|Home | About | Journals | Submit | Contact Us | Français|
Glioblastoma multiforme remains refractory to conventional therapy, and novel therapeutic modalities are desperately needed. TP-38 is a recombinant chimeric protein containing a genetically engineered form of the cytotoxic Pseudomonas exotoxin fused to transforming growth factor (TGF)-α. TGF-α binds with high affinity to the epidermal growth factor receptor, which is uniformly overexpressed in malignant gliomas, often because of gene amplification. Prior to therapy with TP-38, the patient described here was completely refractory to multiple other therapies, with radiographic and pathologic evidence of tumor progression. After therapy, she improved clinically, was weaned off steroids and anticonvulsants, and experienced a progressive decrease in enhancing tumor volume. Despite multiple prior recurrences, she has not progressed for >43 months after TP-38 therapy. Small remaining areas of enhancement demonstrate no evidence of tumor histologically and are hypometabolic on positron emission tomography. This report describes a dramatic and sustained clinical and radiographic response in a patient with a bifrontal glioblastoma multiforme treated with intratumoral infusion of a novel targeted toxin, TP-38.
Despite aggressive surgical resection, high-dose focused radiation therapy, and multiagent chemotherapy delivered at toxic doses, patients diagnosed with malignant primary brain tumors have a median survival of only 40 to 50 weeks after initial diagnosis (Galanis and Buckner, 2000) and only about 25 weeks following recurrence (Wong et al., 1999). Failure of conventional therapies can be attributed, at least in part, to their lack of specificity for neoplastic tissue, which results in dose-limiting systemic or neurologic toxicity (Imperato et al., 1990; McGirt et al., 2002). Specific targeting of tumor cells, for example, through the use of biologic ligands with high affinity for receptors overexpressed by tumor cells, can be exploited to deliver therapeutic moieties more specifically, and this should enhance therapeutic potential and reduce nonspecific toxicity.
To test this hypothesis, we developed TP-38, a 43.5-kD recombinant chimeric protein containing the transforming growth factor (TGF)-α3 ligand fused to a genetically engineered form of the cytotoxic Pseudomonas exotoxin A, PE38 (amino acids 253–364 and 381–613). TGF-α binds with high affinity to the epidermal growth factor (EGF) receptor, which is uniformly overexpressed, often because of gene amplification, in malignant gliomas (Wikstrand et al., 1998, 2002; Wong et al., 1987). The EGF receptor is expressed at only very low levels or is undetectable on normal human glial cells and neurons, which suggests a potential therapeutic window (Libermann et al., 1984; Torp et al., 1991). Replacing the native binding domain of Pseudomonas exotoxin with TGF-α specifically targets the toxin to the neoplastic cells within the brain. Our first assessable patient treated with this agent, who harbored a therapeutically resistant bifrontal glioblastoma multiforme (GBM), experienced a dramatic radiographic and clinical response in the complete absence of any toxicity.
In September 1999, a 59-year-old Caucasian female was evaluated in a local emergency room following a generalized seizure. At that time, she reported a two-week history of progressive short-term memory difficulties, intermittent confusion, and word-finding difficulties. A magnetic resonance (MR) image revealed a multilobular, irregularly enhancing lesion measuring 4.5 cm in greatest dimension in the posterior aspect of the left frontal lobe, with extension into the insular cortex and premotor cortex. Significant vasogenic edema was noted, as well as midline shift measuring 11 to 12 mm. The patient underwent a subtotal resection, with histology demonstrating elongated fibrillar tumor cells associated with pseudopalisading necrosis and vascular proliferation (Fig. 1A). The local neuropathologist diagnosed GBM, and this diagnosis was confirmed by a review of the histology by the neuropathologist at Duke University (R.E.M.). She received chemotherapy according to North Central Cancer Treatment Group (NCCTG) protocol 987252, which included BCNU (carmustine), cisplatin, and oral etoposide. An MR image after three cycles of chemotherapy revealed stable disease. She then underwent conventional external beam radiotherapy and received 6000 cGy in 30 fractions to the contrast-enhanced tumor volume plus a 2-cm margin via a 4-field technique. An MR image taken at the completion of her radiation therapy was also unchanged. The patient then received two additional cycles of BCNU chemotherapy at 134 mg/m2, but this therapy was complicated by grade 3 thrombocytopenia. After her second course of BCNU, an MR image revealed significant enlargement of the enhancing lesion in the left frontal lobe accompanied by an increase in associated vasogenic edema, including new extension of edema across the corpus callosum.
