As you'll recall, previous research has suggested that metformin inferes with mitochondrial efficiency. I can't help but wonder if that's the real MOA behind this finding. Here's an excellent link that summarizes the risks posed by metformin, including mitochondrial damage. Given that ApoE4 carriers already exhibit decreased mitochondrial efficiency, we may want to think twice about this medication.
At the center of metformin’s mitochondrial damage is its affect on the most basic of mitochondrial functions – ATP (cellular energy) production. Metformin reduces mitochondrial ATP production in skeletal muscle by as much as 48%. Sit with that one for a moment, a 48% reduction in cell fuel. Imagine functioning at only half capacity. This would make basic activities difficult at best and exercising to lose weight a very unlikely proposition. Imagine similar reductions in ATP production were observed in the brain or the heart or the GI tract (which, when on Metformin are likely), the types of disturbances we might see become quite clear: neurocognitive decline, psychiatric instability, neuropathy, heart rate, rhythm and blood pressure abnormalities, along with gastrointestinal distress to name but a few. Underlying all of these symptoms, and indeed, all mitochondrial dysfunction, is an overwhelming sense of fatigue and malaise.[Emphasis mine.]