Unpack the cluster of pieces in front of Black’s king. Actually there are two pins here. One is obvious if you look methodically at what White currently attacks with his sliding pieces: his bishop on g5 pins the f6 knight to Black’s queen. The other is less obvious. Notice that White’s queen attacks Black’s knight on f5, and that behind this knight is another one on f6. If that rear knight on f6 were loose, the knight in front of it obviously would be pinned. But the knight in front equally is pinned if moving it would cause the piece behind it to be overmatched. The rear knight currently is attacked twice and defended twice. If Black moves away the knight on f5, the knight on f6 suddenly is attacked three times, since an attack by the queen has been added. White will take it immediately. So the knight on f5 is pinned, too.
White can use everyone's favorite way of exploiting a pinned piece; he can attack it with a pawn: g2-g4. If it moves, the knight behind it is lost as just discussed. If Black instead plays h5xg4, White imagines how he might recapture and with what consequences. He could play Ne5xg4, and now the knight on f6 would be attacked for a third time even without moving the knight in front of it. Again White takes the f6 knight next move. (Remember that it’s pinned to its queen and so cannot move.)
Examine this one until it sinks in. Once more the point is to see how a piece can be relatively pinned if moving it would cause the piece behind it to be loose—or attacked more times than it is defended.