She then underwent a gross total resection of enhancing tumor with carmustine-loaded polymer wafer (Gliadel; Guilford Pharmaceuticals Inc., Baltimore, Md.) placement. Pathology from this procedure confirmed recurrent GBM (Fig. 1B). Postoperatively, she received temozolomide at 150 mg/m2/day for five consecutive days every 28 days. After two cycles, an MR image revealed new nodular enhancement within the resection cavity, as well as extension of enhancement into the genu of the corpus callosum, consistent with recurrent disease.
She then began tamoxifen at a dose of 80 mg orally twice a day in combination with irinotecan, administered at 340 mg/m2 weekly for four consecutive weeks. The irinotecan was reduced to 300 mg/m2 after she developed grade 3 thrombocytopenia following her first dose. Unfortunately, after one cycle of irinotecan with tamoxifen, an MR image revealed further enlargement of the enhancing lesion including further extension across the corpus callosum.
The irinotecan and tamoxifen were discontinued, and she began cyclophosphamide administered orally at 100 mg/day for the next three months. This therapy was complicated by grade 4 neutropenia requiring granulocyte colony–stimulating factor therapy. An MR image again demonstrated additional enlargement and extension of the enhancing lesion consistent with further progressive disease (Fig. 2, panels at left labeled “Pre,” axial and coronal views).
In March 2001, 18 months after diagnosis and after her fourth consecutive radiographic progression, the patient was enrolled on a phase 1 trial of convection-enhanced delivery (CED) of TP-38 (IVAX, Miami, Fla.) approved by the Duke University Health System Institutional Review Board. Because data from 18F-fluorodeoxyglucose (18F-FDG) PET, MR spectroscopy, or biopsy immediately prior to TP-38 treatment is not available, it remains possible that the targeted enhancing mass consisted entirely of radiation necrosis. This seems unlikely, however, given that in July 2000, this same enhancing area was partially resected and histologically determined to be GBM. Her neurologic examination at the time of TP-38 therapy was remarkable only for grade 1 to 2 deficits in short-term memory and concentration, and her Karnofsky performance status (KPS) was 90%. TP-38 was administered via two multiport pediatric ventricular catheters (CSF-ventricular catheter #41207, Medtronic PS Medical, Minneapolis, Minn.) placed stereotactically into the left and right frontal cortices, respectively (Fig. 3). Twenty milliliters of TP-38 were infused through each catheter. The infusion was administered at a rate of 0.1 ml/h for 1 h, then 0.2 ml/h for 1 h, and then 0.4 ml/h and required 51.25 h for completion. The TP-38 concentration was 25 ng/ml, resulting in a total delivered TP-38 dose of 1 mg. Intracranial pressure was monitored from each catheter, and no significant increases in intracranial pressure were observed throughout the duration of the infusion. The patient tolerated the TP-38 infusion without complication and was discharged.
Over the next six months, the patient underwent a complete physical examination and brain MR imaging at two-month intervals. Her physical examination remained stable while her MR images demonstrated a progressive decrease in the residual enhancing tumor (Fig. 2). An MRI scan approximately 13 months after TP-38 administration revealed continued decrease in size of the bifrontal enhancing lesion; however, a small new nodular area of subependymal enhancement was noted along the roof of the left lateral ventricle. An 18F-FDG PET scan of the brain revealed a level of 18F-FDG uptake comparable to that of gray matter within the residual enhancing bifrontal tumor as well as within the area of new enhancement within the subependyma of the left lateral ventricle. She underwent a stereotactic biopsy of the subependymal lesion in the left lateral vent normal white matter with mild gliosis and no evidence of high-grade glioma (Fig. 4). The 18F-FDG PET brain scan co-registered with the axial MR image taken at this time demonstrated that remaining areas of contrast enhancement on MR imaging were metabolically inactive, suggesting necrosis (Fig. 5). Subsequent 18F-FDG PET scans revealed no evidence of hypermetabolic activity in any areas.
Since the biopsy, the patient has continued to undergo regular evaluations, including a complete physical examination and brain MR image, every 3 to 4 months. She is now >43 months from her TP-38 infusion and >5 years from her initial diagnosis of GBM. She reports that her short-term memory difficulties have improved steadily since treatment with TP-38, and she has not developed any new neurologic symptoms. Her neurologic examination remains otherwise normal and her KPS remains 90%. She has been completely weaned off steroids and no longer requires anticonvulsants. Her MR image demonstrates a continued, gradual decrease in the size of her residual enhancing areas (Fig. 2).
