Temperature and flow rate limit the optimal ex-vivo perfusion of the heart - an experimental study