Hi Mikael,

Personally, I would use cryosections, quicker and much less problematic processing.

But if you are instistant on parafin embedded material you could try using 3% Glutaraldehyde with the addition of Ruthenium Red or Alcian Blue in your fixation process, these are typically used in SEM but should work equally well in parafin embedding. The Ruthenium Red and Alcian Blue help the preservation of the cell wall during fixation and so might aid in preservation of bacterial cell shape in your samples, they also act as fixation agents in mucus secretions and so may also aid in a higher bacterial yeild preventing wash out of your samples.

Best Wishes

Adam.

I routinely examine FFPE tissues, and have never noticed an issue with bacteria becoming rounded. I would see a few possible causes for what you are seeing.  Are you using a isotonic buffer to make up your fixative solutions ? Have you considered that the "roughly round" shapes you are seeing might be transverse sections through rod-shaped bacteria ? Could the gut flora of your animals be predominantly cocci ?

Thanks for the comments! Yes the PFA is PBS buffered (Carnoy obviously not because it is water-free). The 2D sectioning doesn't explain the shape, as these are 4 um sections and I see absolutely no rod-shaped cells. This is mammalian intestine which is to my knowledge not supposed to be so predominantly cocci. 

Adam, how do you treat your cryosections before staining? I guess you use SOME kind of fixation there? Glutaraldehyde might be worth trying too, although I really don't understand why the bacteria would be affected by PFA while the mammalian cells are just fine!

Mikael,

Just one comment: bacteria usually can be seen (better, if applying some specific agents [e.g. dyes, as pointed out by Adam and there are for sure more techniques for the task out in the wild], but - due to their vital size (usually said to be 0.5 - 3 µm in diameter , length up to 5µm) they are visualized at the lower end of resolution capability of LM (at least you have to use objective(s) x50 Oil or x100 Oil). To think over the "roundish" shape () left after your fixation is like solving a crossword puzzle as long as we do not know / be informed about genus/species you are talking about. As of my knowledge many bacteria posess  "only" circular shape when they are not clustered (e. g. as they appear in vessels during septicaemia/DIC, cf. attached image:  compare size of  leaking erythrocytes [bluish-tourquoise within collagen fibril meshwork = red] with bacteria (ok, these ARE cocci, measured by EM to be in a roundish, not perfect circular shape with a size of around 0.60µm up to 1.2 µm)

So if you your sample sections are  at 4µm always keep in mind hampered LM-  too.

There are solutions for your matter, for sure (if I have the time I shall come back again with at least one). Best wishes and regards, Wolfgang

PS: would you mind to provide a / some representative image(s) out of your recent collection...so we could have an imagination of your problem?? Thank you in advance!

Thanks Wolfgang! Please see a few examples attached. These are Gram stained bovine large intestine samples, photographed through a 100X oil immersion objective. I'm not after a specific species or even group of bacteria at this stage, so this is the normal intestinal flora. I compared these to mouse intestines which were perfusion fixed with PFA, and there the bacteria were much more beautifully visible, appearing much larger and often clearly rod shaped. This is basically why I'm thinking there is something wrong with my protocol.

Is this effect only seen in Gram stain or have you tried other stains also? For example HE, Giemsa or Warthin-Strarry?

Would be interesting, if this phenomen only appears with Gram.

Personally I don't see a problem with these. I can see small cocci, larger cocci, short rods or coccobacilli... The fact that you are not seeing gram-positive rods in these particular sections might just be due to sampling variability. Mouse intestinal flora will be different from bovine, so you can't expect to see the same organisms in the same numbers. Bovine flora is known to be rich in cocci (streptococcus, enterococcus, etc...), so it's not surprising that these are most apparent.

Thanks Jean-Martin, I wasn't thinking they could be THAT different! Gudrun, yes I could try the other stains too, and I'll be doing 16S FISH as well.

Dear Mikael,

thank you really for providing the images.

I "see" the same as Jean-Martin Lapointe has said.

