Rapid and Robust Analysis of Cellular and Molecular Polarization Induced by Chemokine Signaling

The first example chosen here concerns the use of human primary T lymphocytes stimulated with CCL19. However, the same strategy can be used with primary mouse T cells, T cell lines or any other cell type responsive to chemokine stimulation. The gating strategy presented here includes a series of windows that select the focused events, then the single cells. Finally the range of fluorescence intensity is followed here as an example on two markers: the MHC Class I HLA-ABC and filamentous actin for resting (Figure 1) or CCL19-stimulated (Figure 2) T lymphocytes.

First, the Circularity index is measured in untreated (Figure 1) or CCL19-stimulated (Figure 2) T lymphocytes. The fact that resting cells have a major single peak with a high Circularity index value indicates that most cells are close to round. However, it is clear that CCL19 stimulation elicits the appearance of a second subpopulation of cells with a low Circularity index. This subpopulation encompasses cells that have adopted a polarized morphology. Also, when representing the Bright Detail Similarity index between HLA-ABC and F-actin staining, one can observe that some cells exhibit a drop in the value of this index upon CCL19 treatment, indicating that both markers show a lower degree of co-localization. Interestingly, from the dot-plots presented previously, one can position the curser on any particular dot to get a picture of the cell consisting of a brightfield image and an image for each of the stainings performed. Figure 3 thus presents examples of CCL19 stimulated T lymphocytes that fall in the gate of non polarized (Figure 3A) or polarized (Figure 3B) cells. One can clearly visualize the differences in morphology between both subpopulations. From the statistics tables below the Circularity histograms from Figures 1 and 2, one can plot the percentage of cells that fall in the polarized gate in the absence or presence of CCL19 (Figure 3C). As expected, it is clear that chemokine stimulation increases the percentage of polarized T cells (from 11.2% to 50.6%). One can also plot the mean Circularity index of the whole cell population obtained in both conditions (Figure 3D). CCL19 stimulation elicits a drop in this index. However, as only half the cells are polarized, the change in the mean Circularity index observed upon CCL19 stimulation is small.

Similarly, Figure 4 represents examples of CCL19-stimulated cells, in which both the HLA-ABC and the F-actin stainings do (Figure 4A) or do not colocalize (Figure 4B). From the statistics tables below the Bright Detail Similarity histograms shown in Figures 1 and 2, one can obtain the proportion of cells that exhibit an index value below 1.5, representing the cells that exhibit a very low degree of co-localization between HLA-ABC and F-actin. Figure 4C shows that the fraction of cells that falls in this gate increases upon CCL19 stimulation (from 3.4% to 13.9%). Collecting the values of the index for each individual cell, one can also plot the mean Bright Detail Similarity index of the whole cell population obtained in both conditions (Figure 4D). CCL19 elicits a drop in the index, however, for the same reasons as previously, changes are small. By looking at the images of the corresponding cells, it appears that this drop in HLA-ABC – F-actin co-localization is mainly due to the clustering of F-actin in one pole of the cell. Conversely, the subcellular distribution of HLA-ABC MHC Class I molecules does not seem to be grossly affected by CCL19 stimulation.

That is the reason why we next decided to test further the methodology by analyzing two markers (F-actin and phospho-ERM) known to segregate in opposing poles of chemokine-stimulated T cells. Human T cells stimulated with CXCL12 polarized strongly, as quantified in Figure 5A with the same gating strategy than before (from 5.95% to 79.10% polarized cells after 8 min of stimulation with 50 ng/mL CXCL12). In these conditions, there is a strong decrease in the mean Bright Detail Similarity index as shown in Figures 5B and 5D, meaning that F-actin and phospho-ERM colocalize less after CXCL12 stimulation. As expected, the percentage of cells exhibiting a segregation of the markers chosen here is higher (48%, Figure 5C) compared to the ones previously studied (14%, Figure 4C). Accordingly, the drop in the mean Similarity index of each condition is also more obvious (Figure 5D). Therefore, the imaging flow cytometry technique presented here is suitable to quantify the degree of co-localization of two stainings in different settings.

Finally, we decided to illustrate to what extent the perturbation of some cytoskeletal element or signaling pathway known to affect changes in T cell morphology upon chemokine stimulation, could be measured by imaging flow cytometry. Control T cells were first compared with some pretreated with the actin-disrupting drug latrunculin A (Figure 6A). Cells were stimulated with CXCL12, stained with fluorescent phalloidin and analyzed by imaging flow cytometry. Histograms showing the Raw Max Pixel intensity of the F-actin staining are shown for both cell populations (Figure 6A, top). As expected from the severing activity of latrunculin A towards actin filaments, the intensity of the phalloidin staining is lower in these cells (right panel) compared to control T lymphocytes treated with the vehicle DMSO (left panel). Moreover, latrunculin A-treated T cells exhibit a larger Circularity index indicating that they remain rounder than control cells (Figure 6A, bottom). This can be quantified by a gate set on the histograms to measure the percentage of cells that exhibit morphological alterations. Although 32% of control cells get polarized in this condition, only 10% of latrunculin A –treated T lymphocytes exhibit any shape changes (Figure 6B). Therefore, it is shown here that the imaging flow cytometry technology is suitable to measure differences in morphological modifications when one studies the potential effect of some cytoskeletal elements in this phenomenon.

