The clinical effectiveness of opioid maintenance for heroin dependence is believed to result from a medication's ability to decrease mu-opioid receptor (muOR) availability thereby replacing agonist effects, alleviating withdrawal symptoms and attenuating heroin effects. We empirically tested this hypothesis in five heroin-dependent volunteers who were successively maintained on 32, 16, 2, and 0 mg daily buprenorphine (BUP) tablet doses. We predicted and confirmed that higher BUP doses would decrease in vivo muOR availability (measured with PET and [(11)C]carfentanil), increase plasma levels of BUP and its metabolite nor-BUP, and decrease withdrawal symptoms and hydromorphone (HYD) responses. Relative to placebo, BUP significantly decreased mean (+/-SEM) whole-brain muOR availability 41+/-8, 80+/-2, and 84+/-2% at 2, 16, and 32 mg, respectively. Regions of interest (ROIs) (prefrontal cortex, anterior cingulate, thalamus, amygdala, nucleus accumbens, caudate) showed similar dose-dependent effects. Changes in muOR availability varied across ROIs (prefrontal cortex, 47% vs amygdala, 27%) at BUP 2 mg, but were more homogeneous across ROIs at BUP 32 mg (94-98%; except thalamus, 88%). Relative to placebo (0 ng/ml), peak plasma levels of BUP and nor-BUP were comparable and dose-dependent (0.5-1, 5-6, and 13-14 ng/ml at 2, 16, and 32 mg, respectively). muOR availability decreases were negatively correlated with BUP plasma level and positively correlated with questionnaire-based opioid withdrawal symptoms and attenuation of HYD symptoms. These findings suggest that high-dose BUP maintenance produces near-maximal muOR occupation, muOR availability correlates well with plasma levels, and BUP-related opioid symptoms and antagonist blockade exhibit concentration-effect relationships.