A critical role for DAP10 and DAP12 in CD8+ T cell-mediated tissue damage in large granular lymphocyte leukemia

Abstract

Large granular lymphocyte (LGL) leukemia, or LGLL, is characterized by increased numbers of circulating clonal LGL cells in association with neutropenia, anemia, rheumatoid arthritis, and pulmonary artery hypertension (PAH). Emerging evidence suggests that LGLL cells with a CD8(+)CD28(null) phenotype induce these clinical manifestations through direct destruction of normal tissue. Compared with CD8(+)CD28(null) T cells from healthy controls, CD8(+)CD28(null) T cells from LGLL patients have acquired the ability to directly lyse pulmonary artery endothelial cells and human synovial cells. Here, we show that LGLL cells from patients possess enhanced cytotoxic characteristics and express elevated levels of activating natural killer receptors as well as their signaling partners, DAP10 and DAP12. Moreover, downstream targets of DAP10 and DAP12 are constitutively activated in LGLL cells, and expression of dominant-negative DAP10 and DAP12 dramatically reduces their lytic capacity. These are the first results to show that activating NKR-ligand interactions play a critical role in initiating the DAP10 and DAP12 signaling events that lead to enhanced lytic potential of LGLL cells. Results shown suggest that inhibitors of DAP10 and DAP12 or other proteins involved in this signaling pathway will be attractive therapeutic targets for the treatment of LGLL and other autoimmune diseases and syndromes.

Figure 1

CD8⁺ T cells from LGLL patients display enhanced lysis of a normal pulmonary artery endothelial cell line. (A) Percentage of CD56⁺, CD8⁺, and CD4⁺ T cells by flow cytometric analysis in unsorted and sorted populations after negative selection using RosetteSep. (B) CD8⁺ T cells were purified from LGLL patients and healthy blood donors. Direct cytotoxicity was performed using 5-hour ⁵¹Cr release assays. CD8⁺ T cells were incubated with ⁵¹Cr-labeled CRL-2598 endothelial cells at an E/T ratio of 50:1, and the CD8⁺ T-cell–mediated cytolysis of CRL-2598 cells was compared among LGLL patients (n = 9) and healthy controls (n = 9). (C) CD8⁺ T cells from LGLL patients are cytotoxic toward synovial cells. CD8⁺ T cells were purified from LGLL patients and healthy blood donors. Direct cytotoxicity assays were performed using 5-hour ⁵¹Cr release assays. CD8⁺ T cells were incubated with ⁵¹Cr-labeled HTB293 synovial cells at an E/T ratio of 50:1. CD8⁺ T-cell–mediated cytolysis of HTB293 cells was compared among 9 LGLL patients and 9 healthy controls. (D) CD8⁺ T cell–mediated cytolysis of CRL-2598 cells was compared among patients presenting with LGLL and PAH (n = 3), PAH without LGLL (n = 2), and healthy controls (n = 3). The mean and SEM of triplicate wells are shown.

Figure 2

Activated NK cells, but not freshly isolated NK cells, lyse CRL-2598 cells. NK cells isolated from healthy blood donors were either untreated (fresh) or treated with 100 U/mL IL-2 for 3 days and then tested for their ability to lyse ⁵¹Cr-labeled CRL-2598 cells or K562 (NK-sensitive target) tumor cells. (A) Fresh NK cells were used in direct cytotoxicity assays against K562 and CRL-2598 at the E/T ratios indicated. Results from 1 representative experiment are shown; results from 2 different control subjects yielded similar results (not shown). (B) Percentage specific lysis of CRL-2598 cells was compared between IL-2–treated and fresh NK cells using an E/T ratio of 50:1. Data include results from 3 healthy control subjects. The mean and SEM of triplicate wells are shown. (C) Percentage specific lysis of CRL-2598 cells by the activated NK-cell lines, NK92 or NKL, at E:T ratios of 20:1, 10:1, 5:1, and 2.5:1. The mean and SEM of triplicate wells are shown.

Figure 3

CD8⁺ T cells from LGLL patients maintain constitutively activated ERK1/2 and PI3K. CD8⁺ T-cell lysates were prepared from LGLL patients and healthy donors. The activity of ERK and PI3K in these cells was assessed by Western blot analyses using antibodies specific for (A) phosphorylated ERK1/2 (pERK) and (C) phosphorylated AKT. Membranes were stripped and reprobed with antibodies that detect (B) pan-ERK and (D) pan-AKT to demonstrate total protein levels.

