T-cell calcium dynamics visualized in a ratiometric tdTomato-GCaMP6f transgenic reporter mouse

Essential revisions:

The authors should provide quantification of the data described in relation to Figure 10. A breakdown of the different patterns with number mice/LN analyzed and number of events observed. If statistically valid evidence of sub cellular Calcium is found then this should be discussed in the context of other invert studies that look at such local Ca²⁺ signals. If it turns out that this is not possible, a more conservative interpretation should be presented and the advantages of the reporter would still be clear.

We believe that a detailed characterization of patterns of cellular behavior in relation to sparkles would be beyond the scope of this paper. Instead, we focused on ensuring that the signal-to-noise issues are addressed thoroughly and that the green/red ratio quantitation validates that sparkles are truly subcellular local signaling events. We are not certain about which “invert” studies the comment refers to. Local subcellular Ca²⁺ signals have been detected using chemical indicators in a variety of cell types over the last 20-30 years. We further note that the accepted companion paper illustrates the relation of sparkles to cellular motility (Figure 7), presents an independent method for analyzing sparkles (Figure 8), and a peptide-MHC-dependent mechanism (Figure 9). Figure 9 of the current manuscript illustrates detection of abundant sparkles in homozygous Cd4-Salsa6f lymph node; image processing minimized fluctuations in background fluorescence in order to identify cell-wide Ca²⁺ signals and sparkles.

In response to the reviewer comment, Figure 10 was revised and now includes a pixel intensity histogram, and two illustrations of how sparkles are localized in adoptively transferred Cd4-Salsa6f^+/+ T cells. The amplitude histogram (new Figure 10A) illustrates sparkle signals that are more than 5 standard deviations above background noise. The remaining panels illustrate image processing on two representative sparkles to show that green/red ratios exhibit regional variation that are statistically significant beyond p values of 0.0001. The text in the Materials and methods has been amended to clarify sparkle identification, and we apologize for the confusion. As motile T cells extend, they adopt a polarized morphology, in which most cytoplasm is concentrated at the front and back with little on the sides of the cell (see our accepted companion eLife motility manuscript for additional example images). Because of this arrangement of cytoplasm, it is possible that apparent sparkles might arise from global Ca²⁺ signals in which asymmetry in the distribution of T cell cytoplasm limits detection of fluorescence from other regions of the cell, or alternatively from polarized T cells in which half of the cell lies outside the imaging field. Bright local Ca²⁺ events (>5.4 SD) were excluded from the set of sparkles if they coincided with a cell-wide event, even if the intensity of cell-wide event was much lower (2.1 SD threshold). We show that sparkles are not subregions of cells with slightly higher amounts of fluorescence that we produced by inappropriately thresholding cell-wide Ca²⁺ signals. To address this potential concern, we selected example cells in which the front and back of the cells were clearly visible. We used a simple quantitative approach to establish that Ca²⁺ signals detected can selectively occupy either the front or back of these cells. Local signals with peak SNR of 6.6 and 7.3 were chosen, comparable in size to local signals reported in whole Cd4-Salsa6f lymph nodes (Figure 9). For both cells in Figure 10, five views are now shown: the median filtered maximum projection red TdTomato and green GCaMP6f channels (new Figure 10B,C,H,I); the green GCaMP6f channel processed image (new Figure 10D,J); two-channel composite images which have been green channel thresholded but are a clear way to visualize the locations of Ca²⁺ signals in cells (Figure 10E,K; compare to panels D and J to ensure that the distribution of green fluorescence has not compromised by being displayed in this manner); and the G/R ratio (Figure 10F,L). G/R ratio largely reflects the location of detected sparkles.