Medulloblastoma: need for targeted treatment

The prognosis for children suffering from a brain tumor has improved greatly over the past few decades. For medulloblastoma, the most common malignant brain tumor in childhood, the 5-year progression-free survival for standard-risk patients has risen to almost 80% with the introduction of the Packer treatment regimen [1]. This protocol consists of neurosurgery, aiming for complete resection, postoperative radiotherapy and concurrent vincristine, followed by cisplatin, vincristine and chloroethylcyclohexylnitrosourea (CCNU, lomustine) chemotherapy. Of course, a survival of 80% means there is still much room for improvement. However, with increased survival, another question concerning childhood brain tumor treatment also becomes more and more important: what effect does the treatment have in the later life of survivors?

Within the treatment of malignant pediatric brain tumors, radiotherapy takes up a central role in a child’s chance of survival. However, it is also the main cause of late adverse side effects after treatment. These side effects range from cognitive damage and hormone dysfunction (e.g., growth disorders and early puberty) to loss of hearing and taste [2–4]. The severity of side effects and the extent to which they occur are dependent on the amount of radiation that is received [4–6]. Even with the standard radiation dose having decreased since the introduction of the Packer protocol, side effects can still be substantial. Ris and colleagues reported an average loss in IQ scores of over four points per year for the first 3 years after treatment with this protocol. In specific subgroups (i.e., younger children or those with a higher baseline IQ score at the start of treatment), this loss was even more severe [6]. The side effects are caused by damage that is inflicted on the healthy tissues within the irradiated field. In general, side effects are caused by lack of specificity of treatment for the tumor, and improving specificity would allow for greater survival and/or fewer and less severe treatment-related side effects. Within major research fields, such as that of non-small cell lung cancer, abundant attention has been focused on exploring possibilities of targeted treatment, leading, for instance, to the introduction of epidermal growth factor receptor (EGFR) inhibitors in clinical trials [7]. However, for children with medulloblastoma, treatment still consists of relatively unspecific conventional modalities. Little gain has been made in the area of targeted treatment, although it is here that opportunities for improvement lie. To achieve greater treatment specificity, several routes are open: the use of tumor-targeted toxic therapies, neuroprotectors and radio- and/or chemosensitizers.

Tumor-targeted therapy

In order to target toxic therapy towards medulloblastoma cells, the differential expression patterns of tumor and healthy tissue have to be exploited. Therefore, it is very important to know which proteins are differentially expressed in the tumor compared with normal tissue. The signaling pathways that are most often aberrantly expressed in medulloblastoma are the hedgehog, wingless and notch pathways [8]. As such, they are likely candidates for targeted treatment. Although there is considerable interest in exploiting these pathways, to date, not many agents that specifically target them are available. However, those agents that are available have demonstrated efficacy in preclinical models [9,10]. However, many more preclinical data have to be gathered before they can be taken into clinical development. As a primitive neuroectodermal tumor, medulloblastoma is characterized by undifferentiated cells that still have the potential for differentiation. A very elegant approach would be to direct tumor cells towards maturation. This would be beneficial in two respects. Higher differentiation is associated with tumor growth inhibition owing to dimished cell division. Indeed, the expression of differentiation-related genes is correlated to better outcome [11]. Moreover, normal tissue function would be less disturbed with tumor cells (partly) regaining function after forced differentiation. In fact, the aforementioned aberrantly expressed pathways are involved in differentiation of neuronal cells, thus treatment directed at aberrant expression within these pathways will often be differentiation therapy. Several substances have already been found to induce differentiation in medulloblastoma cells in vitro, including, among others, phenylacetate, valproic acid and resveratrol [12–15]. The exact mechanisms by which they do so, however, are still largely unknown. Histone deacetylase (HDAC) inhibition is involved in the effects of at least some of these substances, but it is unclear which downstream pathways are important [12,15]. Further research needs to be performed to unravel the determining mechanisms.

Neuroprotectors

Whereas the development of directly tumor-targeted therapy, as described above, is aimed at finding new agents that have a greater therapeutic index than conventional treatment modalities, another possibility would be to use agents that do not have a direct antitumor effect, but increase the therapeutic index of other treatment modalities. Neuroprotectors are agents that are able to protect healthy brain tissue from the toxic effects exerted by antitumor treatment. In general, there has been much attention over the use of antioxidants as protectants in cancer treatment. However, their use is debated heavily as there is great controversy as to whether they protect normal tissue or the tumor (for a thorough review of the discussion see [16–18]). The issue is complicated and both in vitro and clinical studies demonstrate seemingly contradictory results. Some antioxidants have protective effects that are not related to oxidation, which may cloud the discussion. For instance, retinoic acid also induces differentiation and influences DNA repair activity [19,20]. The antioxidant action of so-called ‘antioxidants’ is not absolute and depends on factors such as tissue oxygen pressure and presence of other antioxidants, which means that they can have a pro-oxidant and, thus, procarcinogenic activity [18,21]. Overall, it must be concluded that whether the antioxidants are beneficial depends very much on the setting in which they are administered. Since the issue is so complex and there are indications that administration of antioxidants can be counterproductive, it remains doubtful whether, in the future, antioxidants will be included in medulloblastoma treatment regimens. Other than for antioxidants, there has been little attention for neuroprotectors in cancer treatment.

