Lack of efficacy of Doxil® in TNF-α-based isolated limb perfusion in sarcoma-bearing rats

Here we show that Doxil® has minimal antitumour activity in the isolated limb perfusion (ILP) setting and its activity was not enhanced by the addition of tumour necrosis factor (TNF). Doxil® accumulation in tumour tissue was low and also not augmented by TNF. In contrast, activity of free conventional doxorubicin was enhanced by TNF. We conclude that application of Doxil® in a TNF-based ILP is not a useful alternative to free conventional doxorubicin or melphalan.

Isolated limb perfusion (ILP) provides an excellent tool in the treatment of locally advanced tumours. During ILP, high local drug concentrations are possible due to minimal leakage into the systemic circulation, and the effect on vital organs is limited, allowing high dosages to be used. We and others have demonstrated that addition of tumour necrosis factor (TNF) to an ILP with melphalan increased the tumour response dramatically as compared to melphalan alone (Lienard et al, 1992;Eggermont et al, 1996).
Melphalan is used most commonly in ILP, but other agents have been applied with varying success in limb or organ perfusion (Rossi et al, 1992;Weksler et al, 1994;Abolhoda et al, 1997). We observed in ILP that local toxicity was dose-limiting at suboptimal doxorubicin concentrations .
The formulation of doxorubicin in long-circulating liposomes (Stealth liposomal doxorubicin, Doxil s ) prolongs circulation time, decreases toxicity and augments localisation in tumour tissue (Gabizon et al, 2003). We hypothesised that the use of Doxil s in ILP may reduce local toxicity while augmenting tumour accumulation and improving tumour response. In this study, we examined the efficacy of Doxil s in a TNF-based ILP in sarcoma-bearing rats.

Animals and tumour model
Male inbred BN rats (soft-tissue sarcoma model, BN175) and WAG/RIJ rats (osteosarcoma model, ROS-1) were obtained from Harlan-CPB (Austerlitz, the Netherlands). Small fragments (3 mm) of tumour were implanted subcutaneously in the right hindleg, as previously described (de Wilt et al, 1999). Tumour growth was recorded by calliper measurements, and tumour volume was calculated using the formula 0.4(A 2 B) (where B represents the largest diameter and A the diameter perpendicular to B). All animal studies were done in accordance with protocols approved by the Animal Care Committee of the Erasmus University Rotterdam, the Netherlands (Workman et al, 1998).
A tumour response indicates either a partial remission (PR, decrease of tumour volume between -25 and 90%) or a complete remission (CR, tumour volume less than 10% of initial volume).

Isolated limb perfusion protocol
Rat limbs were perfused as previously described (de Wilt et al, 1999). Tumour necrosis factor (50 mg), Doxil s or doxorubicin (400 mg BN175 and 200 mg ROS-1) were added as boluses to the oxygenation reservoir. Control rats were perfused with Haemaccel or placebo liposomes alone. The concentration of TNF was adapted from previous animal studies, and doxorubicin concentrations that yielded no local toxicity were used. All animal studies were approved as stated above (Workman et al, 1998).

Assessment of doxorubicin accumulation in solid tumour during ILP
Accumulation of doxorubicin in tumour and muscle was determined directly after ILP, as previously described (Mayer et al, 1989;Van Der Veen et al, 2000). As the ILP included a thorough washout, there was no intravascular doxorubicin present. All animal studies were approved as stated above (Workman et al, 1998).

Statistical analysis
The results were evaluated for statistical significance using the Mann -Whitney U-test with SPSS for Windows. P-values below 0.05 were considered statistically significant.

