA alone as a therapy for gliomas. First, the IKVAV sequence must be attached to the PA to exert its effects which means that only cells contacted by the PA would be treated. Second, IKVAV-PA was able to suppress most but not all proliferation of NSCs. These concerns suggested that an added approach would be necessary. BMP signaling suppresses proliferation and tumor formation by human BTSCs (9), but delivery of BMPs to brain tumors is problematical since soluble growth factors delivered to tumor beds are rapidly cleared (1,2). However we have now created a BMP-binding/releasing PA that can deliver BMPs to cells over a prolonged period of time. Culture of NSCs or human ESCs in BMP-PA replicated effects of addition of soluble BMPs to culture medium. It then became possible to consider use of the two PAs together.

We have developed fifteen BTSC lines from tumor tissue harvested from patients undergoing surgical resection for GBM. In culture, single BTSCs expand to form clonal spherical cell aggregates, that we and others have termed gliomaspheres. When dissociated, human gliomasphere cells can divide indefinitely to give rise to gliomaspheres in a serial fashion, consistent with their stem-cell like identity. In addition, they give rise to differentiated progeny resembling neurons, astrocytes and oligodendrocytes when cultured in conditions that differentiate normal NSCs. Finally, BTSCs give rise to brain tumors that resemble the parent tumor when implanted in brains of immunocompromised NOD/SCID mice, demonstrating that the BTSCs are stem-cell like and tumorigenic. However we found that treatment of the cells with either BMP-PA or IKVAV-PA caused the BTSCs to differentiate into neurons and mature appearing glia (Fig. 3). Despite the entirely different signaling pathways activated by IKVAV and BMP, the characteristics of neurons that they induced from human BTSCs were similar. These exciting but very preliminary findings suggest that these nanoengineered materials could potentially provide the first meaningful treatment for GBMs. Testing this highly exploratory but potentially exciting approach is the focus of this proposal.