SPK-241

The spicamycin analogue KRN5500 alters glycoprotein processing and induces damage in the endoplasmic reticulum (ER)‐Golgi apparatus in cancer cells. In the present study, we explored the cytotoxic effects of KRN5500 on multiple myeloma (MM) cells and the bone marrow microenvironment with special reference to ER stress. Cell proliferation assay showed that KRN5500 induced G1 arrest and apoptosis in MM cells in a time‐ and dose‐dependent manner. KRN5500 enhanced ER stress independently of caspase activation in MM cells. This cell death was observed even in the presence of bone marrow stroma cells or osteoclasts. Notably, KRN5500 induced cell death also in osteoclasts. In vivo effects of KRN5500 were evaluated using two xenograft models established in severe combined immunodeficient (SCID) mice by either subcutaneous injection of RPMI 8226 cells or intra‐bone injection of INA‐6 cells to subcutaneously implanted rabbit bones (SCID‐rab model). KRN5500 significantly inhibited tumour growth in both animal models, and decreased the number of osteoclasts, which resulted in prevention of bone destruction in the SCID‐rab model. These results suggest that KRN5500 exerts anti‐MM effects through impairing both MM cells and osteoclasts. Therefore, this unique mechanism of KRN5500 might be a useful therapeutic option in patients with MM.

One tenth of the lethal dose to 10% of mice is one of the conventional parameters used to derive a safe starting dose in phase I trials of cytotoxic agents. There is no consensus on which preclinical models and parameters should define the starting dose for molecularly targeted agents.

Reports of 81 first-in-human phase I trials evaluating 60 different molecularly targeted agents administered as monotherapy were reviewed. The maximum-tolerated dose (MTD) was defined as the highest safe dose administered to patients, whereas the maximum-administered dose (MAD) was recorded if the MTD was not reached.

Fifty-seven of the 81 trials specified the animal model used to determine the starting dose, with 29 (51%) of 57 based on rodent data and 28 (49%) of 57 based on non-rodent data. A wide range of toxicologic parameters was used to select the starting dose. The starting dose exceeded the human MTD in three (3.7%) of 81 trials, and in all three trials, nonhematologic toxicity was dose limiting. The median number of dose levels to reach MTD or MAD from starting dose was five (range, one to 14 dose levels), and the median ratio of MTD or MAD to starting dose was 12 (range, < 1 to 300). Hypothetical doubling of the starting dose appeared to be safe, whereas tripling of the starting dose was unsafe.

The derivation of starting dose for first-in-human phase I trials of molecularly targeted agents in patients with cancer is safe but is based on diverse practices using a variety of preclinical toxicologic parameters.