In the example presented by Claudiu—"Does the word cheese indicate something yellow?" I would follow the Chicago Manual of Style convention of italicizing words used as words:
Does the word cheese indicate something yellow?
I would do so not because this particular option is inherently superior to other ways of calling out a word used as word, but because I've done it for 35 years, to the point where the practice has become almost automatic for me.
I don't, however, use italics in all instances that involve calling out a word used as word in an answer. In single-word requests, where (in effect) the single word used as word is the crux of the answer, I follow the very frequent practice of SWR answerers on this site of putting the suggested word in boldface. I think of this as a convenience for readers who may be looking for the SWR suggestion—and nothing else—in each answer they scan. It's especially helpful I think for people who are considering their own SWR answer and want to scan the already-posted answers to see whether someone else has already suggested the word they want to nominate. For example:
For a single word meaning "a food consisting of the coagulated,compressed, and usually ripened curd of milk separated from the whey," I suggest cheese.
The other place where I use boldface formatting is when I am reporting multiple early instances of a word or phrase in answer to an etymology question. Here again the point is to give readers a quick signal as to where the relevant word or phrase appears in the block quote that I've retyped. People who want to read the examples carefully are welcome to do so; but for many readers, I suspect, only the immediate context—or perhaps only the bare existence of the word or phrase in the cited quotation—is of interest. It's very easy for italics to get lost in a paragraph-length block quote, and of course italics in the original source are far more common than boldface there. For example, from Transactions of the Department of Agriculture of the State of Illinois, volume 9 (1880):
The exportation of butter from January 1, 1879, to November 27. has reached 34,705,284 pounds; the excess over last year for same time, 13,518.230 pounds. The exportation of cheese for same time is 120,350,857; a falling-off of 8.638,316 pounds, as compared with last year, as per New York price-current report of November 27, 1879.
I'm certainly willing to reconsider my boldfacing practices if they strike enough people here as being obnoxious or counterproductive in some other way, but I think that they usefully serve the crucial purpose of making particular kinds of answers easier to read.