Optimizing Federated Learning using Remote Embeddings for Graph Neural Networks