Publishing a result that is aesthetically pleasing, but has no practical value

Yes, this is fine. Math papers very often contain results just because they are interesting or instructive, even if they do not seem to be "useful".

You could mention this when introducing the result, with something like "The following theorem may help to illustrate the connection between blah blah blah..."

Authors also sometimes signal this sort of thing by describing a result as "pleasant", "amusing", etc, though "elegant" is probably a little too egotistical.

If the referee feels it's too much of a digression, they might suggest you take it out. But I don't think this would be the difference between acceptance and rejection.

To add to Nate Eldredge's correct (and useful!) answer and to Alexander Woo's sarcastic quip highlighting the same point, one should keep in mind that pure mathematics is, by its very definition, the part of mathematics that seeks to study mathematical structures for the sake of the pure intellectual and aesthetic value of the mathematical ideas one is trying to discover. Yes, it helps that a lot of pure mathematics has turned out to be useful beyond the wildest dreams of the people who discovered it -- a totally weird phenomenon that no one seems to understand -- but that is not the primary concern (or even the secondary or tertiary concern, usually) of the pure mathematician.

Lack of (caring about) usefulness is a feature, not a bug.

If your result "moreover connects several distant theorems together." I'd like to know that. You may not find a useful application of that result, but knowing what you just stated may help me to come up with something useful.

So publish it.