Are you postulating biology very similar to Earth, where megafauna are composed of proteins, with mineralized bones (not necessarily calcium)? I can't think of plausible biochemistries for that (I'm not a biochemist), but I wouldn't rule it out entirely.
If you're thinking of an un-Earth-like biology, then I don't see why the rigid bones couldn't be made literally out of metal. Iron at least, possibly aluminum. More exotic metals tend to have problems with the quantity available on a planet, but aren't absolutely impossible either.
Some functions of human bone would have to be moved elsewhere in the body (if those functions are even necessary). You'd not put something like marrow into these, as bending the bone might crimp off structures and flow necessary to survival. If the organs were in a ribcage or ribcage analog, those organs have to be designed in such a way that they are somewhat functional even if pinched or obstructed. This presents some serious challenge, as things like the lungs need the bronchial tubes to be open (there are solutions though, multiple redundant openings/nostrils placed such that all are unlikely to be blocked, and such that they're not long and narrow... on a human you might put these between the ribs directly into the lung). The heart itself (or any muscular pump) would be another. Organs like the liver and pancreas however, should be quite tolerable of being twisted into other shapes for a period of a few days or weeks.
So, on that note... the body would need a way of re-straightening these bones. If they aren't bent too badly, the organism might limp along for a long time without remedy. But in such cases, they weren't in too much danger from those injuries even with the brittle style of bone. Where this gives them a distinct advantage are those injuries where you or I would be crippled and require amputation (or serious, modern, medical attention).
Some biological mechanism would need to identify the region of this metal bone that was deformed, and the body would begin dissolving the metal. Likely until that portion of the bone was gone entirely. For a dent in a skull-like bone, it'd just eat it away until there was a large hole where the dent used to be, for a long bone until that bone was severed. Only then would it begin the process of reforming the metal.
The real question is how the mechanism compels the organism to reorient itself so that the new bone is straight and not malformed. We don't have such a mechanism (though we'd benefit from that too... humans eventually learned to set and splint simple breaks so that the bone would heal straight, but other animals can do no such thing). Perhaps these metal turtles go into a sort of hibernation or inactivty, and unconscious/reflexive movements reposition them for optimal results for the process.