Blocking the ART2.2/P2X7‐system is essential to avoid a detrimental bias in functional CD4 T cell studies

Abstract

Murine T cell subsets differ in their expression level of P2X7. Depending on several parameters like extracellular NAD+, P2X7 can be ADP-ribosylated rapidly by adjacent ARTC2.2 resulting in susceptibilities to apoptosis to a varying extent. This detrimental effect can be prevented when drugs like KN-62 are present during cell preparations. A consistently met problem when analyzing intestinal immune cells is the low yield in viable cells when compared to cell preparations of spleen or peripheral lymph nodes (pLN). Published evidence suggests that this may at least in part be caused by the activity of the surface receptor P2X7 1. P2X7 represents an ATP-gated ion channel executing variegated functions in immune cells 2. An effect of a continuous activation of P2X7 is induction of cell death that is considered a naturally existing tool to shape immune responses. High local concentrations of ecto-ATP come into existence under distinct circumstances like inflammation or injury that are distinguished by a burst of cell necrosis 1. Similarly, in the course of preparing cell suspensions for subsequent in vitro investigations the inherently triggered release of ATP can cause cell death as an unwanted side effect that impacts on the reliability of the experimental data 3. In mouse but not human, this problem is massively aggravated by the concomitant activity of ART2.2, an ecto-enzyme using NAD+ as a substrate to ADP-ribosylate a variety of targets such as P2X7 2-4. This modification results in a long lasting P2X7 channel opening thereby propelling cell death. For experimental work, the devastating power of the ART2.2/P2X7-system is defined by three facts: (i) significantly lower concentrations of ecto-NAD+ compared to those of ecto-ATP can cause cell death 3, (ii) ART2.2 can ADP-ribosylate in cis neighboring P2X7 also at lower temperatures efficiently 5, and (iii) once ribosylated, the detrimental effects cannot be reverted by washing away NAD+ as is possible in case of ATP that needs to be present permanently to keep the ion gate open 3. During investigations addressing functions of follicular helper T (TFH) and follicular regulatory (TFR) cells isolated from different secondary organs, we noticed that RNA levels coding for ART2.2 and P2X7 were significantly higher in cells of Peyer's Patches (PP) than of pLN (data not shown). Moreover, mice lacking P2X7 harbor significantly more TFH cells in their PP than wild type animals 6. To investigate this in more detail, we made use of a C57BL/6 strain expressing GFP in the Foxp3 locus (Foxp3 mouse) 7 and focused on four different CD4 T subsets, GFP−PD1loGL7− naïve T cells, GFP+PD1loGL7− Treg cells, GFP−PD1hi TFH cells, and GFP+PD1hi TFR cells. PD1 is a key marker of TFH and TFR cells and PD1hi cells represent most matured follicular T cells also expressing highest levels of CXCR5 among the CD4 T cells analyzed (Fig. 1A and data not shown) 8, 9. In contrast, the role of GL7 as a marker for TFH cells residing in the germinal center is discussed controversially. Cells from PP and from immunized pLN were included in the study. Whereas all PP cells under scrutiny expressed ART2.2 highly, there was a significant proportion of pLN TFR cells with decreased ART2.2 levels (Fig. 1B). Regarding P2X7, naïve CD4 T cells are virtually devoid of this receptor (Fig. 1C). Expectedly, Treg cells express P2X7 10 but TFH as well as TFR cells possess even higher P2X7 surface levels. Notably, on average PP cells express higher amounts of both, ART2.2 and P2X7 (Fig. 1D). This is in accordance with reports showing that P2X7 expression is sensitive to retinoic acid, a vitamin A metabolite produced by intestinal dendritic cells 11, 12. These results would suggest a different sensitivity to ART2.2/P2X7-mediated cell death depending on both, the CD4 T cell subset looked at and the origin of the material (as shown for intestinal CD8 T cells 12). A panel of compounds exists that inhibit either ART2.2 or P2X7 activity 3. Among these we selected the cheapest commercially available substance, KN-62, that blocks P2X7. Before disrupting them, intact organs (PP and pLN) were distributed equally into two dishes containing medium with or without KN-62 (Fig. 1E). Following preparation of single cell suspensions, cells were plated and incubated for another hour before their status was checked via annexin V/DAPI staining. This confirmed an overall higher