Quantifying collective behavior in mammalian cells.

C ollective social behaviors generate one emergent response from many individual actors without central coordination: bees swarm, birds flock, and fish school. Singlecelled bacteria and amoebae collectively swarm, detecting chemical signals emitted by close neighbors to synchronize their motion. Within organisms, collective behaviors drive multicellular processes during migration and wound healing. Similar to unicellular organisms, mammalian cells coordinate their responses during these complex processes by detecting changes in concentration of signaling molecules. However, the signaling pathways used by mammalian cells are typically more interdependent than those in unicellular organisms, and gene expression varies with the time-dependent concentration of signaling molecules. Understanding the role of intercellular communications for collective behavior in mammalian cells thus demands sensitive measurements of responses to chemical signals. In PNAS, Sun et al. (1) quantify the spatiotemporal evolution of calcium response to ATP stimulation in fibroblasts. They find that the response of dense colonies of cells to ATP stimulation is faster and more synchronized than that of isolated cells. By quantitatively correlating responses between neighboring pairs of cells, they identify pacemaker cells; by physically separating cells in a hydrogel matrix, they confirm that cells communicate via gap junctions to synchronize their responses. These experiments show that both temporal and spatial dynamics of collective response in mammalian cells depend on intercellular signaling. In bacteria, transitions from individual to collective behavior are frequently coupled to changes in cell density. For example, bacteria transition from mechanisms of individual motility (2) to collective motility modes, such as twitching (3), when the cell density exceeds a critical threshold. Similarly, initially free-living planktonic bacteria form surface-associated communities called biofilms when the density of cells and the population size exceed critical thresholds, so that the concentration of signaling molecules (autoinducers) in the space between bacteria becomes sufficiently large (4) to initiate changes in phenotype. This process allows bacteria to sustain a single behavior from a collection of noisy and stochastic inputs (5). Changes from individual to social behavior in the unicellular amoeba Dictyostelium discoideum are also coupled to cell density. Synchronized waves of a molecular signal (cAMP) direct individual cells to form aggregated fruiting bodies. The proximity of nearby cells in dense colonies plays two important roles in this process. First, random fluctuations in the production of cAMP by neighboring cells below the activation threshold can initiate synchronization throughout the population (6). Second, a subpopulation of hypersensitive first-responder cells can also initiate synchronized waves (7). Studies in unicellular organisms thus highlight the roles of cell density and individual variation in intercellular signaling leading to collective behavior. Direct tests of the importance of these factors for intercellular signaling in mammalian cells, which exhibit tightly controlled signaling pathways because the penalties for deviant behavior are severe, require experiments that carefully measure collective responses. Here, Sun et al. (1) develop techniques to quantify spatiotemporal variations in chemosensing in fibroblasts, which synthesize the structural components of the ECM. When the ECM is damaged, fibroblasts must first migrate to the wound and then deposit new collagen and other ECM components to heal it. Because this process requires coordinated action by multiple cells, fibroblasts are a useful model system in which to investigate collective response to chemical signals. Fibroblasts stimulated by ATP release calcium, a second-messenger signal for processes ranging from proliferation to collagen synthesis, and the intracellular calcium concentration can be measured using fluorescent calcium indicators. Analogies to unicellular organisms suggest potential factors for collective chemosensing in mammalian cells. Because collective behavior emerges in bacteria and amoebae when the density of cells is sufficiently high, Sun et al. (1) first create adjacent 2D communities of fibroblasts of different cell densities in a microfluidic chamber. The fluorescence intensity of individual cells oscillated over time when exposed to an external flow of ATP, showing that ATP stimulation induced release of calcium. On average, the activation time required for a cell to reach a threshold intensity was shorter for cells in the high-density colony than for cells in the low-density colony, independent of flow direction. Differences in distributions of activation times for colonies of low and high densities showed that the response process was not random. Together, these observations indicated that cells must communicate to generate responses that are internally correlated. Fibroblasts communicate with each other using two main mechanisms. Adjoining fibroblasts can form gap junctions, channels that connect the cytoplasm of the cells and allow direct diffusion of small molecules; in addition, diffusing molecular messengers can transmit information through the extracellular space. Because gap junctions A B