Integrating conflicting chemotactic signals. The role of memory in leukocyte navigation

Abstract

Leukocytes navigate through complex chemoattractant arrays, and in so doing, they must migrate from one chemoattractant source to another. By evaluating directional persistence and chemotaxis during neutrophil migration under agarose, we show that cells migrating away from a local chemoattractant, against a gradient, display true chemotaxis to distant agonists, often behaving as if the local gradient were without effect. We describe two interrelated properties of migrating cells that allow this to occur. First, migrating leukocytes can integrate competing chemoattractant signals, responding as if to the vector sum of the orienting signals present. Second, migrating cells display memory of their recent environment: cells' perception of the relative strength of orienting signals is influenced by their history, so that cells prioritize newly arising or newly encountered attractants. We propose that this cellular memory, by promoting sequential chemotaxis to one attractant after another, is in fact responsible for the integration of competitive orienting signals over time, and allows combinations of chemoattractants to guide leukocytes in a step-by-step fashion to their destinations within tissues.

Figure 1

Neutrophils migrating in a uniform field of chemoattractant display directional persistence. Neutrophils 400–700 μm from their starting well were tracked as they migrated under an agarose gel containing a uniform concentration of LTB4 (a–d). (a) Average migration angle during the initial 2 min of observation. A majority of cells point forward (0°), or away from the starting well. (b) Decay of cellular persistence. At each time point, cells that had not yet turned >90° from their starting angle were considered persistent. (c–e) Plot of each cell's FMI during the initial 2-min period versus its FMI during the subsequent 4-, 7-, or 9.5-min period. The FMI is the ratio of net forward progress to total path length, as diagrammed (initial index = w/x; subsequent index = y/z). The plot of initial versus subsequent FMI provides information about cell turning behavior. Possible paths of several example cells are shown (A, B, C, and D), as discussed in the text.

Figure 1

Neutrophils migrating in a uniform field of chemoattractant display directional persistence. Neutrophils 400–700 μm from their starting well were tracked as they migrated under an agarose gel containing a uniform concentration of LTB4 (a–d). (a) Average migration angle during the initial 2 min of observation. A majority of cells point forward (0°), or away from the starting well. (b) Decay of cellular persistence. At each time point, cells that had not yet turned >90° from their starting angle were considered persistent. (c–e) Plot of each cell's FMI during the initial 2-min period versus its FMI during the subsequent 4-, 7-, or 9.5-min period. The FMI is the ratio of net forward progress to total path length, as diagrammed (initial index = w/x; subsequent index = y/z). The plot of initial versus subsequent FMI provides information about cell turning behavior. Possible paths of several example cells are shown (A, B, C, and D), as discussed in the text.

Figure 2

Neutrophils migrating in an LTB4 gradient display a chemotactic bias, independent of the effects of directional persistence. Neutrophils 400–700 μm from the starting well were tracked as they migrated under an agarose gel in response to a distant LTB4 source. The plot shows each cell's FMI over an initial 2-min period versus its FMI during the subsequent 9.5-min period. The y-intercept of the regression line is significantly greater than zero (0.34, with a 95% confidence interval of 0.24–0.43), indicating a chemotactic bias towards the LTB4 source.

Figure 3

Neutrophils migrating from a local chemoattractant source can chemotax towards a distant chemoattractant. (a) Neutrophils were tracked as they migrated from a well containing IL-8 (1 pmol), LTB4 (0.3 pmol), or medium only towards a distant well containing IL-8 (1 pmol) or LTB4 (0.3 pmol). The bar graph shows the chemotactic bias, or the y-intercept of the initial versus subsequent FMI plot, exhibited by cells migrating under various conditions. The initial versus subsequent FMI plot for cells migrating from medium only towards LTB4 is shown in Fig. 2; the other plots are not shown. The average velocity of cells was similar under all conditions (24–27 μm/min, as indicated in the figure). We tested each condition on the same day, using neutrophils from the same blood donor, and scored 30–60 cells per condition. (b) Neutrophils were tracked as they migrated towards an IL-8 or LTB4 source 1.5 mm away. Cells 400–700 μm from the chemoattractant well at the time point relevant to the experiment described in (a) were analyzed and their chemotactic bias was determined. Bar lengths represent the chemotactic bias, or y-intercept of the initial versus subsequent FMI plot. Error bars represent the 95% confidence interval for the y-intercept. Distances from each well to the center of the videotaped field are indicated on the diagram (in micrometers).

Figure 4

Neutrophils can integrate directional signals from chemoattractant sources presented at an angle. A neutrophil-containing well and two chemoattractant source wells were placed in an equilateral triangle. Chemoattractant source wells contained medium, IL-8 (10 pmol), or LTB4 (10 pmol). Cells were allowed to migrate for 150 min, after which they were fixed, stained, and photographed. In the presence of a single chemoattractant source (top left and middle), cells migrated toward that source only. In the presence of two identical chemoattractant sources (bottom left and middle), the majority of cells migrated furthest in the direction of one chemoattractant source or the other. In the presence of two different chemoattractant sources (right), cells migrated in a broad front, with the majority of cells migrating furthest in an intermediate direction between the two sources. Each condition was performed in triplicate; photographs show representative results.

