Determination of unbound platinum concentrations in human plasma using ultrafiltration and precipitation methods

Quantification of the unbound portion of platinum (Pt) in human plasma is important for assessing the pharmacokinetics of the chemotherapeutic drug cisplatin. In this study, we sought to compare the recovery of unbound Pt using Nanosep (R) filters to 1) traditional filters (Centrifree (R), Centrisart (R), Amicon (R)) or trichloroacetic acid (TCA) protein precipitation, and 2) unbound, bound, and total Pt concentrations in clinical specimens. For the tested filters, the impact of 1) molecular weight cut-offs, 2) centrifugation force, and 3) total Pt concentration on Pt binding in human plasma was evaluated. Pt was quantified using inductively coupled-plasma mass spectrometry. In human plasma spiked with 0.9 mu g/mL Pt, the percent of unbound Pt increased at higher centrifugation speeds. By comparison, the percent of unbound Pt was highest (42.1%) following TCA protein precipitation. When total Pt was <= 0.9 mu g/mL, unbound Pt (similar to 20-30%) was consistent across filters. Conversely, when plasma was spiked with Pt exceeding 0.9 mu g/mL, the percent of unbound Pt increased from 36.5 to 48% using ultrafiltration, compared to 63.4% to 79% with TCA precipitation. In patients receiving cisplatin-containing chemotherapy, the fraction of unbound Pt at concentrations exceeding 0.9 mu g/mL ranged between 35 and 90%. Moreover, the unbound fraction of Pt in plasma correlated with the concentration of unbound (R-2 = 0.738) and total Pt (R-2 = 0.335). In summary, this study demonstrates that 1) the percent of unbound Pt is influenced by total and unbound Pt levels in vitro and in clinical specimens, and 2) ultrafiltration with Nanosep (R) filters is a feasible method for quantifying unbound Pt concentrations in human plasma.