A phase I study of nelfinavir concurrent with temozolomide and radiotherapy in patients with glioblastoma multiforme

Introduction

Glioblastoma multiforme (GBM) is the most frequent primary malignant brain tumor in adults. The standard of care of newly diagnosed glioblastoma following surgical resection to the extent safely feasible is concurrent temozolomide with radiotherapy followed by monthly temozolomide for 5 days each month for at least 6 cycles [1]. This course of therapy results in a median survival of 14.6 months and a progression-free survival of about 6.9 months as compared with median survival of 12.1 months with radiotherapy alone [1]. However, this standard therapy still only results in a modest survival benefit, hence a clear need for new approaches exists. The addition of a molecularly-targeted therapy to the standard treatment is one such approach.

GBM has been notably resistant to therapies, and in particular, radiotherapy. Ideal radiotherapy for the treatment of tumors, including GBM, is selective treatment effect on tumor cells, with relatively minimal effect on normal cells. Radiosensitizers are agents that enhance selective targeting of tumor cells using radiotherapy. One way radiosensitizers may have this differential effect is through selective targeting of pathways more highly expressed in tumor cells.

In GBM, poorer prognosis is associated with PTEN and EGFR or EGFRvIII (truncated EGFR) tumor mutations, which occur in a third of patients and in up to 40 % of patients respectively [2–4]. EGFR and Ras have been shown to modulate tumor radiosensitivity [5–7]. EGFR has a number of downstream effectors, including Ras and PI3K. EGFR and Ras-mediated radioresistance is mediated, at least in part by PI3K, and phosphorylated Akt (PAkt) is a good marker for this effect [8]. We have previously shown in head and neck cancer that P-Akt is a good predictor of (worse) clinical response to radiation [9]. We and others have shown that blocking PI3K-Akt pathway enhances radiation response in vitro and in vivo [8, 10, 11]. Radiosensitization occurs (when this pathway is blocked) in cells in which this pathway is constitutively activated but does not occur in cells (such as normal tissues) in which this pathway is not activated [8, 9, 11]. Inhibition of this pathway, therefore, is a promising approach for radiosensitization. One difficulty in implementing this therapeutic strategy has been obtaining the means to block this deregulated Akt signaling pathway in patients.

We have found that one class of commonly used drugs, the HIV protease inhibitors (HPIs) interfere with PI3K-Akt signaling. These drugs given in combination with reverse transcriptase inhibitors are the mainstay of the current therapeutic regimens for HIV infected patients. The HPIs are peptidomimetics that inhibit the HIV aspartyl protease, a retroviral enzyme that cleaves the viral gag-pol poly-protein and is necessary for the production of infectious viral particles. One prominent side effect of HPI treatment is insulin resistance [12, 13]. Since Akt plays a key role in the coordinated regulation of growth and metabolism by the insulin/IGF—signaling pathway [14, 15], we explored the possibility that HPIs might inhibit PI3K-Akt signaling in tumor cells and radiosensitize human tumors as a result.

Nelfinavir (NFV) is an inhibitor of the human immunodeficiency virus protease. HPIs have been used continuously in patients with well-characterized pharmacokinetics. The apparent volume of distribution was 2–7 L/kg and the terminal half-life in plasma was typically 3.5–5 h. NFV has an excellent tolerability and safety profile characterized by low rates of toxicity [16]. The most common acute side effects of NFV are gastrointestinal and are mostly mild or moderate in intensity and generally easily managed. Patients who have been on long-term NFV for 1–2 years can have hyperlipidemia, insulin resistance, and fat redistribution. It inhibits cytochrome p450 resulting in high serum concentrations of drugs metabolized by this pathway. There are reports of HIV patients on protease inhibitors who have received radiation therapy; no increase in side effects from the radiation have been reported and clinical outcome may be improved [17]. The standard dose of NFV in HIV patients is 1,250 mg bid given orally [18].

