White’s bishop has a clean line to h8, which is half of a mating idea. Completing it would require getting a heavy piece onto h8, using a heavy piece to put pressure down the g-file, or attacking the king with a knight. Since all the files leading to Black’s king are closed and White has no apparent way to open any of them with sacrifices, think about the knight. You can move it to e7 in two steps: Nd5, then Ne7. And then see that moving the knight off of c3 would also unmask the discovered attack QxQ. Or you might have started by seeing the kernel of a discovery for White, then looked to see what the knight could do once it left c3, and realized it could threaten to execute a classic knight-and-bishop mate by leaping to e7. The point either way is the same: after White plays Nd5, Black has to take some measure to prevent White from mating next move with Ne7; after he does so (say, with c6xN), he loses his queen to QxQ.

This is yet another case of using a discovered attack to buy two moves to get a knight into mating position—though this time the likely result is a gain of material rather than mate. This often is how mating threats function: you may not carry them out, but they force your opponent to make sacrifices to stop you. The position also is a reminder that even a knight far away from the action must not be overlooked as a resource. This is an extreme case, but it shows that even from c3 a knight still may be only a move away from threatening to mate the castled Black king.