Radiosensitizers

To increase the therapeutic index of conventional treatment, another option is to use agents that specifically render tumor cells more sensitive to treatment-induced cell death rather than protecting normal cells. As radiation is thought to be the main cause of adverse side effects in children, so-called radiosensitizers could significantly improve therapy for medulloblastoma. There is considerable interest in radiosensitizers in many tumor types, which offers the opportunity to make use of experiences that exist in tumors other than medulloblastoma. To achieve radiosensitization, proteins that are involved in tumor cell response to radiotherapy must be targeted. If there is a difference in expression of these proteins between normal tissue and tumor cells, or if tumor cells are more dependent upon the pathway(s) involved, selectivity for tumor cells can be achieved. The EGFR family of receptors is a group of such targets that have been implicated in radiosensitivity and are expressed aberrantly in medulloblastoma [22,23]. Radiation can activate these receptors directly and this is thought to induce accelerated tumor repopulation between radiotherapy fractions [24]. Also, there are indications that the radiation-induced activation directly stimulates DNA repair activity [25,26]. The extent to which these effects play a role is dependent upon the levels of EGFR family members in the tumor cells. When tumor cells show overexpression, inhibition of the EGFR family should be able to selectively enhance the tumor’s response to irradiation. Indeed, we have found that inhibition of the EGFR family by CI-1033 can induce substantial radiosensitization in medulloblastoma cells (Luttjeboer M and colleagues, unpublished observations). Cyclooxygenase (Cox)-2 is another protein that is thought to be involved in the response to radiation and has been investigated as a target for radiosensitization in glioma and other tumors [27]. Cox-2 is not (or minimally) expressed under normal conditions, but is expressed in disease conditions, such as rheumatoid arthritis or cancer. For medulloblastoma, overexpression of Cox-2 has been reported, and it could, therefore, be a target for radiosensitization [28]. The mechanism behind radiosensitisation via Cox-2 inhibition is unclear, but inhibition of DNA repair has been reported [29]. One disadvantage is that the specific Cox-2 inhibitors that are available are used in concentrations that are much higher than is necessary to inhibit Cox-2 in vitro and, therefore, the radiosensitization found may be an aspecific by-effect [27,30]. As Cox-2 is induced by nuclear factor (NF)κB, there is a possibility that overexpression of Cox-2 is a surrogate marker for high NFκB activity. High NFκB activity is a well-known cause of apoptosis resistance and its inhibition is, therefore, investigated extensively as a treatment strategy for cancer [31]. To date, this has mostly been as a single-agent approach, although recently there has been increasing interest in the use of NFκB inhibitors as chemo- and radiosensitizers. With NFκB inhibitors such as bortezomib already in clinical use, the development of other treatment protocols comprising these inhibitors in a multiagent approach could happen relatively quickly. Therefore, further preclinical research into the use of NFκB inhibition as a strategy for radiosensitization in medulloblastoma is certainly warranted.

From bench to bedside

There is ample opportunity for the development of a targeted treatment strategy to improve outcome for medulloblastoma patients, although there is still much work to be carried out in the laboratory. In future, a combination of one or more targeted agents with conventional treatment modalities will probably be used to treat children with medulloblastoma. How long it will take before we get there, however, is also a political question. Medulloblastoma is a tumor that is typical in childhood. It occurs very rarely between 20 and 40 years of age, and not at all in later life. Moreover, it is well known that pharmacokinetics differ in adults and children. This makes it essential for clinical trials to be held in pediatric patients. To date, this has been an obstacle for the development of improved therapy. In some EU countries, Phase I clinical trials in incompetent subjects, including children, are prohibited by law. Also, there is little incentive for pharmaceutical companies to develop new medulloblastoma medication as the market for such medication would be small. Therefore, the EU’s move towards a system such as that of the USA is welcomed, where pharmaceutical companies are obliged to test new drugs in the pediatric population when there is a rationale to do so, while offering extended patent rights in return. After all, it is only when both laboratory and clinical research can be used to their full capacity that we will be able to find a better cure for children with medulloblastoma.