Targeted toxins, like TP-38, are emerging as potent and specific antineoplastic agents. Dramatic clinical responses in therapeutically refractory patients have been well documented in leukemia (Kreitman et al., 2000, 2001) and lymphoma (Frankel et al., 2000; Olsen et al., 2001). The success of these agents has been related to several factors. First, these agents are several orders of magnitude more potent than conventional chemotherapeutics. For example, while alkylating chemotherapeutic agents require thousands of molecules to kill a single tumor cell, a single genetically engineered toxin molecule can react enzymatically with multiple intracellular targets, which resulted in the complete interruption of cellular protein synthesis, leading quickly to cellular apoptosis. Second, these toxins can be cytocidal regardless of cell cycle kinetics or proliferation rate. Third, factors such as hypoxia that limit the activity of radiation and chemotherapy have no effect on the activity of these toxins, and in fact, no mechanism of resistance to these toxins has been described.
Results with targeted toxins in most studies treating solid tumors, however, have not been as encouraging (Frankel et al., 2000). This may be because these relatively large molecules fail to diffuse well in solid tumors. Because drug diffusion into solid tumors is dependent on a high concentration gradient, is inversely related to the size of the agent, is inhibited by high intratumoral pressures, and is usually slow relative to tissue binding or clearance, delivery of macromolecular toxins to solid tumors is impeded (Jain, 1990, 1996). Diffusion of systemically delivered targeted toxins into brain tumors is even further hampered by the blood-brain barrier.
The specificity of agents for brain tumor therapy, however, can be significantly increased by regional administration directly into the anatomically isolated intracranial or intrathecal compartments (Brown et al., 1996; Pastan et al., 1995). Direct intracerebral delivery, as used in this study, bypasses the blood-brain barrier, delivers high concentrations of therapeutic agent directly to the tumor, and reduces systemic exposure to drug-induced toxicity (Morrison et al., 1994). Delivery of macromolecules to solid tumors can be further enhanced by CED, a pressure gradient–dependent microinfusion of therapeutic agents (Bobo et al., 1994). CED has been predicted to produce a bulk flow current that has the potential to homogeneously distribute even large molecules much greater distances throughout the extracellular spaces of the brain (Morrison et al., 1994). The potential benefit of CED in the treatment of brain tumors has already been demonstrated by us and others in experimental animals (Bobo et al., 1994; Grossi et al., 2003; Heimberger et al., 2000; Lieberman et al., 1995), and CED of target toxins has been shown to be efficacious in other patients with malignant gliomas (Laske et al., 1997; Rand et al., 2000). In the trial by Laske et al., nine out of 15 assessable patients had a 50% reduction in tumor after receiving a chemical conjugate of transferrin and a mutant form of diphtheria toxin. In that study, most of the patients had much smaller tumors than the case reported here, however. Furthermore, a number of their patients developed a delayed toxicity, including neurologic deficits such as weakness, likely related to the specific binding of the transferrin to the normal cerebral vasculature (Hagihara et al., 2000; Laske et al., 1997). Interestingly, in all reported brain tumor series including our own (Sampson et al., 2003), responses have been slow to manifest, often requiring many months to become apparent. This raises the possibility that (a) elimination of contrast enhancement may be significantly delayed, thus confounding radiographic response evaluations in clinical trials in patients with malignant gliomas, or alternatively (b) that a significant bystander effect, potentially immunologic in nature, may be playing a role in the efficacy of these agents. Of interest in this regard is the finding of oligoclonal banding in this patient’s cerebrospinal fluid. Although preliminary results of the trial in which the patient reported here participated have been previously reported (Sampson et al., 2003), that brief and preliminary report provided no details on the clinical history nor the pathologic or radiographic findings for this patient. Therefore, this patient has been presented here separately to allow readers to make independent judgments and derive independent hypotheses that might explain the sustained responses that can be seen with these agents.
Despite the encouraging results represented by our case study and other results published on the use of intracerebral infusion of targeted toxins for malignant brain tumors, a number of shortcomings remain. For example, little work has been done to optimize the parameters relevant to this novel mechanism of drug delivery to the brain. Optimization of these parameters and elucidation of the mechanism of tumor response initiated by these novel compounds are needed to unveil the true promise of these reagents.
TP-38 is a recombinant toxin that specifically targets the EGF receptor, which is uniformly overexpressed on malignant gliomas. Intracerebral infusion of this agent using CED in the patient reported here has been associated with a sustained clinical and radiographic improvement without any evidence of toxicity.
1The authors acknowledge support for this work from the National Institutes of Health through Grants K23 RR16065 and 5RO1-CA97611-05.
3Abbreviations used are as follows: CED, convection-enhanced delivery; EGF, epidermal growth factor; 18F-FDG, 18F-fluorodeoxyglucose; GBM, glioblastoma multiforme; KPS, Karnofsky performance status; MR, magnetic resonance; TGF, transforming growth factor.