 As one point of lacking information left at this stage:

you stated in your Re# 03 but sent images from bovine large intestine samples (I guess fixed by immersion, material from slaughter house? what was the delay in initial fixation, how big the size of specimen blocks to be fixed....etc.). If you compare (correctly) perfusion fixed material with immersion fixed material, mostly always you'll find the "morphological preservation of tissue and tissue contents" better in the former than in the latter. And just one hint in addition to my attached image in Re # 04: if you have a close look to the intraluminal meningococci (the type of infection was tested / found by bacterial culture, parts of which also were processed for TEM in this interesting case out of 5 or 6 patients suffering from purpura fulminans) you'll see that there is a totally unstained "ring like" space around the bacteria. Because I was at bedside when the sample was taken from the patient I know that the primary fixation “started” directly after punch excision, transferring the biopsy a a whole within 10 seconds into the fixative (which nevertheless only means "classical" immersion fixation). As correlative TEM on this particular area was done I can tell you that the "empty spaces" are not an optical "fata morgana" or "just an artifact" of fixation... instead I found that the empty space is a consequence of lacking "material" around the bacteria due to failing retention of this particular material by the fixative. Using other (more "sophisticated", i. e. specially composed) fixative(s) I could find retained (electron dense) material (though condensed but present) which might correspond to some structural properties of bacteria, namely "outer membrane vesicles" said to be parts of bacterial secretion system(s), i.e. release of (endo-)toxins.

Unfortunately such morphological aspects you can reveal only by ancillary TEM-methods (cf. also my Conference paper/Poster, saved to my publication profile: Purpura fulminans: ultrastructural findings by transmission electron microscopy in four cases, with special reference to fixation of skin biopsies, @ https://www.researchgate.net/publication/215824406_Purpura_fulminans_ultrastructural_findings_by_transmission_electron_microscopy_in_four_cases_with_special_reference_to_fixation_of_skin_biopsies?ev=prf_pub)

So IMHO an optimal initial fixation will influence (positively) certainly the outcome of your study (also in terms of bacterias’ morphological presentation after staining).

Naturally I second also the proposal of Gudrun to try different staining techniques, including the Warthin-Starry-method as well as others.

Best wishes and good luck, Wolfgang

Poster Purpura fulminans: ultrastructural findings by transmission ...

Thanks for the very interesting discussion! This is slaughterhouse material, with fixation starting a few hours after death. I know this is not optimal, but that's what we have to work with. I somehow thought the bacteria would be just fine, although the intestinal epithelium of course starts to degrade immediately.

To preserve the intestinal contents during the processing, I cut a piece of the intestinal wall about 2 cm x 1 cm, fold it like a piece of paper so that the intestinal contents remain in between the intestinal wall, and pack that quite tightly into a standard histological cassette for the period of fixation. There is perhaps 2 mm thickness of tissue from the surface to the intestinal contents between the folded sheets. So it's not VERY thick for the fixative to penetrate.

Dear Mikael,

thank you for your comment on specs, spec.prep. and initial methods.

Yes, I find this discussion also interesting (have dealt with a lot of such in past times on other RG-threads). Due to the retarded initial fixation ("a few hours"later) I nevertheless would recommend (eventually taking specimens in the slaughter house, if possible) or, in the lab on arrival, at least superficial application of fixative prior to dissecting your final specs. So the fixative can work from the luminal side at least for some minute [will penetrate/ fix some 40-50 µm! of cell layers) before you "fold/roll" the resulting tissue piece of interest. If you don't do this it would mean that - given 2-2.5 mm of [folded] thickness - fixative arrives in the center of the specimen after another hour of fixation only (approx. 1 -1.5 mm/hr @ RT. Please keep in mind that penetration rate of a fixative (here FA) is not equivalent to "fixation rate". Until the first drop of fixative bacteria certainly will be in a "living"state and therefore also will change perhaps their morphology (which then can be seen in the microscopic ). Additionally, tissular degradation processes could influence "behaviour" as well as morphology of the bacteria.

Other authors found increased values of FA-penetration : e.g. cited only one:

"The commonest primary fixative is 4% formaldehyde solution which will penetrate just 2-4 mm in 24 hours. Small specimens measuring less than 10 mm in diameter will fix by simple diffusion. Larger specimens should be thinly sliced, partly dissected or perfused to facilitate rapid and even fixation. cf.: 

R D Start, C M Layton, S S Cross, J H F Smith:  Reassessment of the rate of fixative diffusion. Jr Clin Pathol 1992;45:1120-1121 (hopefully open access also for you) @ http://jcp.bmj.com/content/45/12/1120.full.pdf

best wishes and regards,

Wolfgang

I 

Thanks again, that's a very good point!