I've heard various recommendations concerning threshold levels of infestation in drone brood, the most recent being from the Maine state apiariest who said that 10% was his own personal threshold for treating- he cited finding 5 cells with varroa out of 50. For what it's worth, find 2 out of 20 is also 10% so I gather, if you were him, then you'd treat. Since "treating" in this case involves simply removing the frame of drone brood, I see little reason NOT to remove it.
Off the top of my head I think there's about 4900 drone cells on both sides of a deep frame so fully capped, there's potentially about 500 varroa in your comb. I'd personally be inclined to pull the comb. Taking 500 mites out of your hive this time of year would be a big downer for the mite population explosion. If left to hatch, upwards of 1300 mites could emerge; and so it goes.
Michael is quite right- the bees WANT drones, as many as 10% of the population in a healthy happy hive so it's important that you let them raise some.
In any case, you need to monitor your mite population either with sugar shakes or 24 hour drop counts so you have an idea of what their numbers are doing. It could well be that Michael is right- that the mite population isn't anything to sweat about NOW and you could leave the drones in to complete their development and emerge. In other words, 3 weeks or a month from now you might be able to trap a lot more mites in that same comb. I'm not questioning that. However, without an idea of the mite population trend in your hive now, I personally don't think I could pass up the opportunity to yank so many mites so early in the season.
George-