Dear Netpersons

A few weeks ago there was a querry in re patchy keratin signal.  I
cleverly deleted the message and don't know with whom I was communicating.
Still, I have a question for general consideration:

I, too, am seeing uneven keratin signal.

I am using rat olfactory epithelium (positive control).  The animals are
perfused with 4% buffered paraformalin.  The heads are cleaned of skin,
muscle and mandible then postfixed by immersion overnight at 4 deg. C.
The turbinates and associated bone are decalcified in three 1 liter
changes of 0.25 M EDTA over two days.  The EDTA is removed by washing the
tissue in three changes of PBS pH 7.4 over 8 hours.  The tissue is held in
70% EtOH until processing in a Tissue Tek VIP.  The entire processing time
is 4.5 hours.  The dehydrating agent is Fisher's Histo Grade Isopropanol,
the Xylene as well.  The paraplast is run at 59 deg C.  I've minimized
treatment times esp. in the xylene and paraplast to 40 min and 60 min,
respectively.  The tissue cuts well without apparent processing artifacts
(e.g. venetian blind effect).

This tissue has in the past responded well to my small battery of
antibodies (positive controls) but in the last 6 months has been giving
patchy results.  

Over the last two weeks, I've run three assays all of which gave similar
results.  The conditions and results of one are:

I have brand new DAKO wide spectrum polyclonal anti-cytokeratin antibody;

The pattern is the same whether I use two different FITC conjugated
secondaries (one is brand new) or a biotinylated secondary with ABC
incubation and DAB
coloring;

The pattern is the same with two batches of tissue:  one of which is a
month old the other of which is 2 years old.

A monoclonal anti-NCAM (neural cell adhesion molecule) (Sigma) works
almost but not quite fine on the two tissue batches.  The anti-NCAM should
and does light up the receptor neurons in the sensory epithelium and axon
bundles in the lamina propria of the two year old tissue sample but lights
up ONLY the axon bundles in the new tissue sample.

The NCAM results suggest the problem is in the processing.

The cytokeratin results suggest the problem may be in the processing but
not
necessarily the same problem suggested by the NCAM results.  

Here are my thoughts:

The cytokeratin signal pattern relects a difference in lots of the
antibody (it used to never fail)

The processing reagents are of a different and lower quality than the same
label reagents available in the past.

Something is wrong with the fixation.  Specifically, the fixitive was too
old.

All ideas will be greatly appreciated.

Yours against the wall

John Carroll Dennis
Anatomy, Physiology, and Pharmacology
109 Greene Hall
Auburn University, AL  36849