INTRODUCTION: Differences between tibial and femoral joint surfaces and knee compartments concerning coupled bone and cartilage turnover or bone-cartilage cross talk have not been previously examined, although the mechanical and biological interaction of the mineralized subchondral tissues with articular cartilage is of great importance for advancing osteoarthritis. MATERIALS AND METHODS: Therefore, with the help of immunohistochemistry and real-time polymerase chain reaction (RT-PCR), human knee joint cartilage tissue was investigated for expression of key molecules of the extracellular matrix and cartilage composition (collagen type I and II, aggrecan) plus proteoglycan content (colorimetric analysis). Furthermore, we correlated the results with 3D microcomputed tomography of the underlying subchondral bone (high-resolution micro-CT system). Measurements were performed in dependence of the anatomical site (femoral vs. tibial, medial and lateral each) to identify regional differences during the osteoarthritic process. From an enduring series of 108 patients undergoing implantation of TKA, 34 osteochondral samples with lesions macroscopically classified as ICRS grade 1b (group A) and 34 samples with ICRS grade 3a/3b lesions (group B) were compared with 21 healthy controls. RESULTS: Concerning 3D analysis, the medial femoral condyle and tibia showed the most significant increase in bone volume fraction and a decrease in the trabecular number in group B frequently accompanied by subchondral bone resorption pits and enchondral ossification. Under physiological conditions, tibia plateaus show lower bone volume fraction than the corresponding femoral site and this difference enlarges with advancing OA. Partially even contradictory behavior was observed such as trabecular separation at the lateral tibial and medial plateau in osteochondral OA samples of the same patients. Collagen type II expression levels show faster and varying changes than type I during the OA process, leading to a lower positive or negative correlation with bone microstructural analysis, especially on the tibia plateau. CONCLUSIONS: Structural bone and cartilage parameter changes showed varying developments and correlations among each other in the different compartments of the knee. As a clinical conclusion, therapies to postpone or prevent cartilage degeneration by influencing the loss of mineralized bone could be site dependent.