I reject the original dichotomy.

Biology doesn't know what a "mental illness" is. Biology has variation of various sorts. All of them tend to induce various compensatory mechanisms. Some of them increase reproductive fitness; others don't. You don't get labels, only phenomena.

Thus, "mental illness" is a human construct. You don't do tests and discover, "Oh, this is a mental illness!" We declare some things broken-in-such-a-way-that-we-want-to-fix-them-like-we-want-to-fix-a-broken-bone-or-cancer-or-a-yeast-infection-or-something. Others we say are part of normal variation, even if they're pretty far out, especially if they're not causing anyone any problems.

Some people don't want to have children. Is this a mental illness? Even if it's caused by a different level of neuromodulators, and likely could be "fixed" in many people if we could deliver the right cocktail of neuromodulators and psychotherapy? What about shyness? PTSD? Synesthesia? Fear of spiders (the ordinary kind, not paralyzing terror)? Fear of speaking up in public (often treatable with alcohol)? Belief in God? Disbelief in God? Photic sneeze reflex?

Of course you can go into people's brains and discover that hm, they seem to release relatively small amounts of oxytocin, and that correlates with the decreased strength of their emotional attachments with other people and increased antisocial and asocial behavior. But does that make them ill or does that just tell us that, like with absolutely everything, the way they are is because of how they are: because of some mechanism that is part-of-them? Either way, if we can change it with oxytocin, and they say they want it, should we withhold it from them (once we determine it's safe, reasonably likely to be efficacious, etc.)?

The bottom line is that some issues that are very strongly dependent on mind cause serious problems for the people who have those traits. Sometimes they're only problems because society is the way it is now, and they weren't problems before, and might not be again in the future. Regardless, they cause problems for people now. We know because they tell us. Shouldn't we try to help them?

Sometimes we help by changing them (prosthetic limbs). Sometimes we help by changing conditions in society (disabled access ramps).

It doesn't fundamentally matter whether we decide to use the "illness" label or just the "gosh-what-can-we-do-here" label. Gender dysphoria exists, causes serious problems for many people who express it, and is recalcitrant to most attempts to improve things.

If we develop a drug that induces gender phoria towards one's existing gender, or a drug that induces attraction to the opposite sex in those who don't feel it, or a propensity to believe in God in those who don't, it doesn't follow that gender dysphoria, homosexuality, or atheism are mental illnesses any more than not loudly singing off-key is a mental illness because we can "fix" it with alcohol. And if we can't develop a drug to treat suicidal tendencies, it also doesn't follow that suicidal tendencies are not a mental illness.

Illness is a societal determination of what is a really serious problem and warrants immediate and substantial attention to try to make things better.

So.

This whole axis of debate is a red herring. Right now, we know gender dysphoria is a thing. And if someday we manage to develop a drug or cognitive behavioral therapy or whatever to reduce some cases of gender dysphoria in a non-surgical, non-hormonal, non-"affirming" way, that just gives people who feel it more options about how to choose their path in life. (Assuming we let them rather than mandate a particular path--there may be reasons of safety or cost to facilitate one way over the other.)