Neuronal tubulin is the standard in the field for performing microtubule polymerization assays. One use of neuronal tubulin concerns screening for tubulin ligands which have anti-tumor activity. Neuronal tubulin is ideal for preliminary screens where a large number of compounds have to be screened for initial tubulin binding activity. However there has been poor correlation between IC50 values determined from dose response curves on neuronal tubulin versus tissue culture or patient studies. This is due to several reasons including blood brain barrier diffusion, neurotoxicity, resistant phenotypes and possibly differential tubulin isotype expression. It is known that the tubulin isotypes present in neuronal tissue differ from those in cancer cells. Bovine neuronal tubulin has mainly beta II (58%) and beta III (25%) tubulins (Banjeree and Luduena, 1992) in combination with alpha-tubulin to make the typical heterodimer, this is in contrast to HeLa cells which have mainly beta I (90%) and beta IV (9%), and MCF-7 cells which have mainly beta I (50%) and beta IV (39%). Using tubulin isolated from brain tissue, HeLa and MCF-7 cells, we developed a micro-tubulin polymerization assay that is suitable for measuring the IC50’s of compounds on small quantities of cancer cell tubulin. The assay was highly dependent on protein concentration, for example the IC50 for vinblastine was five fold higher when 2.5mg/ml tubulin was used compared to 1.5mg/ml. Establishing conditions to reproducibly measure drug activity is the key to this technology. Using IC50 measurements on neuronal and cancer cell tubulins we compared them by dividing the neuronal IC50 by the cancer cell IC50, this created a Tubulin Ligand Index (TLI) which could be used to compare different drugs. Using this parameter we measured the activity of known drugs and their analogs in order to find which site on tubulin showed the most potential for increasing activity targeted toward cancer cell tubulin. Paclitaxel and its analogs showed more variation in their TLI values (10-20 fold), compared to either vinblastine (1-3 fold) or colchicine analogs (1-3 fold). We are comparing this activity with tissue culture studies to determine whether tissue culture is a true representation of tubulin ligand anti-cancer activity. There is room for improving current anti-cancer compounds using this assay so that the difference between cancer cell and neuronal tubulin specificity is closer to 100 fold. Hopefully in the future this will result in greater anti-tumor specificity and lower side effects.

[Proc Amer Assoc Cancer Res, Volume 46, 2005]