Human Studies

The extensive number of animal studies partially detailed in the previous section has created a fairly comprehensive understanding of the principles and potential of CED. This data has in turn generated exciting clinical studies which may soon validate the utility of CED as a neurosurgical tool. In this section, completed and ongoing human neuro-oncology studies which utilize high-flow microinfusion as a strategy for drug delivery are reviewed.

In our opinion, the factor that will most limit the use of CED as a tool for the treatment of brain tumors is the effectiveness of the infused agent. So far, the substances that have been tried vary from the mundane to the exotic and include: chemotherapy drugs, designer toxins, radiotoxic conjugates, and liposomes bearing modified viral genomes. The most widely studied agents are the receptor-directed toxins, of which a modified diphtheria toxin

FIGURE 20.3 Tf-CRM107 protein-toxin conjugate components. (A) Native diphtheria toxin. (B) Point mutation disables cell binding domain. (C) Conjugation to transferrin or antibody confers specificity. Illustration reprinted with permission from Science, October 1987, volume 238.

(CRM107) conjugated to transferrin (Tf-CRM107, ''TransMID'') is the prototype and the subject of a phase III clinical trial (Xenova).

Details of the development of CRM107 for CED illustrate the concepts important to understanding the designer toxin therapies. Tf-CRM107 (140000 molecular weight) is a genetically modified diphtheria toxin conjugated to the transferrin (Tf) ligand (see Fig. 20.3). Theoretically, transferrin binds to transferrin receptors which are upregulated in tumor cells but not normal brain, a phenomenon conferring relative specificity. After receptor-mediated endocytosis, the toxin, subjected to a lower pH, unfolds and is inserted into the endocytotic vesicle membrane. The alpha subunit is then released into the cytoplasm and catalyzes the ADP ribosylation of ribosomal elongation factor-2 (EF-2), a protein necessary for protein synthesis. Due to its capacity for catalytic inactivation of numerous ribosomal EF-2 molecules, diphtheria toxin is very potent. Only a single molecule is necessary to kill a cell.

The phase I trial was performed in 15 evaluable patients with malignant gliomas of which nine were glioblastomas, five were anaplastic astrocytomas, and one was an anaplastic oligodendroglioma [39]. Overall, median survival was greater than 75 weeks, median time to progression was 38 weeks, and in nine of the patients, tumor volumes were reduced by more than 50 per cent. Based on magnetic resonance imaging, there were two complete responders, though progression eventually occurred (see Fig. 20.4). Systemic toxicity was manifested by transient elevations of serum alanine and aspartate aminotrans-ferases in fourteen patients. Mild hypoalbuminemia also occurred in twelve patients. Overall, the infusions were well tolerated, though peritumoral toxicity, associated with higher toxin concentrations, occurred in three patients one to four weeks after Tf-CRM107 introduction. Biopsied regions of magnetic resonance abnormalities in three patients who developed significant neurological deficits revealed thrombosed cortical venules and capillaries.

The encouraging results were further confirmed in a phase II trial (unpublished data presented in poster format at the Annual Meeting of American Association of Neurological Surgeons, 2000). Thirty-four out of forty-four patients were evaluated after two stereotactic intratumoral infusions, spaced four to ten weeks apart, of Tf-CRM107 using dual silastic catheters. The phase II toxin concentration was less than the concentration that was associated with peritumoral toxicity in the phase I trial. Five patients were complete responders (15 per cent), seven were partial responders (20 per cent), nine had stable disease (26 per cent), and thirteen had progressive disease (39 per cent). The median survival time was 37 weeks for all the patients and 13 were still alive one year after the initial infusion [40]. Overall, the

FIGURE 20.4 T1-weighted gadolinium-enhanced coronal MRI images depicting a complete response of a glioblastoma multiforme in a 48-year old female to convective delivery of Tf-CRM107. (A) Tumor prior to treatment. (B) Four days after treatment there is increased enhancement. (C) At seven months, the enhancement is resolving. (D) Complete resolution at 14 months with recurrence at 23 months (recurrence not shown). Figure reprinted with permission from Nature Medicine, volume 3, 1997: page 1365.

FIGURE 20.4 T1-weighted gadolinium-enhanced coronal MRI images depicting a complete response of a glioblastoma multiforme in a 48-year old female to convective delivery of Tf-CRM107. (A) Tumor prior to treatment. (B) Four days after treatment there is increased enhancement. (C) At seven months, the enhancement is resolving. (D) Complete resolution at 14 months with recurrence at 23 months (recurrence not shown). Figure reprinted with permission from Nature Medicine, volume 3, 1997: page 1365.

results of the phase II trial were comparable to those obtained in the phase I trial. Currently, a multicenter phase III trial is in progress to compare the efficacy of the Tf-CRM107 receptor directed toxin to the best standard of care in patients with non-operable, progressive or recurrent glioblastoma multiforme.

