In my opinion, this is quite normal. If you think about it, that is essentially what preprints are for: making a version of the article—often very similar to the one submitted to the journal—available before peer review. Some journals even encourage or publish preprints, while others do not.
The main point of caution, in my view, is trust in the colleague you are sharing it with. You need to be confident that they will not circulate the manuscript further or upload it to sites like Scribd or Academia.edu, which could create problems regarding authorship or even raise plagiarism concerns.
If it is someone from your own research team, I don’t think it should be an issue.
Hello! As long as you are the sole author and your colleague is not your supervisor, I do not see why you would do that. Your colleague can read the article after publication.
I agree with Felipe Aqueveque. The manuscript is your work (ok, be sure any co-author is happy) so why on earth shouldn't you share it with a colleague? Knowledge creation is a social process not limited solely to the formalities of reviewer comment but open to anyone who you consider to be interested and helpful. Just one caveat, of course: if it's in submission to one journal, then as a courtesy to them, you should not send it out to another until the first journal's review process is complete.
Nowadays almost all our manuscripts from our group go to open public repositories (e.g. bioRxiv) even before submission and we do not submit to journals not allowing this. Then the published paper is linked to in the repo when it comes out. Exceptions are some invited papers and book chapters.