Good luck with the project. As discussed in RMS there are many descriptive tools you can use to interpret principal components, and there is a case study of sparse principal components in the book. One simple approach is to run stepwise variable selection just as a descriptive tool to predict one of the cluster summary scores.