Hyperglycaemia does whatever you want it to. Want to show it increases glycolysis and/or oxidative phosphorylation? No problem. Want to show it decreases both? Equally no problem. Choose your tissue, choose your duration, choose your insulin level, choose your glucose level, choose your tissue culture medium before test, choose... With the correct combination you can show anything.
But certain patterns emerge from lots of papers. In the short term hyperglycaemia increases both glycolysis and oxidative phosphorylation. Acute hyperglycaemia in neurons induces an equally acute hyperpolarisation of the inner mitochondrial membrane (a pre requisite for reverse electron flow through complex I), followed by a burst of free radicals (from reverse electron transport in the face of a low NAD+/NADH ratio?), followed by a collapse of the inner mitochondrial membrane potential (from free radical induced loss of cytochrome c?), soon to be followed by apoptosis, as you might expect
These guys set out the events nicely but suggest the mechanism is unclear. I would be willing to bet on G-3-P dehydrogenase as driving reverse electron flow using the high membrane potential from glycolysis. It seems that, under "mitochondrial preparation" conditions, ignoring reverse electron flow, G-3-P dehydrogenase also spills a reasonable dose of free radicals not only inwards towards the matrix but also outwards to the inter membrane space, in roughly equal amounts. As does complex III of course, but complex III is not specifically driven by a short side branch of hyperglycaemia-induced hyperactive glycolysis. Goodness only knows if this happens in-vivo, but let's accept that it does. Cytochrome c is on the outer surface of the inner mitochondrial membrane and spilling free radicals outwards seems a good way to oxidise the cardiolipin anchors and release one of the most important pro apoptotic proteins we have, cytochrome c.
So acute hyperglycaemic injury, in a cell type where glucose entry is essentially concentration driven, is potentially apoptotic if the injury is severe enough. Lesser but sill significant injury may come from spills of superoxide from complex I on to the mitochondrial DNA, another potentially interesting effect. Research on G-3-P dehydrogenase is still in its infancy and there are no clear cut answer as to how important this scenario might be, but I rather like it. Is it true? Who knows. It's hard to tell.
Exactly how difficult it is to transfer information from "preparations" to any semblance of "in vivo" is reviewed by Martin Brand. I like this chap, he really looks at the limitations of how much we currently know (not much, it appears) plus he came up through Naked Mole Rat research, another positive. Here's his summary of where free radicals might be produced:
Outwards spillage, directly on to cytochrome c, from G-3-P dehydrogenase and complex III...
It's quite clear that hyperglycaemia is not invariably acutely fatal to all neurons on first exposure. It takes years of following the advice of the ADA and AHA to develop diabetic neuropathy or to kill off enough central neurons (around 70%) to get the clinical label of Alzheimers and, while recurrent hyperglycaemia might get us there directly, the indirect effects are much more interesting to a mitochondriac like myself.
Chronic hyperglycaemia is where we have a depressed inner mitochondrial membrane potential, reduced glycolysis and electron transport with subsequent failure to generate superoxide.