advertisement

Editors Selection IGR 9-3

Basic Research Animal Studies: Semiautomated method to measure ON axons

Comment by John Danias on:

27639 A semiautomated targeted sampling method to assess optic nerve axonal loss in a rat model of glaucoma, Marina N; Bull ND; Martin KR, Nature Protocols, 2010; 5: 1642-1651

Find related abstracts

Marina et al. (1617) describe in detail a semi-automated targeted sampling method to measure optic nerve axons. They also provide methodological details for inducing IOP elevation in albino rats using perilimbal application of laser. They use this glaucoma model to determine the accuracy of their quantitative method in comparison to a method employing random sampling and a method employing full manual counting of the whole optic nerve.

The authors are to be congratulated for their decision to provide their detailed protocols in print and to provide additional supplemental data (including a video of the application of the laser in rat eyes) that would allow anyone to replicate their experimental setup. Such transparency and detail in research methodology are rare (in part because of space restrictions in printed journals), but critically important in allowing others to replicate (and thus validate) experimental findings. The protocols provided are clear and concise and should be easy to follow in any lab with appropriate instrumentation. The laser-induced glaucoma model described is, according to the authors, an acute IOP-elevation model. The IOP graph they provide suggests a rapid high IOP elevation immediately after the laser application. The authors do not provide IOP data on days 2-7 after lasering but extrapolate assuming a slow decrease in IOP. Personal experience suggests that IOP decreases to normal levels much faster than implied by the graphs. The authors also report up to 50% axonal loss using this method, but fail to disclose the significant variability between individual eyes (although this variability is clearly evident if one inspects the axonal counts in table 1).

The semi-automated method described for counting axons is a middleof- the-road approach between the very labor-intensive 'manually count all axons' approach and the qualitative 5-point-scale approach. It seems to perform better than the random sampling method, even though it requires significantly less effort. However, it does have its own limitations as the authors readily admit: it may be insensitive to small amounts of damage, is dependent on uniform and good staining of the nerve, and is subjective as to the location and size of axonal damage. All of these factors affect reproducibility. According to the authors when they compared counting results from two different observers using the same samples and the same method the square of the correlation coefficient (r²) was only 0.69 (typically one wants an r² of at least 0.8). Despite this fact, the method described seems to be a reasonable approach in quantifying the amount of axonal loss in the laser induced glaucoma model that can be easily implemented using public domain software.

• ← Previous
• Next →