The naturally processed CD95L elicits a c-yes/calcium/PI3K-driven cell migration pathway

Abstract

Patients affected by chronic inflammatory disorders display high amounts of soluble CD95L. This homotrimeric ligand arises from the cleavage by metalloproteases of its membrane-bound counterpart, a strong apoptotic inducer. In contrast, the naturally processed CD95L is viewed as an apoptotic antagonist competing with its membrane counterpart for binding to CD95. Recent reports pinpointed that activation of CD95 may attract myeloid and tumoral cells, which display resistance to the CD95-mediated apoptotic signal. However, all these studies were performed using chimeric CD95Ls (oligomerized forms), which behave as the membrane-bound ligand and not as the naturally processed CD95L. Herein, we examine the biological effects of the metalloprotease-cleaved CD95L on CD95-sensitive activated T-lymphocytes. We demonstrate that cleaved CD95L (cl-CD95L), found increased in sera of systemic lupus erythematosus (SLE) patients as compared to that of healthy individuals, promotes the formation of migrating pseudopods at the leading edge of which the death receptor CD95 is capped (confocal microscopy). Using different migration assays (wound healing/Boyden Chamber/endothelial transmigration), we uncover that cl-CD95L promotes cell migration through a c-yes/Ca²⁺/PI3K-driven signaling pathway, which relies on the formation of a CD95-containing complex designated the MISC for Motility-Inducing Signaling Complex. These findings revisit the role of the metalloprotease-cleaved CD95L and emphasize that the increase in cl-CD95L observed in patients affected by chronic inflammatory disorders may fuel the local or systemic tissue damage by promoting tissue-filtration of immune cells.

Figure 1. Cl-CD95L elicits CD95-CAP formation.

(A) Activated (PHA/IL2)-peripheral blood T-lymphocytes (PBLs) from healthy individuals were incubated with or without 100 ng/ml of cl-CD95L for 30 min. Cell morphology was followed using phase contrast microscopy and CD95 was stained using a mouse anti-CD95 mAb (APO1-3) followed by a goat Alexa488-coupled anti-mouse IgG mAb. Red arrows depict emitted pseudopods (Bars = 5 µm). (B) PHA/IL2-activated PBLs were incubated with the cytotoxic Ig-CD95L for indicated times, and cells were analyzed as described in (A). White arrows depict blebs emitted by apoptotic cells (bars = 5 µm). (C) The fibroblastic cell line PS120 devoid of endogenous CD95 (PS120^control) or reconstituted with human wild type CD95 (PS120^CD95) were untreated or treated with 100 ng/ml of cleaved CD95L for indicated times, and cell shape and CD95 distribution were analyzed.

Figure 2. Cleaved CD95L triggers a pseudopod-localized actin/PI3K/Akt signal and a Ca²⁺ entry through activation of the CRAC channel Orai1.

(A) H9 T-cells were transiently transfected with the actin marker Lifeact-GFP. Living cells were harvested (Ficoll) and treated or untreated for 30 min with 100 ng/ml of cleaved CD95L. Cells were fixed, and CD95 was stained using anti-CD95 mAb (APO1-3) and revealed with the secondary Alexa555-coupled Goat anti-mouse antibody (red). Nuclei were stained using DAPI (blue). Bars = 5 µm. Images were acquired with an ApoPLAN 63× objective. (B) The leukemic T-cell line H9 was incubated with 100 ng/ml of the naturally processed CD95L for indicated times. Cells were lysed, and for each lane, 100 µg of protein were loaded. Proteins were resolved by SDS-PAGE and anti-whole Akt and anti-Akt phosphorylation (serine⁴⁷³) immunoblots were performed. Akt phosphorylation stands for its activation. (C) The pleckstrin homology (PH) domain of Akt binds PIP3 produced by PI3K activation. As a consequence, the probe PH_Akt-GFP stains PIP3. H9 T-cells were transiently transfected with the PH_Akt-GFP construct (green). Living cells were isolated (Ficoll) and incubated with 100 ng/ml of the naturally processed CD95L (cl-CD95L) for 30 min. Cells were fixed, and CD95 was stained as previously mentioned (red). Pictures were taken by confocal microscopy. Cell morphology was followed using differential interference contrast microscopy. Nuclei were stained with DAPI (blue). Bars = 5 µm. (D) Calcium measurement in single cell. Jurkat T-cells were loaded with the permeant calcium probe Fura-2AM, and in parallel, CD95 was followed using an anti-CD95 mAb and an Alexa555-coupled goat anti-mouse mAb as described in Materials and Methods. Cells were bathed in a Ca²⁺-free medium and pre-incubated (lower panel; cl-CD95L) or not (upper panel; unstimulated) with cl-CD95L (100 ng/ml). 2 mM Ca²⁺ was added when indicated by the black-filled rectangle. Ratio images (F340 nm/F380 nm) were taken every 5 s, and the images shown correspond to the indicated period of time. Grey level intensities were translated to false colors according to the colors scale shown at the right, and [Ca²⁺]_i was estimated from the ratio values and calibration experiments as described in Materials and Methods. Red arrows indicate the CD95-CAP. (E) Jurkat T-cells were treated or untreated with 100 ng/ml of cl-CD95L for 60 min. Cells were fixed and Orai1 and CD95 were stained as described in Materials and Methods. Cell morphology was analyzed using phase contrast microscopy. (F) Left panels: GFP-, GFP-Orai1, or GFP-Orai1_E106A-expressing Jurkat T-cells were loaded with 1 µM of the calcium probe Fura-2AM for 30 min at RT. Cells were bathed at 37°C in a medium containing 2 mM [Ca²⁺]_e and then treated with 100 ng/ml of cl-CD95L (black arrow). GFP#84 corresponded to control cells expressing GFP. GFP-Orai1_E106A#155 and #169 corresponded to two independently isolated clones expressing the non-functional CRAC channel. GFP-Orai1#142 corresponded to Jurkat cells overexpressing human Orai1. Values were recorded every 10 s. Right panel: For each experiment, the area under the curve (AUC) was measured for 1,000 s, and the statistical analyses of the AUC values were performed for indicated cells using non-parametric two-tailed Mann-Whitney tests. ** and *** indicate p values≤0.01 and 0.001, respectively.

