In two-echelon vehicle routing problems, customers are supplied from distribution centers through intermediary satellites. First-echelon vehicles transport goods to satellites, at which second-echelon vehicles collect the goods and deliver them to the customers. The flow of goods must respect operational and load synchronization constraints, i.e., one must decide on the assignment of each customer to a satellite and ensure that incoming and outgoing quantities at each satellite coincide. We extend the basic version by integrating reverse flows and satellite capacities which introduce several additional synchronization issues. To solve the arising problem, we employ a matheuristic solution approach that is based on different decompositions of the overall problem. We evaluate the performance of our algorithm on instances for special cases of our problem that are already discussed in the literature and present computational studies focusing on the impact of reverse flows and satellite capacities.