Effect of GSK3beta inhibition on neurite initiation

To test the prediction potential of the proposed model we took an interest in glycogen synthase kinase (GSK) - 3β, a serine/threonine kinase that regulates a number of diverse cellular processes like glycogen metabolism, cell-fate determination and hormone signalling, just to mention few. Unlike many kinases, GSK3β is constitutively active in cells and it is mainly regulated by inactivation (Cohen and Frame 2001). According to our model the inactivation of GSK3β is one of the key events during neurite initiation process and there are several pathways in the model that lead to GSK3β inactivation. If the model is correct than "overinhibition" of GSK3β should lead to uncontrolled sprouting of many new neurites.

Li ion is well known inhibitor of GSK3β and is commonly used in these types of studies (Stambolic, Ruel et al. 1996). Li ion effect on neurite outgrowth has already been tested by other researches and it is known that Li ion inhibits neurite outgrowth in PC12 cells (Burstein, Seeley et al. 1985). We were able to confirm that result which does not speak in favour of our proposed model..




24 hours

NGF + 8mM LiCl


24 hours


However, using time lapse microscopy, we noticed that the cells treated with LiCl develop a large number of small neurites in the first 4-8 hours of treatment which are unable to grow very far. During following hours they undergo regression until the cell is finally completely devoid of neurites.

video showing NGF induced neurite outgrowth

260x332 pixels
3.5 Mb

video showing NGF induced neurite outgrowth in the presence of 8mM LiCl

280x212 pixels
4.0 Mb

Although Li ion is very good inhibitor of GSK3β, it also moderately inhibits several other kinases (Davies, Reddy et al. 2000) and it is a very good inhibitor of several phosphatases involved in the inositol and glucose metabolism (Quiroz, Gould et al. 2004). Deregulation of inositol metabolism by Li ion could be the reason why we see a regression of newly formed neurites and ultimate inhibition of neurite outgrowth. Because of this we modified the treatment and washed away LiCl after 4 hours after which we left the cells for another 24 hours to differentiate. Under these conditions we observed a major increase in the number of neurites per cell. The percentage of cells bearing 4 or more neurites increased from 11.5% +/- 2% in cells treated only with NGF to 32% +/- 6% in cells which were treated with NGF and 16mM LiCl. The increase in the number of cells bearing 4 or more neurites happened in dose dependent manner and it was accompanied by the corresponding drop in the percentage of cells bearing 1 or 2 neurites. In control experiments, treatment of cells with 16mM NaCl had no effect on the number of neurites.


Because of the concerns about the specificity of Li ion and to see if we can observe the same cellular response with another, structurally unrelated inhibitor, we performed similar experiments with CHIR99021. CHIR99021 is the most specific inhibitor of GSK3β available today. It exhibits from 500 fold to more than 10000 fold selectivity for GSK3β versus other protein kinases and it's Ki is in nanomolar range (Ring, Johnson et al. 2003). We observed very similar cellular response as with LiCl treatment. The percentage of cells bearing 4 or more neurites increased from 12% +/- 1% in cells treated only with NGF to 55% +/- 2% in cells which were treated with NGF and 20 ÁM CHIR99021. It was not necessary to wash CHIR99021 away to get the effect, although similar effects are observed even if the compound was washed off after 4 hours (data not shown). These results are, indeed, in agreement with what one could predict from the proposed model and they speak in favour of the GSK3β involvement in neurite initiation in PC12 cells.


PC12 cells morpholgy after 24h treatment with NGF


PC12 cells morpholgy after 24h treatment with NGF in the presence of 20 mM GSK-3β inhibitor CHIR 99021