Easily accessible normal tissues expressing the same molecular site(s) of drug action as malignant tissue offer an enhanced potential for early proof of anticancer drug mechanism and estimation of the biologically effective dose. Studies were undertaken in healthy male volunteers to assess the tolerability of single and multiple (four in 24 h) 3 mm punch biopsies of the buccal mucosa, and to determine the feasibility of detecting and quantifying a range of proliferation, cell-cycle arrest and apoptosis markers by immunohistochemistry (IHC) for use as potential pharmacodynamic (PD) end points. The biopsy procedure was well tolerated with 100% of volunteers stating that they would undergo single (n = 10) and multiple (n = 12) biopsies again. Total retinoblastoma protein (pRb), phosphorylated pRb (phospho-pRb), total p27, phosphorylated p27 (phospho-p27), phosphorylated-histone H3 (phospho-HH3), p21, p53, Cyclin A, Cyclin E, Ki67 all produced good signal detection, but M30, cleaved caspase 3 and terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labelling did not. Total pRb, phospho-pRb, total p27 and phospho-p27 were quantified further in a multiple biopsy study to allow components of variability to be addressed to inform future sizing decisions on intervention studies. Neither site of biopsy within the oral cavity, nor the nominal time of biopsy had any significant impact on any of the four markers expression levels. Inter- and intrasubject coefficients of variation (CVs) that could be used to size future intervention studies for pRb, phospho-pRb, total p27 and phospho-p27 were 14, 19, 18 and 16%; and 18, 29, 25 and 19%, respectively. In conclusion, quantitation of such markers in 3 mm buccal punch biopsies would be suitable to explore as PD end points within intervention studies of drugs acting on these pathways.