In papers I've participated in recently we've called it TNFα. No reviewer has corrected us as far as I can recall.
Near as I can tell, we've done this because we do not read the HUGO Gene Nomenclature Committee database any more often than we read the dictionary cover-to-cover. Instead, we read and cite papers in the literature, and since many of those are older than this apparent change, and even more recent ones are in turn citing older papers, TNFα is the only nomenclature I knew of until your question posted here.
I could have guessed that a "TNFβ" or something similar exists/existed at some point, but I really have no knowledge or familiarity with it, certainly far less than with TNFα, so a renaming there is not really of any interest or relevance to me.
Biology is a massive field and every niche has its own set of terminology that is more or less familiar.
If people in the field decide over time to start using the new name, then that will propagate the new usage. I don't see a strong motivating factor to rename in the meantime: the old name isn't misleading or easily confused with something else, it's not derogatory or offensive, the new name doesn't provide additional context or function. If anything, there's a temporary motivation to not change, in that use of the old term far outnumbers use of the new.
I suppose you can think of me however you'd like based on my explanation here - I don't have much control over that.