Figure 1. Number of Mtb internalized determines probability of macrophage death.
The death frequency for 720 infected MDM was graphed over time as a function of the sum of Mtb internalized (raw data provided in Figure 1—source data 1). To obtain death frequencies, MDM were divided into 10 groups according to the number of Mtb internalized. The frequency of dead cells was determined at 10 movie frame (1.7 hr) intervals for each group (x-axis), and plotted against the mean sum of Mtb internalized for the group (y-axis). Interpolation between points was performed to obtain a surface. The color scale represents the frequency of cell death from low (blue) to high (red).
DOI: http://dx.doi.org/10.7554/eLife.22028.004
Figure 1—source data 1. Intracellular Mtb fluorescence through time, Movie frames of Mtb phagocytosis, and MDM frame of death, if it occurs, for IFNγ untreated MDM.
elife-22028-fig1-data1.xlsx (5.9MB, xlsx)
DOI: 10.7554/eLife.22028.005
Figure 1—figure supplement 1. Measures of cell death.
Time of cell death induced by Mtb and determined by macrophage dynamics was correlated to macrophage death as determined by the uptake of the dead cell stain DRAQ7. Time of DRAQ7 uptake was correlated with time of macrophage detachment (A) or time of cessation of movement of internal cellular structures (B).
Figure 1—figure supplement 2. Fluorescence as a measure of Mtb numbers.
we tracked the increase in Mtb by fluorescence (blue squares) versus by colony forming units (CFU, red circles) over three days of growth in culture. Means and standard deviations of triplicates. One of two independent experiments.
Figure 1—figure supplement 3. MDMs were ranked according to the sum of Mtb internalized (black), number of Mtb in the last aggregate internalized (red), number of Mtb in the first aggregate internalized (green), or randomly (blue).
Ranks (x-axis) were then plotted against the cumulative number of dead cells present up to each rank (y-axis). Ranking according to the sum of Mtb internalized gave the best separation between the top 50% and the bottom 50% of ranks (ratio top/bottom = 2.6), followed by a number of Mtb in the last clump internalized (2.4), number of Mtb in the first clump (1.4), or randomized ranking (1.0). All values were significant relative to the randomized ranking (p<0.0001), as determined from 10,000 randomizations.
Figure 1—figure supplement 4. Outcomes of macrophage-Mtb interactions.
(A) A subset of cells is shown as an example. Macrophages were numbered on the y-axis based on the sum of Mtb internalized, with cells internalizing the largest number at the top of the chart. The fate of each macrophage over time is represented by one line which intersects the y-axis at the rank of the cell. The time of macrophage death, if it occurred, is at the point where the line changes from green to dark blue. The timing and number of bacteria per pickup event is indicated by the x-axis location and shade of the marker, respectively. Pickups of cell-free Mtb are marked as circles, and pickups of dead infected cells as stars. (B) The full dataset of 759 cells from five independent imaging experiments. For cells which died and detached, the dark blue line was extended from the point of cell death to movie end.
Figure 1—figure supplement 5. Analysis of differences between macrophages grouped by the sum of Mtb internalized.
(A) Fraction of cells which died over time in each of the 10 infected cell groups (n = 72, group mean and standard deviation of Mtb per macrophage shown in the legend on the right), and the uninfected bystander group (n = 39). Lines are cubic fits as a guide to the eye. (B) Table of p-values for differences between groups as determined by bootstrap. The fraction of cells in each group was compared to the fraction of dead cells in infected groups with lower mean Mtb internalized, or to bystanders. Each group was resampled by drawing either 72 cells (for comparison to infected groups) or 39 cells (for comparison to bystanders), with replacement, 100,000 times. The p-value was calculated as the number of times the resampled fraction of dead cells at the end of the movie was lower than in the comparison group, divided by 100,000 (code provided as Source code 2). As the threshold for statistical significance, we used α/n = 0.005, where α = 0.05 for single comparisons was adjusted for n = 10 comparisons by the Bonferroni method. Table shows the groups in the first column and the comparison groups in the first row. p-values which passed the significance threshold are shown in bold.