Secondly, the acquisition of a large number of cells with this method also opens the possibility to analyze rapidly small subsets of cells present as a minority. In Figure 7, an example is presented where our analysis is specifically restricted to the behavior of T cells co-transfected with GFP together with plasmids encoding for dominant negative mutants of the polarity complex Par6/PKCζ. The histogram corresponding to the non-transfected cells has allowed us to gate specifically on the relevant GFP⁺ T cells (Figure 7A, left). The intensity of GFP measured as histograms that quantify the distribution of the fluorescence Intensity of the channel 2 indeed shows that only a small fraction of the whole cell population has been transfected (Figure 7A). The degree of morphological polarization of these transfected cells was then quantified by measuring only the Circularity index of the GFP⁺ cells (Figure 7B). An additional gate for T cells exhibiting low values of Circularity allowed estimation of the percentage of polarized cells in each condition (Figure 7C). The results show that disruption of the function of the Par6/PKCζ signaling pathway inhibited T cell polarization induced by both CCL19 and CXCL12 chemokines, in line with our previous findings⁷.

Figure 1. Representative non-stimulated staining results on primary human T cells. The upper left panel shows the histogram repartition of the Gradient RMS calculated with the M01 mask on the bright field images (Channel 1). The upper right panel shows the Area of the M01 mask eroded of 2 pixels, based on the brightfield images (Channel 1) of the In Focus cells gated from the previous plot. The middle panels show the Raw Max Pixel values of both the HLA-ABC and Phalloidin stainings (Channels 2 and 4, respectively) either as a dot plot (left panel) or as histograms (middle and right panels) for the individual, in-focus cells selected at the previous step. From this population of individual, in-focus cells with correct staining, is calculated the Circularity feature on the M02 mask eroded of 2 pixels of the HLA-ABC staining (lower left panel) and the Bright Detail Similarity R3 feature (lower right panel) showing the degree of similarity between the HLA-ABC and the Phalloidin stainings. For these two latest plots, statistics indicating the repartition of the population (% gated, mean of the feature, S.D.) are indicated in a table below. Please click here to view a larger version of this figure.

Figure 2: Representative staining results on primary human T cells stimulated with 200 ng/ml CCL19 for 8 min. The panel layout and gating strategy are identical to Figure 1. Please click here to view a larger version of this figure.

Figure 3: Quantification of the morphological status of T lymphocytes stimulated by CCL19. (A,B) Examples of cells from Figure 2 outside or in the “polarized” gate, respectively. (C) Histogram showing the percentage of polarized T cells obtained in the absence of CCL19 (Figure 1, -) or with 200 ng/ml CCL19 (Figure 2, +). (D) Histogram representing the mean values of the circularity index ± SE in the individual, in-focus cell population in the absence (-) or presence (+) of CCL19. *** p < 0.001. Please click here to view a larger version of this figure.

Figure 4. Quantification of the degree of co-localization between the HLA-ABC and phalloidin stainings. (A,B) Examples of cells from Figure 2 outside or in the “not colocalized” gate from the Bright Detail Similarity histogram, respectively. (C) Histogram showing the percentage of T cells exhibiting no co-localization between both markers in un-stimulated (-) or CCL19-stimulated (+) conditions. (D) Histogram representing the mean values of the Similarity index ± SE in the individual, in-focus cell population in the absence (-) or presence (+) of CCL19. *** p < 0.001. Please click here to view a larger version of this figure.

Figure 5. Analysis of F-actin – phospho-ERM exclusion upon CXCL12 stimulation of human primary T cells. (A) Percentage of polarized T cells in the absence or presence of CXCL12 (10 or 50 ng/mL). (B) Imaging flow cytometry histograms showing the repartition of the individual, in-focus cell population according to the Bright Detail Similarity index that compares the extent of co-localization between the phospho-ERM and phalloidin stainings, in the absence of CXCL12 or with 10 ng/ml or 50 ng/ml CXCL12. (C) Quantification of the proportion of cells exhibiting no co-localization of the pERM and the phalloidin stainings in T cells stimulated or not with CXCL12. (D) Quantification of the mean values of the Similarity index ± SE in each condition. *** p < 0.001. Please click here to view a larger version of this figure.

Figure 6. Effect of actin polymerization on T cell polarization. (A) F-actin staining intensity (upper panels) and Circularity (lower panels) of primary human T cells treated with Latrunculin A (500 nM, 30 min) or DMSO and then stimulated with 100 ng/ml CXCL12 for 8 min. (B) Histogram showing the percentage of polarized T cells treated with DMSO or Latrunculin A after 8 min of stimulation with CXCL12. Please click here to view a larger version of this figure.

Figure 7. Expression of the PKCζ kinase-dead mutant or the N-terminal mutant of Par6 blocks T cell polarization after CCL19 or CXCL12 treatment. (A) Imaging flow cytometry histograms showing the expression intensity of GFP, in non-transfected primary T cells or primary T cells co-transfected with GFP and an empty vector (Control) or the kinase-dead mutant of PKCζ or the N-terminal part of Par6. (B) Imaging flow cytometry histograms representing the Circularity index of the GFP-positive cell population expressing the different constructs at the basal level (NS: No Stimulation) or after CCL19 or CXCL12 stimulation (100 ng/ml for 8 min). (C) Quantification of the percentage of morphologically polarized T cells expressing the different constructs at the basal level or after CCL19 or CXCL12 stimulation. Please click here to view a larger version of this figure.