Underlay femtocells promise to enhance the coverage and rate performance of next-generation wireless networks. Nevertheless, many concerns still remain in the context of shared-spectrum operations and quality of service (QoS) provisioning at the macrocell users. In this paper, we introduce a novel approach for cooperative femtocell-to-femtocell interference management based on interference draining. Accordingly, a group of femtocells can decide to cooperate and improve their downlink rate, by exploiting the available frequency-spatial directions, and still guarantee a minimum target QoS at the closest MUEs. To address this problem, we use tools from cooperative game theory that enable the femtocells to decide, in a distributed manner, on whether to cooperate or not. In the proposed framework, each femtocell access point individually decides its own cooperative strategy, and maximizes a utility function that captures the cooperative gains and the limitations due to the macrocell users QoS targets. We show that, using the proposed approach, the femtocells can self-organize into a network partition composed of disjoint groups of femtocells which form the recursive core of the cooperative game. Simulation results show significant gains in terms of average payoff per femtocell, reaching up to 23%, for a network of K = 140 FBSs, relative to the non-cooperative approach.