Towards Differential Privacy in Sequential Recommendation: A Noisy Graph Neural Network Approach