Pharmacokinetics of methotrexate in red blood cells

Background & Objectives:
Low-dose methotrexate (MTX) given once weekly is the gold standard in the therapy of rheumatoid arthritis (RA). However, MTX doses required to achieve adequate disease control are highly variable between patients and difficult to predict. MTX polyglutamate (MTXPG) concentrations inside red blood cells (RBCs) have been discussed as a potential biomeasure predicting pharmacodynamic (PD) outcomes in RA [1].
Our objective was to describe the population pharmacokinetics (PK) of MTX and its polyglutamated metabolites in RBCs as a first step towards a full pharmacokinetic-pharmacodynamic (PKPD) model for MTX in RA.

Materials & Methods:
Individual RBC concentrations for MTX and MTXPGs were available from 48 patients from two studies [2,3]. No plasma MTX/MTXPG data were available from these patients. A previously published 2-compartment model for the plasma PK of MTX [4] was used to predict plasma MTX concentrations in these patients. Additional compartments for MTX and MTXPG were added to represent the RBC distribution and metabolism sites where up to four additional glutamate moieties are added to the parent drug. Therefore, five catenary RBC compartments (one for the parent and one for each polyglutamated metabolite) were required in total. Covariates tested on the apparent volume of distribution inside RBCs (VRBCs) included total body weight, lean body weight (LBW), blood haemoglobin concentrations ([Hb]) and mean corpuscular volume (MCV). Residual unexplained variability was modelled as a combined error.

For the parent model, LBW, [Hb] and MCV were all significant covariates on VRBCs. Between-subject variability (BSV) on the clearance of MTX from RBCs (CLRBCs) was not supported in the final model for MTX parent alone. The final parameter estimates (population mean (BSV)) were: kin = 1.35 x 10-4 hr-1 (63.7%), CLRBCs = 7.05 x 10-4 L/hr (-), VRBCs = 0.287 L (110%), CVprop = 20.3% and SDadd = 3.58 nmol/L.
Results for the parent-metabolite model for MTXPGs are pending.

RBC concentrations of MTXPGs are a tempting biomeasure for MTX monitoring due to their relative ease of access. Whether RBC MTXPG concentrations truly have predictive performance for disease control due to an indirect association with PD outcomes needs to be tested in future work, ideally based on a population PKPD model. This work is considered to be the first step towards such a model.

1. Dervieux T, et al. (2004). Arthritis & Rheumatism 50(9):2766-2774
2. Dalrymple J, et al. (2008). Arthritis & Rheumatism 58(11):3299-3308
3. Stamp L, et al. (2011). The Journal of Rheumatology (Online first)
4. Hoekstra M, et al. (2004). The Journal of Rheumatology 31(4):645-648