In heterogeneous cellular networks for mobile communications, small cells are deployed within the coverage range of primary macro cells to provide for a localized capacity boost in traffic hotspots. Sharing of communication channels among the small-cell and macro-cell tiers is spectrally efficient, but causes failures of control signaling and data channels due to unmitigated co-channel interference. Consequently, the small-cell coverage range and capacity deteriorate. In Long Term Evolution (LTE) networks, the performance of control channels such as the physical downlink control channels (PDCCH) is of particular concern because they are protected only by simple interference averaging based on pseudo-random subcarrier allocation. Observing that the randomization algorithms are primarily seeded by the physical cell identifiers (PCIs) and cell radio network temporary identifier (C-RNTIs), we show that efficient interference-aware scheduling of control transmissions can be enabled by optimized allocation of PCIs, C-RNTIs and PDCCH resources. Simulations of a 3 GPP-compliant heterogeneous network show that the small-cell size can be doubled for a better macro-cell traffic offload by trading the number of active PDCCHs for a higher small-cell expansion bias. Alternatively, the small-cell PDCCH capacity can be at least tripled for high-load applications such as Voice over LTE by using selective macro-cell PDCCH muting.