After cutting, let the sections dry overnight at room temperature under a ventilator.

Fix with cold acetone (10 min, 4C). Some people apply a "double acetone fixation" method: 2x10 minutes acetone fixation with air-drying in between. This may improve the tissue morphology.
 

The quality of the acetone should be P.A. grade and don't use it twice. Re-distilled acetone cannot be used for fixation of cryo's. Traces of water in the acetone ruins the tissue morphology.
 

Be aware that acetone is not a real fixative like NBF. Acetone just solves the fatty membranes and coagulate the proteins. Cryostat tissue sections remain quite vulnerable with respect to surfactants like Triton-X100, Tween-20. These should be avoided. Be aware that in some autostainer wash buffers surfactants may be included.
 

To solve this problem of tissue vulnerability an extra fixation (after acetone-fixation and air-drying) with Zamboni's (1 min, RT) is optional. This extra fixation step also improves the tissue morphology. However, depending on the antigen to be IHC stained, this step may also decrease the IHC staining intensity.
 

The application of methanol, either as a fixative or as base for endogenous peroxidase activity blocking is (at least to my vision) absolutely excluded. For example, most human CD markers are completely destroyed by methanol. On the other hand I am aware of investigators who are using successfully an acetone/methanol mixture for fixation of mouse cryo's for staining CD4, CD8 etc.
 

Acetone is not a good fixative when staining nuclear antigens. This fixation will lead to quite fuzzy stained nuclei. Instead, a NBF-fixation (5 min, RT) works out well. Although NBF-fixation is used: do NOT apply heat-induced antigen retrieval.
 

Some antigens associated with fatty structures (for example, oxLDL, or apoB) will solve into acetone. In this case NBF (5 min) can be tested. Again, NO antigen retrieval!

Endogenous peroxidase activity can be blocked with 0.1% Na-azide + 0.3% peroxide in PBS or TBS (20 min, RT). This kills at least the endogenous peroxidase activity in erythrocytes effectively, but in neutrophils some will remain. If endogenous peroxidase activity of neutrophils is a real problem, glucose oxidase blocking method is optional (see Histonet archives).
 

At the end of the IHC staining procedure, after the chromogen step wash your slides with tap water. NEVER use distilled water as this will ruin the tissue section completely!!!!

Unlike with FFPE sections, aqueous solutions (depending on the chromogen used of course) is no problem for mounting cryostat tissue sections.
 

The concept of first cutting a tissue section and than fixation means "post-fixation". This is something different from a FFPE section that is first fixed as a block, than embedded and cut
("pre-fixation"). This means that soluble antigens may leak away from a "post-fixed" tissue section. Be aware of this when IHC staining any small protein (<50KD), cytokine, chemokine, hormone,
etc.

References:

1. McMillan EM, Martin D, Wasik R, Everett MA (1981) Demonstration in situ of "T" cells and "T" cell subsets in lichen planus using monoclonal antibodies. J Cutan Pathol. 8(3):228-34. PubMed Abstract

2. Matsumoto Y (1985) Simultaneous inhibition of endogenous avidin-binding activity and peroxidase applicable for the avidin-biotin system using monoclonal antibodies. Histochemistry. 1985;83(4):325-30. PubMed Abstract