Figure 5

Neutrophils' chemotactic bias depends on the direction from which they have migrated. Neutrophils were tracked as they migrated into a central position between their starting well and a well 1.5 mm away. (a) Chemotactic bias in the central region between two chemoattractant source wells. Chemotactic biases are shown for a single population of migrating cells (top, cells 600–900 μm from each source) and for two fluorescently labeled neutrophil populations allowed to migrate simultaneously (bottom, cells 550–950 μm from each source). (b) Representative micrograph of simultaneous migration at the end of the observation period showing that the two neutrophil populations had intermingled significantly (chemotactic bias of each population in this experiment shown below micrograph). Bars represent the magnitude of cells' chemotactic bias (y-intercept of the initial versus subsequent FMI plot) under different conditions, calculated as a weighted, pooled average from two experiments performed on different days with different blood donors. Error bars show the 95% confidence interval. Results of individual experiments are shown in Table . Distances from each well to the center of the videotaped field are indicated on the diagram (in micrometers). (c) In this model, cells can vectorially add orienting signals from opposing sources, but cellular perception of orienting signals is determined, in part, by the cell's chemoattractant exposure history. Thus, when cells are migrating in the presence of equivalent chemotactic signals from opposing IL-8 and LTB4 sources, a cell migrating into a central region from the IL-8 direction is less sensitive to the IL-8 signal, and chemotaxes towards LTB4. Conversely, a cell migrating from the LTB4 direction is selectively less sensitive to the LTB4 signal, and chemotaxes towards IL-8. Bar, 100 μm.

Figure 5

Neutrophils' chemotactic bias depends on the direction from which they have migrated. Neutrophils were tracked as they migrated into a central position between their starting well and a well 1.5 mm away. (a) Chemotactic bias in the central region between two chemoattractant source wells. Chemotactic biases are shown for a single population of migrating cells (top, cells 600–900 μm from each source) and for two fluorescently labeled neutrophil populations allowed to migrate simultaneously (bottom, cells 550–950 μm from each source). (b) Representative micrograph of simultaneous migration at the end of the observation period showing that the two neutrophil populations had intermingled significantly (chemotactic bias of each population in this experiment shown below micrograph). Bars represent the magnitude of cells' chemotactic bias (y-intercept of the initial versus subsequent FMI plot) under different conditions, calculated as a weighted, pooled average from two experiments performed on different days with different blood donors. Error bars show the 95% confidence interval. Results of individual experiments are shown in Table . Distances from each well to the center of the videotaped field are indicated on the diagram (in micrometers). (c) In this model, cells can vectorially add orienting signals from opposing sources, but cellular perception of orienting signals is determined, in part, by the cell's chemoattractant exposure history. Thus, when cells are migrating in the presence of equivalent chemotactic signals from opposing IL-8 and LTB4 sources, a cell migrating into a central region from the IL-8 direction is less sensitive to the IL-8 signal, and chemotaxes towards LTB4. Conversely, a cell migrating from the LTB4 direction is selectively less sensitive to the LTB4 signal, and chemotaxes towards IL-8. Bar, 100 μm.

Figure 5

Neutrophils' chemotactic bias depends on the direction from which they have migrated. Neutrophils were tracked as they migrated into a central position between their starting well and a well 1.5 mm away. (a) Chemotactic bias in the central region between two chemoattractant source wells. Chemotactic biases are shown for a single population of migrating cells (top, cells 600–900 μm from each source) and for two fluorescently labeled neutrophil populations allowed to migrate simultaneously (bottom, cells 550–950 μm from each source). (b) Representative micrograph of simultaneous migration at the end of the observation period showing that the two neutrophil populations had intermingled significantly (chemotactic bias of each population in this experiment shown below micrograph). Bars represent the magnitude of cells' chemotactic bias (y-intercept of the initial versus subsequent FMI plot) under different conditions, calculated as a weighted, pooled average from two experiments performed on different days with different blood donors. Error bars show the 95% confidence interval. Results of individual experiments are shown in Table . Distances from each well to the center of the videotaped field are indicated on the diagram (in micrometers). (c) In this model, cells can vectorially add orienting signals from opposing sources, but cellular perception of orienting signals is determined, in part, by the cell's chemoattractant exposure history. Thus, when cells are migrating in the presence of equivalent chemotactic signals from opposing IL-8 and LTB4 sources, a cell migrating into a central region from the IL-8 direction is less sensitive to the IL-8 signal, and chemotaxes towards LTB4. Conversely, a cell migrating from the LTB4 direction is selectively less sensitive to the LTB4 signal, and chemotaxes towards IL-8. Bar, 100 μm.

Figure 6

Cellular memory can guide leukocyte navigation through complex chemotactic fields. Leukocytes migrating into a complex chemotactic field navigate to their targets using cellular memory. Cells expressing chemoattractant receptors (a–c) for ligands A–C can navigate to their targets, regardless of where they enter the tissue. A cell that enters near the stromal cell secreting agonist B first migrates up the B gradient (cell 1). As the cell migrates within range of the A gradient, loss in cellular sensitivity to B enhances the cell's migration towards A. Following the A gradient draws the cell close enough to perceive a dominant attractant (C) from its end target. A cell that enters between the two stromal cells may initially migrate up the steepest local gradient it encounters. If it migrates up gradient A (cell 2), it quickly approximates the gradient of the dominant agonist C, which directs it to its target. If it migrates up gradient B (cell 3), any loss in sensitivity to B will increase the influence of gradient A. As the influence of A increases, the cell may migrate within range of the dominant agonist, C, and be attracted towards it destination. Note that if both stromal cells secreted attractant B, cell 3 would likely continue up the steepest local B gradient, and would be unlikely to wander within range of the chemoattractant from its target (C). This system is resilient: if the end target were to move to a site near the stromal cell secreting B, cells could be easily shunted towards the new target. Similarly, stromal cells secreting A and B could recruit cell 4, a different leukocyte subset (expressing receptors a, b, and d), to a target site near stromal cell B. In the absence of an end target, cells would be expected to linger between agonists A and B, as they would become relatively more sensitive to one of the agonists as they approached the source of the other agonist.