RESULTS
Tumour response to Doxil s in TNF-based ILP Perfusion with Doxil s , TNF or buffer alone resulted in progressive disease in all soft-tissue sarcoma-bearing rats ( Figure 1A). Perfusion with Doxil s plus TNF resulted in a short growth delay followed by rapid outgrowth of the tumour, and all rats showed progressive disease. Application of free conventional doxorubicin resulted only in a slight inhibition of the tumour growth, and no rats showed a tumour response. Isolated limb perfusion with conventional doxorubicin combined with 50 mg TNF increased the antitumour activity with a response rate of 83% (PR and CR combined) (Po0.01 compared with doxorubicin alone).
Isolated limb perfusion in osteosarcoma-bearing rats with buffer or conventional doxorubicin alone had no significant effect on tumour growth ( Figure 1B). Isolated limb perfusion with TNF alone resulted in a response rate of 25%. Isolated limb perfusion with conventional doxorubicin combined with TNF further increased the tumour response to 83% (Po0.05 compared with TNF alone or doxorubicin alone). Isolated limb perfusion with Doxil s only, induced slight tumour growth delay comparable to free conventional doxorubicin. Strikingly, ILP with Doxil s plus TNF diminished the tumour response, and none of the rats showed a tumour response.

Accumulation of doxorubicin in solid tumour after ILP
We observed that addition of TNF did not significantly augment the accumulation of Doxil s in soft-tissue sarcoma or osteosarcoma when compared to ILP with Doxil s alone (data not shown). Levels of doxorubicin were significantly increased when TNF was added to ILP with free doxorubicin .

DISCUSSION
In the present study, we demonstrated that ILP treatment with Doxil s combined with TNF in sarcoma-bearing rats does not provide a useful alternative to free conventional doxorubicin. The lack of efficacy of Doxil s is not due to failure of the drug to be active at the tumour site, as dramatic synergy between Doxil s and TNF after systemic treatment has been shown . Rather, we speculate that the liposomes are unable to extravasate into the tumours during the relatively short ILP interval.
In spite of the indicated usefulness of doxorubicin in ILP for the treatment of sarcoma, we and others observed dose-limiting local toxicity after a TNF-based ILP with conventional doxorubicin (Di Filippo et al, 1999;Van Der Veen et al, 2000). Biodistribution and pharmacokinetic studies with Doxil s demonstrated a favourable profile of the liposomal formulations over the free drugs, that is, circulation time was extended, toxicity reduced and tumour localisation was increased (Gabizon et al, 2003). Therefore, we envisioned that Doxil s could be a good alternative to free conventional doxorubicin in ILP. However, Doxil s failed to induce any response in sarcoma-bearing rats, even when applied in combination with TNF. Minimal accumulation of Doxil s in tumour after ILP with or without TNF was observed, whereas considerably higher levels of doxorubicin were found in tumour after ILP with free doxorubicin plus TNF. A possible explanation for the difference in accumulation between free conventional doxorubicin and liposomal doxorubicin is the particle size. Whereas distribution of a small molecule like doxorubicin is diffusion dependent, the transport of particulate matter is convection dependent (Jain et al, 1990;Hobbs et al, 1998). This would indicate that, during ILP, drug distribution is mostly diffusion dependent and not convection dependent.
The increased tumour localisation of Doxil seen after systemic administration is reportedly due to the ability of the pegylated liposomes to avoid accumulation in the liver and spleen and other parts of the mononuclear phagocytic system (MPS), which results in a long circulation time and extravasation through the leaky vasculature of tumours (Gabizon et al, 2003). The volume of distribution of Doxil s is markedly smaller than that of doxorubicin given systemically, reflecting the broad tissue distribution of the latter. The short 30-min circulation time imposed by the ILP procedure is likely inadequate for the  Figure 1 (A) Tumour volumes of subcutaneous implanted soft-tissue sarcoma BN175 after isolated limb perfusion with perfusate alone (n ¼ 6), 400 mg Doxil s (n ¼ 4), 50 mg TNF (n ¼ 6), Doxil s plus 50 mg TNF (n ¼ 8), 400 mg free doxorubicin (DXR) (n ¼ 7), or a combination of TNF and free DXR (n ¼ 6). (B) Tumour volumes of subcutaneous implanted osteosarcoma ROS-1 after isolated limb perfusion with perfusate alone (n ¼ 6), 200 mg Doxil s (n ¼ 6), 50 mg TNF (n ¼ 8), 200 mg Doxil s plus 50 mg TNF (n ¼ 7), 200 mg free doxorubicin (DXR) (n ¼ 6), or a combination of TNF and DXR (n ¼ 6). The mean tumour volumes are shown7s.e.