Typically edge effect is caused by drying (tissue and/or reagent) or uneven reagent coverage. You have to be especially careful if you are heating reagents. (TUNEL usually requires heated Tdt enzyme). Do you see the same pattern with other stains on adjacent sections of this same tissue? Cause could be prior to staining?

Sorry just realized that this is positive control so should be staining all and not just edge. I would still investigate other stains and processing prior to TUNEL. When you see this pattern with other staining (IHC, IF etc) it usually indicates poor fixation or processing prior to staining. 

thanks for sharing!

We see this all the time in the liver. First, since this is your positive control, all of your nuclei should be as bright as the edge cells. I suggest treating the section in a way that allows the terminal transferase better access to the nucleus. This could involve limited proteinase K treatment or the use of citrate buffers and heat. There is another possibility for negative staining on the positive control slide, in that over-digestion of the DNA with DNase can lead to loss of the substrate for TdT. This is not your problem, since the edge cells would be most susceptible to over-digestion. Until all of your cells are brightly positive in the DNase-treated slide, your results with experimental slides won't be reliable because actual dead cells are not visualized if the enzyme can't permeate.

This still leaves the edge effect. In the liver, we simply move to another part of the slide that doesn't have an edge. I know in the brain this may not be an option because the region you study may be at the edge of a ventricle, etc. I agree with Dr. Burton that this may reflect pre-fixation drying. However, you not that this is perfusion-fixed brain, so this is less likely. You could try different fixatives (freshly made paraformaldehyde, with or without picric acid) or different freezing methods (dry ice versus CO2 or dipping into liq N2). The edge of the tissue may be more likely to have ice crystal formation, faster freezing should help. Is the artifact worse on the edge that hits the cutting blade first? You could re-orient your block for cutting.

Good luck, try out different protocols; TUNEL staining requires optimization, even if you've done it before.

Many thanks for all of your answers.

I haven't checked whether our perfusion is effective, but after perfusion, I put all dissected brains in 4% paraformaldhyde overnight at 4 centigrade. Say if there is truly perfusion problem which i cannot guarantte to avoid, is it possible that only fixative immersion causes (poor fixation) this effect?

If drying or post-mortem changes are the problem, then rapid fixation (perfusion) should limit these changes. Fixation by only immersion in PF is not as good. You can cut the brain in half before immersion-fixation to help with penetration and fixation, but I strongly encourage perfusion.

Do you have the same edge artifact without DNase treatment? If not, then your fixation protocol is probably not the source of trouble. (Fixation also maintains cellular structure, which is not as critical for TUNEL as it is for localization of a protein withint the neuron).

I had a similar issue with my positive controls - it's a permeabilisation problem. I resolved this issue by microwaving my slides in 0.1 M citrate buffer pH 6.0 to ~ 86 C and rapidly cooling by immediate transfer into PBS. If you're not as concerned with quantitative analysis, you could also try fixing in acetone with no permeabilisation.

Refer to "TUNEL: Improvement and Evaluation of the Method for In Situ Apoptotic Cell Identification" by Negoescu et al. publication for more information on optimizing permeabilisation.

Thank you Justin Mott!

I didn't get the same edge effect without DNase. It's hard to tell whether TUNEL can work in other groups, because actually no significant signal was detected and no one can be sure cell death really occurred except positive control group.

Hi Kaleb Short!

The microwave method seems very promising according to the reference you kindly gave. But did you solve the problem on cultured cell? Because we have observed normal staining on cell. Thanks.

Hao Chen,

The DNase positive control is very helpful in this case! If you have performed the unmasking and TUNEL staining the same in the positive control and experimental groups (except DNase of course), then this is good evidence of the absence of TUNEL-positive cells in the experimental groups.

One bit of caution. Did you find a way (maybe the citrate approach from Kaleb Short) to get the DNase to be positive throughout the section? If the DNase slides are only strongly positive at the edges, then your TUNEL staining is not yet reliable for the whole slide (for any of the groups). Optimize the DNase slide, then use exactly that procedure for the experimental slides.

To Justin Mott:

No unfortunately I haven't got good positive signal in positive control (DNase treatment) but only edge effect. That is why I am seeking help here. Because right now I'm not in SUSTC (south university of science and technology of china), distant from our lab, I plan to first collect some advice before I return. I am grateful for the caution.

Hao

Hao Chen - I've used the microwave/citrate permeabilisation protocol on mouse heart cryosections. 

Be sure Tritón X100 is fresh. You have to prepare permeabilization medium just after use it.