Distributed Potential iLQR: Scalable Game-Theoretic Trajectory Planning for Multi-Agent Interactions