TABLE 1

Input parameters for the initial tizanidine PBPK model with the well stirred model

ParameterValueMethod/Reference
Molecular weight (g/mol)253.7PubChem
log Po:w1.84Average of 5 in silico estimates
Compound typeMonoprotic base
pKa7.46FDA review (NDA 21-447)
BP1.46Predicted with Simcyp V21R1
fu,p0.7(Shanker et al., 2009; Lombardo et al., 2018)
Main plasma binding proteinHuman Serum AlbuminAssumed
Absorption modelFirst order absorption model
Ptrans,0 (10−6 cm/s)249Predicted based on Log Po:w (Sugano, 2009)
Peff,human (10−4 cm/s)2.31Predicted using MechPeff model (Sugano, 2009; Pade et al., 2017)
fa0.94Predicted using Simcyp V21R1
ka (1/h)1.01Predicted using Simcyp V21R1
Lag time (h)0.5Reported upper limit of lag time for tizanidine tablets (FDA review, NDA 21-447)
fugut1Assumed
Qgut (L/h)10.88Predicted using Simcyp V21R1
Distribution modelFull PBPK model
VSS (L/kg)2.42Predicted by Method 2 (Rodgers and Rowland, 2006) with a Kp scalar (0.28) to recover clinically observed Vss
CYP1A2 CLint (μl/min/pmol)6.76Optimized based on clinical studya
Renal Clearance (L/h)3.4Clinical observations
  • BP, blood-to-plasma ratio; fa, fraction absorbed; fugut, unbound fraction in gut enterocyte; fu,p, unbound fraction in human plasma protein; Kp, tissue: plasma partition coefficient; ka, first order absorption rate constant; Peff,human, effective permeability in human jejunum; Po:w, neutral species octanol: buffer partition coefficient; Ptrans,0, intrinsic transcellular passive permeability; pKa, the negative base-10 logarithm of the acid dissociation constant; Qgut, a nominal flow in the gut model; Vss, volume of distribution at steady state.

  • aFor the tizanidine-ciprofloxacin DDI, study specific CYP1A2 CLint of 11.4 μl/min/pmol was used in the simulation to recover the observed tizanidine AUC in the control arm of the clinical study. The simulations for DDIs with the other perpetrators were performed using the global (default) CYP1A2 CLint of 6.76 μl/min/pmol.