Following the encouraging results of the earlier Tf-CRM107 studies, other protein-toxin conjugates designed for local delivery are presently undergoing human trials. It should also be noted that a commonly used term for these agents, "immunotoxin", is a misnomer because this term would refer to a toxin conjugated to an antibody. Immunotoxins thus differ from Tf-CRM107 in terms of the targeted receptor and the nature of the toxin itself.

One example of another cytotoxin that is being delivered locally to glioblastomas in clinical trials is the recombinant fusion protein IL13-PE38QQR which is composed of the immune regulatory cyto-kine, interleukin-13 (IL-13), conjugated to a mutated form of Pseudomonas exotoxin [41]. At concentrations of 0.5-2.0 mg/ml, prolonged survival was observed among the 46 patients enrolled, as was histopatho-logical tumor effect. This approach has been further extended by combining IL13-PE38QQR with the novel fusion protein DAB389EGF, with the combination showing enhanced cytotoxicity against glioblas-toma cells in vitro, thus suggesting a synergistic effect of the combined agents. Ongoing phase I/II and III clinical trials are in progress to determine the effect of infusing the IL-13 cytotoxin into recurrent gliomas as a stand-alone therapy, prior to surgery, or after surgery [42].

A trial using a similar fusion protein, NBI-3001, an interleukin-4 Pseudomonas exotoxin combination, was performed in 31 patients with histologically verified supratentorial grade 3 and 4 astrocytomas [43]. Overall, the median survival was 8.2 months with a median survival of 5.8 months for the gliobla-stoma patients. Systemic toxicity was not evident and treatment-related adverse effects, occurring in 39 per cent of the patients, were limited to the central nervous system.

Some investigators have focused the use of high-flow microinfusion on conventional chemotherapeutic agents. For instance, an organic solvent-based BCNU solution is being convectively administered into glio-blastoma multiforme tumor beds. The agent being used for this purpose is DTI-015 (Direct Therapeutics, Inc., San Bruno, California), which consists of BCNU in 100 per cent ethanol at a typical concentration of 45 mg/ml [44]. Diffusion-weighted magnetic resonance tracking performed in a phase I/II dose-escalation study revealed that the volume of distribution is up to 20 times the volume infused [45]. In patients who received equal or less than the maximum tolerated dose, median survival was 55 weeks. A phase III clinical trial of the technique is reported to be under planning, as are European trials.

Lidar et al. convectively administered another chemotherapeutic agent that is typically given syste-mically, paclitaxel, to fifteen patients with glioblasto-mas and anaplastic astrocytomas [46]. The authors achieved a relatively good effect, five complete and six partial responders, but this result came at the expense of a high incidence of adverse effects. Complications included transient chemical meningitis, infectious complications, and transient neurological deterioration. Of interest particularly to future studies, successful distribution correlated with hyperintense signal changes on diffusion-weighted magnetic resonance imaging, whereas non-effective CED was associated with ''leakage of the convected drug into the subarachnoid space, ventricles, and cavities formed by previous resections, and was seen in tumors containing widespread necrosis.''

Yet one more application that is being investigated is the use of CED to distribute radiotoxic species designed to preferentially target glioblastoma multiforme cells. Merlo et al. infused the peptide vector 90Y-labeled DOTA0-D-Phe1-Tyr3-octreotide (DOTA-TOC) into two cohorts of patients, one with low-grade gliomas and one with anaplastic gliomas [47]. Six of the eleven patients showed disease stabilization and tumor shrinkage. Papanastassiou et al. treated seven patients by convective infusion of I131 labeled monoclonal antibodies, and Riva et al. similarly treated fifty patients [48,49]. Remarkably, in the latter study, median survival was 17 months in patients with bulky tumors, and 26 months in patients with minimal disease. Moreover, median time to progression was three months in recurrent and seven months in newly diagnosed gliomas. Based on these results, Riva et al. suggested that direct intra-parenchymal infusion might be particularly useful in newly diagnosed brain tumor cases in which debulking surgery had already been performed.

Using CED to transport gene therapy vectors, such as viruses and liposomes to brain tumors is perhaps the most intriguing application. Feasibility of using bulk flow to distribute nano-sized particles through the extracellular space appears promising, and, in fact, human clinical trials are in progress [28,33,50]. In a phase I/II study Voges et al. convectively administered liposomes carrying the herpes simplex I genome incorporating the thymidine kinase gene to eight patients with recurrent glioblastoma multiforme [51]. Two of the patients had a 50 per cent reduction in tumor volume and six patients had a focal response.

0 0

Post a comment