Alessandra Livraghi wrote:
> Here I am again. Some time ago I was asking for some histochemical method to
> label mitochondria in  human tissues...I got some good advices and the
> staining with SDH and NBT works pretty well.
> Now I'm back on the same issue, but I'm looking for some GOOD advices about
> antibodies to stain mitochondria in paraffin embedded and/or frozen sections

The very brief (4 min) formaldehyde treatment of your 
cryosections will not fix anything, and the 100% alcohol
that follows will wreck the organelles. Mitochondria are 
fragile little things, destroyed or deformed by some fixatives 
and destroyed or deformed by water etc if you don't fix them.

Dehydrogenase histochemistry on unfixed or almost unfixed
sections certainly shows mitochondrial enzyme activity, but
the sites of deposition of the stain do not necessarily 
correspond to the mitochondria that you'd see in an electron
micrograph or in a cell vitally stained with janus green B or
one of the modern vital fluorochromes, or with a classical LM
stain for properly fixed mitochondria. These statements apply
equally to immunostained mitochondrial antigens in inadequately
fixed cells.

The best fixatives for mitochondria in light microscopy are
archaic mixtures containing potassium dichromate and osmium
tetroxide applied to minute fragments of tissue, and thoroughly
washed out before processing into paraffin. Manfred Gabe, who
was an expert, wrote in his 1976 book that the clearing agent
should be cedarwood oil (which is viscous and not all removed
during infiltration with wax). I don't know what all this does
to the antigens; probably noone has tried to find out.

A neoclassical approach is to fix in neutral formaldehyde (added
calcium ions improve the picture) or glutaraldehyde (or a mixture 
of the two), and post-chrome the specimen. After some days in an 
aqueous dichromate solution the hydrophilic phospholipids (those 
of mitochondria and myelin) are rendered insoluble in organic 
solvents and therefore stainable in paraffin sections. 

I have used Bensley's copper-haematoxylin and Baker's acid 
haematein for this purpose, and I recommend the latter because
it's easy to do and is a thoroughly investigated and fairly 
well understood histochemical method. 

I've never tried Altmann's stain, but have been told that the 
bright red mitochondria on a pale yellow background are 
impressive. Baker's acid haematein gives blue mitochondria 
with a dirty-yellow background. The  yellowy background is 
due to haematein acting as an anionic dye in its own right, 
without any "mordant" metal ions in the staining solution. The 
blue color is that of the complex of haematein with the bound 
chromium in the mitochondrial lipids. It's much the same 
colour as the aluminium-haematein complex in cell nuclei in 
sections stained with H & E (or is it H&E as a single wordlet?).

There's plenty of published work in the field of mitochondrial
staining. If you have access to a library I'll be happy to 
provide references to books, peer-reviewed original papers and 
review articles in this field.

-- 
-------------------------
John A. Kiernan
Department of Anatomy and Cell Biology
The University of Western Ontario
London,   Canada   N6A 5C1
   kiernan@uwo.ca
   http://publish.uwo.ca/~jkiernan/