We tested NFV and found that it inhibited Akt at concentrations that are routinely achieved in patients. NFV also sensitized tumor cells both in vitro and in vivo to radiation [19]. NFV acted synergistically with radiation to delay GBM xenograft tumor growth in mice [19]. Therefore, it is a clear potential agent for treatment of glioblastoma. There is a high frequency of Akt activation in GBM, linked to the pathogenesis of the tumor [19]. NFV has been shown to be distributed in brain tissue while on therapy and is likely to have increased brain penetrance during fractionated radiotherapy when disruption of the blood–brain-barrier may occur. In addition, it has not been shown to sensitize normal tissues to radiation [19]. NFV has been safely administered to HIV positive patients over the last decade with minimal side effects. Thus, if NFV could safely be combined with the standard chemoradiation protocol, the possibility of enhanced efficacy could be explored. We conducted a phase I study of the combination of NFV with temozolamide and radiation therapy in patients with glioblastoma to identify the maximally tolerated dose (MTD), define the toxicities, and explore potential clinical benefit.

Methods

Results

From July 2009 through April 2011, 21 patients were entered into the study at the Abramson Cancer Center at the University of Pennsylvania. Their characteristics are given in Table 1. All patients completed the study and were evaluable. None were lost to follow-up. Four patients discontinued treatment because of toxicity (three patients with grade 3 hepatotoxicity and one patient with grade 3 diarrhea). Otherwise, discontinuation was attributable to progressive disease: one patient was dose discontinued after being hospitalized for severe agitation associated with tumor progression and another patient with grade 4 DVT/PE. At progression of disease, patients received standard salvage therapy which included interventions such as reresection, bevacizumab infusion, and reirradiation.

A schematic of treatment delivery is depicted in Fig. 1. During the first part of the trial, no DLT was observed at the 625 mg twice daily dose level in the first 3 patients treated at that dose level (Table 2). No DLT was seen in the first 6 patients treated at 1,250 mg twice daily dose level. The 1,250 mg twice daily dose level was then expanded to include a total of 18 patients. In this cohort, four DLTs were observed. Three patients experienced grade 3 hepatotoxicity and one patient grade 3 diarrhea (Table 3). Of the three patients affected by hepatotoxicity, two were receiving NFV and the study drug was discontinued. The third patient experienced elevated liver enzymes 2 weeks after NFV treatment was completed. Liver enzymes of all three patients reverted to normal without additional specific treatment.

The only probable treatment related non-hepatotoxic grade 3/4 related adverse event recorded was grade 3 diarrhea in one patient. Serious non-hepatotoxic adverse events which may or may not have led to dose change or discontinuation occurred in 10 total events in 8 patients. These were attributable to diarrhea in 1, deep venous thrombosis/pulmonary embolus in 3, hyperbilirubinemia in 1, seizure in 1, transient ischemic event in 1, hyperkalemia in 1, and agitation requiring hospitalization in 2. The transient ischemic event, hyperkalemia, seizure and agitation were considered to be not related to treatment, but it was not thought possible to fully assess the relationship of the remaining events to treatment. In particular, however, there was no excess of thrombotic/embolic events, no hemorrhagic events grade ≥2, and no cardiac or hypertensive adverse events.

Although the study was not designed specifically to look at drug efficacy, this preliminary data was collected. The median overall survival was 13.7 months (range 3.2–39.2 months). The median PFS was 7.2 months (range 2.6–18.9 months). The patterns of progression included out of treatment field recurrence in 3 patients among which two patients had diffuse leptomeningeal disease (Fig. 2), and local tumor progression in 18 patients. The mean PFS among the three patients with out of field (OOF) recurrence was 14.5 months.

Discussion

The present study was intended to investigate DLT’s at two NFV dose levels (625 and 1,250 mg bid) in combination with standard dose temozolomide and radiotherapy and establish the MTD. Treatment in the NFV 625 mg bid cohort was well-tolerated, with no DLT seen in the patients treated at this dose level. At the 1,250 mg bid dose level, four DLT’s were identified, three of which were hepatotoxicity related and one gastrointestinal, giving a DLT rate of 22 % (4 of 18). In addition to the four DLT’s, there were two other serious adverse events for which dose was discontinued—one patient with DVT/PE and another with grade 3 agitation due to tumor progression which was not attributable to the treatment regimen.

