On the 3rd February 1999, Dr Naseem (1) wrote: "Can electron microscopy be done on formalin fixed paraffin tissue?"
It seems that EM results on paraffin processed tissue will only be as good as the processing procedure used by the histology lab. In Robert Sanoianni's (2) experience most histology labs use a very aggressive protocol, which causes extraction of cellular components to the point where only nuclei and possibly cell membranes might be recognizable. Laboratories that handle large volumes and are concerned with quick turnaround generally dehydrate and clear the heck out of specimens so they don't have to go back and reprocess if something was too big, not well fixed, etc. Robert's laboratory provides good fixation, gentle dehydration, clearing and infiltration of routine histology specimens so that when he has to do EM on one of these, there is usually enough cellular matrix left to make the effort worthwhile.
This sentiment is mirrored by Robert Chiovetti (3) who also found that it all depends on the quality of processing during the initial paraffin embedding, which is usually quite rapid. If this is the case, you will likely only see remains of membranes, cellular outlines and nuclear "ghosts," and not much intracellular detail. Extracellular matrix materials may survive the process well enough to distinguish collagen, elastin, etc.
Gayle Jensen (4) says that the quality of the micrographs depends upon the initial fixation of the tissue. She routinely uses Carson formalin for both light and EM.
Philip Oshel (5) suggests you use well-fixed, nicely treated specimens, not the rapid dothisfast! specimens.
Rob Wadley (6) reminds us that the processing of the tissue to paraffin extracts most if not all the lipids in the tissue, so the image generally looks much more open/clear than if the tissue was properly prepared for EM examination.
In spite of the destructive affects that paraffin processing can cause, some ultrastructural features can still be seen. Tim Morken (7) notes that in most cases it is possible to distinguish lymphomas from carcinomas (lack of or presence of desmosomes), finding neurosecretory granules in carcinoids and pre-melanosomes in melanomas. It is really only useful as a "Hail Mary" attempt to get information after the specimen was wrongly submitted and all other tests fail to give answers.
Robert Santoianni (2) has been able to resolve Birbeck bodies, tonofilaments, junctions, microvilli and immune complex deposits (in glomeruli) in paraffin processed specimens.
Philip Oshel (5) recommends that the sample for reembedding be taken from an area within one millimeter of the original block face to give yourself the best chance of decent morphology. Any deeper, and you'll be beyond the recommended specimen size for good EM (1mm max. thickness).
Gayle Jensen (4) cuts the tissue into small blocks and melt them in a 60 degree oven. Then she does "a real no, no" and puts the tissue into a stoppered glass vial with xylene and puts it back into the oven for 1/2 hour. (This has worked for 28 years.) She changes the xylene at least twice. She then rehydrates the tissue beginning with a couple of short changes of ETOH, followed by a few rinses of 80% ETOH, 50% ETOH and distilled water. She then embeds the tissue as usual on a Lynx Processor.
Bob Chiovetti (3) recommends the following:
Tim Morken (7) suggests that you can skip all the reprocessing from paraffin to EM resin by using 1 percent osmium in xylene to deparaffinize. This was pioneered by Kai Chien of Cedars-Sinai in L.A.
Gayle Jensen (4) on the other hand has tried mixing xylene and osmium to eliminate the rehydration step but had no success. She found the tissue to be full of holes.