Here is the full explanation:
The Cavendish bananas that are most often sold in grocery stores do not actually develop seeds. They are naturally sterile (triploid) and can only be propagated vegetatively. However, each aborted ovum has a vascular network consisting of xylem strands and associated cells containing astringent tannins. Upon ripening, the tannins polymerize into a semi-solid mass called 'tannin bodies' which fill the cells. The tannin bodies sometimes incorporate red-brown pigments from polyphenol oxidase activity (like the browning reaction in cut apples) as the cells age, and can therefore be seen as the red-brown bodies in this case. They are associated with the xylem strand which give them a chain-like appearance.
Because I've received some degree of skepticism when I've posted banana material in the past, I decided to conduct an experiment to see if I could recreate their appearance through some laboratory digestion techniques. So here was my process:
Step 1. Sacrifice my banana from lunch for the good of science
Step 2. Add bananas to pre-prepared tubes of proteinase K in buffer. (Unfortunately I didn't have any amylase which would have digested the carbohydrates in the banana. However, this was the best I could do to simulate the digestive process). Vortex to mix and then incubate at 56 degrees Celsius while gently shaking (the standard tissue digestion that we use for PCR pre-processing).
Step 3. Check regularly. I first checked every 10 minutes , but very quickly realized that this was going to be a long process. After the first 4 hours, this is how the banana sections looked:
Step 4. Check again, 24 hours later - looking pretty good!
Step 5. Final check - 48 hours. Success! I think that these look nearly identical to our clinical specimen. What do you think?
Again, it's not a perfect match since the actual patient specimen was subjected to the entire gastrointestinal digestive process. However, the strings of tannin bodies can clearly be seen. Here is how they look microscopically: