LETTERS TO THE EDITOR ritonavir, saquinavir, indinavir and nelfinavir. They contend that the concentrations used in our experiments—particularly those demonstrating proteasomal inhibition—are non-physiological. On this basis they argue that the inhibition of the proteasome is a minor contributor to the hyperlipidemic syndrome induced by these pharmaceuticals. The authors point out that not all PIs inhibit the proteasome to the same extent. This is true; however, it is notable (and consistent with our hypothesis) that the severity of the hyperlipidemias matches the proteasomal inhibition profiles of the PIs (ritonavir being the worst). The authors state the Cmax value of ritonavir to be 15 µM (∼ 10.5 µg/ml). Abbot Phamaceuticals (www. fda. gov/cder/foi/label/2001/20945s5lbl. pdf) state the Cmax and Cmin of ritonavir to be 11.2±3.6 µg/ml and 3.7±2.6 µg/ml. The lowest concentration of ritonavir that inhibited ApoB production in our studies was 25 µg/ml (∼ 2.5-fold higher than Cmax). The authors also point out, correctly, that most of the PI is bound to serum proteins. In fact, the major ‘carrier’of ritonavir in plasma is albumin. For this reason our experiments were performed in the presence of 1% BSA or 10% FBS. Moreover given the numerous metabolites of ritonavir detected in plasma, it is impossible to predict what effect the in vivo catabolism of PIs might have on the manifestation of side effects. For example, the isopropylthiazile oxidative derivative of ritonavir is an effective antiretroviral agent. Furthermore, it is likely that all of these compounds would concentrate in the liver such that hepatic concentrations may be higher than serum levels. For these reasons, we believe the concentrations used in our study to be appropriate and informative regarding the underlying cause of these syndromes.

We thank the authors for their comment regarding the toxicity of ritonavir and saquinavir to several B-and T-cell lines. However, this is not a confounder to our study. We are aware of this study, and in our laboratory we have also observed several fibroblast and CHO cell lines to be sensitive to PIs (ritonavir in particular). However by contrast, primary hepatocytes and hepatocyte lines (HepG2 and McA7777, as used in our study) displayed no loss in viability as judged by Trypan blue exclusion assays, at the concentrations used in our experiments. This presumably reflects detoxification reactions prevalent in hepatic systems.