sensitivity of PP CD4 T cell subsets to apoptosis/cell death when compared to their pLN counterparts (Fig. 1F). Moreover, TFH cells and Treg cells are more affected than naïve CD4 T cells. These observations fit well to the corresponding expression levels of P2X7 except for pLN TFR cells that are less sensitive than pLN Treg cells. This might be explained by the reduced ART2.2 expression of the former (Fig. 1B). Presence of KN-62 during all steps following harvest of organs was capable to protect efficiently all cell subsets (Fig. 1G) proving that the observed susceptibility to apoptosis/cell death is indeed caused largely by P2X7 activity. As an alternative to KN-62, we tested the selective P2X7 antagonist A-438079 in an identical experimental setup and found that this drug also efficiently protected cells from apoptosis (Fig. 1H). Single cell suspensions for follow-up experiments can be prepared relatively easily from lymphoid organs at low temperatures thereby avoiding exhaustive cell death. In contrast, cell isolation from peripheral organs must be done in presence of digestive enzymes at 37°C. When preparing CD4 T cells from liver, we found that presence of KN-62 during all steps potently protected Treg cells from apoptosis (Fig. 1I). Moreover, we observed that splenic Treg cells propagated well in vitro upon stimulation with IL-2 and plate bound anti CD3 antibody regardless whether they were isolated and investigated in presence or absence of KN-62 (Fig. 1J). This indicated that the presence of the drug preserved functionality of the Treg cells under these experimental conditions. To demonstrate further the impact of NAD+, cells were incubated in the presence of exogenously added NAD+ up to a concentration of 100 μM (Fig. 2A and B). This provoked a progressive drop in viability that could be efficiently prevented by KN-62. We performed analyses determining the production in vitro of the cytokines IFN-γ and IL-10 by Foxp3− and Foxp3+ CD4 T cells of PP. This necessitated the incubation of the cell preparations for 4 h at 37°C in the presence of PMA/ionomycin/Brefeldin A as specified in Supporting Information. Unsurprisingly, incubation in medium caused a drastic reduction in frequencies of Foxp3+ cells when compared to controls (cells kept on ice all the time) or cells in medium + KN-62. Furthermore, a drastic loss was observed in cells actively producing cytokines in both subpopulations (Fig. 2C). B cells are resistant to the deleterious effects of the ART2.2/P2X7-system 3 and together with T cells they make up the vast majority of cells isolated from a secondary lymphoid organ. Therefore, the observed drop in CD4 T cell frequencies (Fig. 2D, upper panel) is due to a preferential loss of these cells during incubation. This was also observed for TFH cells and for all CD4 Foxp3+ cells (Fig. 2C and D). However, among the latter PD1lo Treg cells and TFR cells did not display a divergent sensitivity under these experimental conditions. Again, the observed biases were prevented efficiently when KN-62 was present during cell incubation. These observations demonstrate that the ART2.2/P2X7-axis differentially affects survival and function of CD4 T cell subsets. When neglecting this aspect in data interpretation, this may lead to erroneous conclusions regarding prevalence and function(s) of the subsets investigated. This is even more evident when comparing identical cell subsets with each other that were isolated from different lymphoid organs. The discovery that particularly Treg cells are sensitive to destruction via the ART2.2/P2X7-axis led to the development of a nanobody based strategy to block ART2.2 that prevents loss and also functional impairment of these cells in follow-up experiments 10, 13. The nanobody is commercially available and it was shown that injection of approximately 50 μg into an animal just 15 min before sacrificing mice and harvesting Treg cell from spleen efficiently preserved functionality and yield of Treg cells. However, this approach requires an approval to perform an animal experiment. In addition, based on rough estimates the use of KN-62 is approximately ten times less cost intensive per animal analyzed making it an attractive and affordable alternative tool in the study of T cell subsets that are prone to apoptosis in general but also in an organ specific manner. This work was supported by DFG grant BE1886/5-1 to G.B. The authors declare no commercial or financial conflict of interest. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.