Attacking the Spike: On the Transferability and Security of Spiking Neural Networks to Adversarial Examples