Arsenic is a proven human carcinogen and is associated with a myriad of other adverse health effects. This metalloid is methylated in human liver to monomethylarsonic acid (MMA(V)), monomethylarsonous acid (MMA(III)), dimethylarsinic acid (DMA(V)), and dimethylarsinous acid (DMA(III)) and eliminated predominantly in urine. Hepatic basolateral transport of arsenic species is ultimately critical for urinary elimination; however, these pathways are not fully elucidated in humans. A potentially important human hepatic basolateral transporter is the ATP-binding cassette (ABC) transporter multidrug resistance protein 4 (MRP4/ABCC4) that in vitro is a high-affinity transporter of DMA(V) and the diglutathione conjugate of MMA(III) [MMA(GS)(2)]. In rats, the related canalicular transporter Mrp2/Abcc2 is required for biliary excretion of arsenic as As(GS)(3) and MMA(GS)(2). The current study used sandwich cultured human hepatocytes (SCHH) as a physiological model of human arsenic hepatobiliary transport. Arsenic efflux was detected only across the basolateral membrane for 9 out of 14 SCHH preparations, 5 had both basolateral and canalicular efflux. Basolateral transport of arsenic was temperature- and GSH-dependent and inhibited by the MRP inhibitor MK-571. Canalicular efflux was completely lost after GSH depletion suggesting MRP2-dependence. Treatment of SCHH with As(III) (0.1-1 µM) dose-dependently increased MRP2 and MRP4 levels, but not MRP1, MRP6, or aquaglyceroporin 9. Treatment of SCHH with oltipraz (Nrf2 activator) increased MRP4 levels and basolateral efflux of arsenic. In contrast, oltipraz increased MRP2 levels without increasing biliary excretion. These results suggest arsenic basolateral transport prevails over biliary excretion and is mediated at least in part by MRPs, most likely including MRP4.