In the present investigation, we show for the first time that short-term treatment with the intravenously applicable Erufosine causes a transient decrease in the growth of the T98G glioblastoma tumours. This effect was associated with a significant accumulation of Erufosine in the xenograft T98G tumours upon repeated drug applications. Moreover, we observed a trend to an enhanced radiation-induced growth delay of T98G xenograft tumours when fractionated irradiation was combined with short-term Erufosine-treatment. However, these effects failed to translate into beneficial drug effects on fractionated radiotherapy in terms of local tumour control.
We show that repeated parenteral injections of 20 or 40 mg/kg BW Erufosine result in a significant accumulation of this membrane-targeted agent in T98G xenograft tumours. This corroborates findings of Vink and co-workers showing accumulation of the closely related compound Perifosine in KB squamous cell carcinoma xenograft tumours upon repeated oral drug administration . Interestingly, short-term treatment with Erufosine caused a transient growth-inhibition within 5 days. However, the tumours resumed growth within a few days after discontinuation of drug treatment. This suggests that although the drug reaches concentrations in the tumour tissue that are sufficient for cytostatic and/or cytotoxic effects in vivo this treatment schedule is not sufficient to eradicate a considerable number of clonogenic tumour cells and thus, tumour stem cells. In line with our findings a recent study showed growth inhibitory effects of treatment with the closely related Perifosine in another glioma xenograft model (U251). Since the growth delay described by the authors was more pronounced compared to the results obtained in our study we speculate that the higher cumulative dose (475 mg/kg) of Perifosine upon oral administration may be responsible for the improved drug action . Similarly, Li et al.  detected a substantial growth delay in neuroblastoma xenograft tumours in nude mice upon a 30-day treatment with Perifosine. These data suggest that in vivo efficacy of these compounds may depend on an extended treatment schedule.
Otherwise, the heterogeneous outcome of preclinical and clinical studies with Perifosine implicates that the efficacy of treatment with the membrane-targeted APC also largely depends on the cell type. Whereas extended treatment with Perifosine was without single drug effect in prostate cancer xenografts in vivo and no evident clinical effect of Perifosine was seen in pancreatic cancer, squamous cell carcinoma of the head and neck, breast cancer and melanoma [25–28], the drug had clinical activity in hematological malignancies  and to some extent in soft tissue sarcoma or prostate cancer [30, 31].
We also found that T98G xenograft tumours respond to fractionated irradiation with a dose-dependent tumour growth delay. Already with doses of 5 × 3.5 Gy (17.5 Gy total dose) we observed a significant tumour growth delay compared to the non-irradiated controls. This corroborates earlier findings showing exceptional in vivo radiation sensitivity of this glioblastoma cell line but is in contrast to the clinical experience of high radiation resistance in glioblastoma tumours [32–34].
The combination of fractionated irradiation with 8 parenteral injections of 40 mg/kg BW Erufosine 4 days before and during the fractionated irradiation was obviously able to intensify the radiation-induced decrease in tumour volume. The failure to detect more than borderline significant differences in the growth-delay of tumours treated with fractionated irradiation alone and fractionated irradiation plus Erufosine may at least partially be due to the unexpected high rate of local tumour control rates that were observed in response to fractionated irradiation alone in the combination experiment 1): Only a low rate of local controls (14%) had been observed in the dose-finding experiment in the highest dose-group (5 × 4.5 Gy) and no local controls had been observed in the 5 × 3.5 Gy dose-group. In contrast, a largely increased rate of local controls (28.6%) was observed in the combined treatment experiment (experiment 2) upon fractionated irradiation with a dose of 5 × 3.3 Gy. The reason for the discrepancy between the two experiments is unknown. Only 66% of transplanted animals developed subcutaneous T98G tumours which may be indicative of a residual antitumour immunoreactivity in the NMRI nu/nu mice . But Krause and coworkers had already demonstrated in an earlier study that TCD50 values for T98G tumours do not differ in mice that received whole body irradiation prior to implantation of the xenograft tumours . Therefore, we assume that a residual immune response of the host may not be causative for the differential effects observed in the two experiments.
Finally, when analyzing whether the addition of short-term Erufosine treatment to fractionated irradiation would improve tumour control probability by measuring the tumour control rates after additional top up irradiations we found that short-term Erufosine failed to improve the outcome of fractionated radiotherapy in terms of local tumour control. This suggests that although tumour regrowth was slightly retarded by co-treatment with erufosine no additional effect on the eradication of clonogenic tumour stem cells could be achieved.
Notably, the TCD50 value for T98G xenografts detected in the present study (12 Gy Top-up irradiation dose and a total irradiation dose of about 30 Gy under ambient conditions) was in the range of the TCD50 values reported in earlier studies for T98G xenografts after single dose or fractionated irradiation [33, 34, 36] although a comparison between the different studies remains difficult due to variations in the irradiation scheme (single/fractionated/mixed) and oxygenation conditions (ambient/hypoxic). In the present study, Top-up dose irradiations were administered under ambient conditions. Therefore, the lack of a difference in tumour control experiments in spite of a significant effect of Erufosine in growth delay experiments may be due to an increase of tumour hypoxia at the stem cell level induced by Erufosine during fractionated radiotherapy.
Up to now, only few studies are available that tested a potential benefit of APC in combination with radiotherapy in vivo. However, those studies were restricted to the use of the orally available APC Perifosine and the evaluation of tumour growth delay. Consistent with our observations with the novel intravenously applicable APC Erufosine, De la Pena et al.  observed a growth delay of U251 glioblastoma xenografts after repeated oral administrations of the closely related alkylphosphocholine Perifosine but also failed to detect an increased efficacy when Perifosine-treatment was combined with a single dose of 4 Gy at the onset of drug treatment. Given the comparable in vivo radiosensitivity of the two glioblastoma cell lines T89G and U251  the irradiation dose used might be too low for a sustained combination effect, although a growth delay of single modality treatment was detectable.
In contrast, the same protocol of Perifosine-treatment combined with 2 × 5 Gy on days 2 and 4 caused a significant growth delay in a prostate cancer xenograft model compared to untreated controls whereas the single modality treatment was not effective . Notably, even a sustained tumour regression was demonstrated after combined treatment with Perifosine and irradiation (5 Gy on days 2 and 4) in xenografts of squamous cell carcinoma . Importantly, in this study effectiveness of Perifosine-treatment depended on the treatment duration and cumulative drug dose.
Also, the following limitations of the preclinical models have to be carefully considered: (i) In vitro radiosensitivity does not necessarily correlate to the tumour control probabilities determined in vivo. (ii) Translation of findings from xenograft models to the clinical situation is even more difficult, particularly in the case of high-grade glioma, because of obvious differences in the tumour microenvironment and the growth behaviour of the malignant cells in vivo. In this scenario, the xenograft model will underestimate additional modes of drug action, such as inhibition of migration and invasion and antiangiogenic effects [39–42]. However, up to now the use of more appropriate preclinical models for the evaluation of the efficacy of combined treatment approaches with radiotherapy in vivo, e.g. orthotopic or spontaneous tumours, is still sparse because of the limited availablity of image-guided radiation systems for small animals.
In conclusion, the intervenously applicable APC Erufosine with proven ability to cross the blood brain barrier causes a transient decrease in the growth of T98G glioblastoma tumours in vivo but fails to improve efficacy of fractionated irradiation in terms of local tumour control. Further in vivo studies are needed to evaluate whether extended Erufosine treatment may be more effective in terms of radiosensitization and whether the drug may interfere with the known adverse biological factors that limit the efficacy of radiotherapy in vivo.