While it's hard to overstate the role of microbial action, this is pretty much what I think, too.
While temperature differences can affect the rate of production of CO2, I doubt there will be a very large difference in the total amount of CO2 produced time-independent, so if your bottles won't explode after a couple years at 55F, I doubt they'll explode after a couple months at 85F *just because of an increased rate of metabolism*. It is possible that your final amount of CO2 will be slightly different at each temperature, as you will be changing the ratios of the various metabolic processes taking place in the beer, but I can't imagine it'd be so different as to take you from well below to well above the safety limits of the bottles. Also, if we get hot enough, we'll start to slow down the fermentation as we thermally stress the yeasts too much.
Good, new corks are pretty impermeable to CO2 on the timescales we're talking about, so I believe we can treat the bottles as closed systems in both the high-temperature and low-temperature cases.
We can estimate the contribution from the two other listed factors: thermal expansion of the gas in the headspace and the reduced solubility of CO2 in water at higher temperatures. I'm going to tread these things ideally, as that's probably good enough for our purposes. I'm also going to assume that the system is just pure water and CO2.
Thermal expansion:
From the ideal gas law/Stupac's post, P_2/P_1 = T_2/T_1 for a fixed volume and amount of gas. So from 55F to 85F, that's only about a 6% increase in pressure. Not a ton, but I guess enough if you're already near the structural limits of bottle.
Gas solubility/vapor pressure:
Water provides a partial pressure of about 1.5 kPa at 55F, which increases to around 4 kPa at 85F. A 2.5 kPa difference is likely not a significant contributor to the total pressure difference, as that's less than a 1% difference in pressure at the failure point for lambic bottles.
To a decent approximation,
this reference implies that ratio of the partial pressure of CO2 in the headspace to the concentration of dissolved CO2 to the ratio of the concentration of dissolved CO2 increases by almost 50%. Without actually doing the algebra, I'm going to assume the headspace is small enough and that most of the pressure in the headspace at 55F is CO2 (which the vapor pressure of water presented above strongly implies) that we can treat that as the increase in pressure in the headspace directly. So if we are at a safe 600 kPa at 55F, then we'd climb to a very dangerous 900 kPa due to CO2 leaving solution.
Thus, we see that the largest contribution to the exploding bottles is likely due almost entirely to the reduction in solubility of CO2 in the beer as the temperature increases, though there are some other smaller contributions.