Figure 4

A CD8⁺CD28^null phenotype and increased NKR expression characterizes T-LGLL cells. (A) CD8⁺ T cells isolated from LGLL patients and healthy donors were analyzed by 2-color flow cytometry for the cell surface expression of CD28. (B) Dot plots of CD8⁺/CD28⁺ expression are shown for 1 control and 1 LGLL patient. (C) The presence of activating NKRs on CD8⁺ cells were analyzed using flow cytometry. Results from 1 LGLL patient and 1 control are shown. Results from 10 patients are summarized in Table 1.

Figure 5

CD8⁺ T cells from LGLL patients display elevated levels of DAP10 and DAP12. Quantitative RT-PCR was used to determine the relative levels of (A) DAP12 and (B) DAP10 expression in CD8⁺ T cells from 5 LGLL patients compared with healthy donors (controls). The mean and SEM of 5 healthy donors and LGLL patients are shown.

Figure 6

Expression of dnDAP12 and/or dnDAP10 in CD8⁺ T cells from LGLL patients reduces adaptor signaling, cytotoxicity toward CRL-2598 cells, and granule redistribution. (A) Expression of dnDAP12 blocks Syk phosphorylation. AD293 cells were cotransfected with plasmids encoding Syk and 1 of the following FLAG-tagged DAP12 constructs: wild type (WT, lanes 2), DAP12^Y75A (lane 3), or DAP12^Y64A;Y75A (lane 4). Single transfections with Syk and WT DAP12 are included as negative controls (lanes 1, 5, and 6). After cross-linking with anti-Flag antibody, DAP12 was immunoprecipitated before Western blot analyses with an antiphosphotyrosine antibody (4G10). The blot was subsequently stripped and reprobed with an anti-Syk antibody confirming that the phosphorylated protein is Syk (data not shown). (B) Expression of dnDAP10 and dnDAP12 reduce the cytotoxic potential of CD8⁺ T cells from LGLL patients. The cytotoxic properties of genetically modified CD8⁺ T cells from LGLL patients were evaluated in 5-hour ⁵¹Cr release assays. CD8⁺ cells were infected with recombinant vaccinia virus encoding CD56 (negative control), dnDAP10, dnDAP12, or a combination of dnDAP10 and dnDAP12(Y75A), as well as dnMEK (positive control) and mixed at an E/T ratio of 50:1 with CRL-2598 target cells. Mock-infected CD8⁺ T cells from LGLL patients and healthy controls are included for comparison. Representative results from 1 LGLL patient are shown: cells from 5 other patients yielded similar results. The inhibitory effects of dnDAP10 and dnDAP12 occurred at E/T cell ratios ranging from 50:1 to 6:1 (data not shown). The mean and SEM of triplicate wells are shown. (C) Expression of dnDAP12 and/or dnDAP10 in CD8⁺ T cells from LGLL patients reduces cytotoxicity toward synovial cells (HTB293). Direct cytotoxicity was performed using 5-hour ⁵¹Cr release assays as described in Figure 7B. Results from 1 representative patient and healthy control are shown: similar results were obtained from 5 additional patients. The mean and SEM of triplicate wells are shown. (D) Target cell–induced granule redistribution is blocked in CD8⁺ T LGLL expressing dnDAP10 and/or dnDAP12. A proportion of the cells described in panel B were mixed at 1:1 E/T ratio and examined for granule redistribution. Cells were cytospun onto slides and lytic granules visualized with (i) a FITC-conjugated anti–granzyme B antibody (green). (ii) CRL-2598 target cells were prestained with Cell-Tracker Orange (red: Molecular Probes). DAPI was used for nuclear staining (blue) of both cell types. Results using untreated CD8⁺ T LGLL cells are shown at (iii) 0 and (iv) 10 minutes after mixing. Representative results using CD8⁺ T LGLL cells expressing (v) CD56 (negative control), (vi) dnDAP10, (vii) dnDAP12(Y75A), (viii) dnDAP10, and dnDAP12(Y75A) and (ix) dnMEK are shown. Results from 1 representative experiment are shown: results from 4 other patients yielded similar results (data not shown).

Figure 7

Expression of dnDAP12(Y75A) in PBMCs from LGLL patients reduces their production of chemokines. PBMCs from LGLL patients were infected with recombinant retrovirus as described in panel B and supernatants collected at 24 hours and analyzed for (A) RANTES and (B) IL-8 production by ELISA. Viability of PBMCs was confirmed to be more than 80% at time of ELISA analyses by annexin-V–FITC labeling. ELISA results are shown from 1 representative LGLL patient; PBMCs from 2 other patients were assessed yielding similar results. The mean and SEM of triplicate wells are shown.