At 625 mg twice daily, there did not seem to be a greater level of toxicity than that expected with temozolomide alone. In the first 6 patients enrolled at the 1,250 mg bid dose level, there were no DLTs, defining this dose as the MTD. Although only the patients in the 625 mg bid group and first 6 patients enrolled in the 1,250 mg bid group were used to define MTD in our study, treatment and evaluation were continued thereafter. At 1,250 mg twice daily, the total number of NFV doses delayed or discontinued for reasons of hepatotoxicity, diarrhea, or DVT/PE was 5. NFV was approved for clinical use in HIV patients in 1997 [20]. The standard dose of NFV in HIV patients is 1,250 mg bid given orally [18]. The pharmacokinetic properties of NFV were evaluated in healthy volunteers and HIV-infected patients; no substantial differences were observed between the two groups [18]. To date, NFV has had an excellent tolerability and safety profile that is characterized by low rates of toxicity and related discontinuations [16]. The most common acute side effects of NFV are gastrointestinal (GI). The diarrhea related to NFV is mostly mild or moderate in intensity, occurs at the initiation of therapy and is generally easily managed with loperamide and oral hydration. In a 48-week study of NFV administered at 1,250 mg bid, however, 18 % of patients experienced moderate or severe diarrhea; although, for the majority of cases, this adverse effect resolved [21]. In our study, although many patients experienced GI side effects such as diarrhea or nausea/vomiting, most were mild and easily managed with loperamide and oral hydration. Only one patient at the 1,250 mg bid dose level experienced dose-limiting diarrhea. To date, there are little toxicity and efficacy data available in cancer patients who have taken NFV. Even though the number of DLTs at the higher dose regimen was increased, 1,250 mg NFV twice daily in combination with temozolomide (75 mg/m² daily) and radiotherapy is the recommended dose given appropriate concurrent monitoring of liver function tests.

Three patients experienced OOF tumor recurrence, two of whom had leptomeningeal disease. The percentage of patients with OOF recurrences was 14.3 %, and previous studies report between 4.2 and 20 % [22, 23]. The PFS among the three patients with OOF recurrence is more than double the overall mean PFS suggesting the possibility that better local tumor control for a longer period led to increased likelihood of first recurrence elsewhere. It is tempting to speculate that NFV may alter the natural history of OOF tumor progression, but a larger study would be needed to show a statistically significant effect. Genetic and biologic correlates of tumor typing may help identify patients that would benefit from the addition of NFV to standard therapy.

This study was primarily intended to establish safety, but we report preliminary data on efficacy which is difficult to interpret given the small number of patients. Another limitation to the interpretation of our efficacy data is lack of analysis of MGMT methylation status in these patients. Since MGMT-methylated patients tend to have better outcomes after treatment with temozolomide, this may be a confounding factor in patients with longer survival and PFS in our study. In addition, further conclusions about efficacy in a small number of patients would have been enhanced by the characterization of other molecular markers for these tumors, such as phospho-Akt, which might directly determine the clinical outcomes for these patients on NFV. The observed patterns of progression/recurrence were mostly local tumor progression within the treatment field. Further study of NFV in combination with temozolomide and radiation is warranted. This phase I trial provides toxicity data to facilitate such future studies.

Conclusion

The combination of NFV (1,250 mg twice daily dose) with temozolomide (75 mg/m² daily) in combination with radiotherapy is feasible and relatively well tolerated. At the 1,250 mg twice daily dose, main dose-limiting toxicities observed were hepatotoxicity and GI related side effects such as diarrhea. The recommended dose of NFV is 1,250 mg twice daily with weekly monitoring of liver function tests.