Figure 3. Cleaved CD95L promotes cell migration through a DD-dependent signal.

(A) The CD95-deficient PS120 and its counterparts expressing human wild type CD95 or DD-truncated CD95 were seeded in a Boyden chamber in the presence or absence of cl-CD95L (100 ng/ml) and incubated for 24 h. The filter was removed, the upper side containing the non-migrating cells was wiped out with cotton-tipped swabs, and migrating cells in the opposite side of the filter were fixed with methanol and stained (Giemsa). For each experiment, five pictures of random fields were taken and a representative picture was depicted (Bars = 70 µm). (B) Cells were treated as described in (A). To quantify cell motility, Giemsa-stained migrating cells from the lower side of the membrane were lyzed and absorbance was measured at 560 nm. (C) Wound healing assay, a confluent monolayer of the indicated cells was “wounded” with a tip and then cells were incubated for 24 h in the presence or absence of 100 ng/ml of cl-CD95L and pictures were acquired (Bars = 50 µm). Pictures are representative of 5 independently performed experiments.

Figure 4. cl-CD95L promotes a cell migration through a Ca²⁺/PI3K/Akt signal.

(A) The CD95-deficient fibroblastic cell line PS120 transfected with empty vector (PS120^control) or reconstituted with either human wild type CD95 (PS120^CD95) or a death domain-truncated CD95 (PS120^{CD95(Δ1-210)}) was incubated with 100 ng/ml of cleaved CD95L for indicated times. Cells were harvested, lyzed, and 100 µg of protein was loaded per lane. Proteins were resolved in a SDS-PAGE and immunoblots were performed. Phosphorylation of the serine⁴⁷³ on Akt indicates activation of the kinase. Whole Akt is added as loading control. (B) Upper panels: PS120^CD95 cells were pre-incubated with non-cytotoxic and non-cytostatic concentration of PI3K inhibitors (2.5 µM of wortmannin and 5 µM of LY294002), the cell permeant chelator of calcium BAPTA-AM (5 µM), or the inhibitor of IP3-R and SOC channels 2-APB (20 µM). Cells were then untreated (ctr) or treated for 5 min with 100 ng/ml of cl-CD95L and lysed, 50 µg/ml of protein was loaded per lane, and indicated immunoblots were performed. Lower panels: densitometry analyses of the immunoblot bands using ImageJ software and the histograms depict Phospho-Akt/whole Akt ratios. (C) Cell migration of PS120^CD95 was assessed using the Boyden chamber assay. Cells were pre-incubated with non-cytotoxic and non-cytostatic concentrations of PI3K inhibitors (2.5 µM of Wortmannin and 5 µM of LY294002), the cell permeant chelator of calcium BAPTA-AM (5 µM), or the inhibitor of SOC channels 2-APB (20 µM) for 30 min and then stimulated with 100 ng/ml of cl-CD95L for 24 h. For each experiment, 10 pictures of the migrating cells were taken and a representative picture was depicted (Bars = 50 µm). (D) Cells were treated as described in (C). To quantitatively measure cell motility, Giemsa-stained migrating cells from the lower side of the membrane were lyzed, and absorbance was measured at a wavelength of 560 nm. Values represent means and SD of three independently performed experiments. **p<0.01 and ***p<0.001 as calculated using two-tailed non-parametric Mann-Whitney test. (E) The silencing effect of the Orai1-targeting shRNAmir-pGIPZ vectors was analyzed by immunoblot in lentiviral-transduced HEK cells. 48 h after transduction, cells were lysed and 100 µg of lysates were loaded per line. β-actin serves as a loading control. (F) Cell migration of indicated shRNAmir-transduced HEK cells was assessed using Boyden chamber assay. Right panels: HEK cells were transduced with scrambled or Orai1-targeting ShRNA, and 48 h later, cells were incubated for 24 h in the presence or absence of 100 ng/ml of cl-CD95L. Migrating cells were fixed with methanol and stained by Giemsa. For each experiment, five pictures of random fields were taken and a representative picture was depicted (Bars = 70 µm). Left panel: To quantitatively measure cell motility, Giemsa-stained migrating cells from the lower side of the membrane were lyzed, and absorbance was measured at a wavelength of 560 nm. Values represent means and SD of three independently performed experiments. *** p<0.001 as calculated using two-tailed non-parametric Mann-Whitney test.

Figure 5. The src kinase c-yes orchestrates the cl-CD95L-mediated PI3K/Akt signaling pathway and cell migration.

(A) Upper panels: T-leukemic cell line H9 was transiently transfected with Lck-GFP. 24 h after transfection, living cells were treated or untreated for the indicated times in the presence of cl-CD95L (100 ng/ml). CD95 was stained using a mouse anti-CD95 mAb and a goat Alexa555-coupled anti-mouse IgG mAb. Lower panels: Activated PBLs were incubated for the indicated times with 100 ng/ml of cl-CD95L, and then cells were fixed. Lipid rafts were stained using Alexa488-cholera toxin B subunit and CD95 was followed as previously mentioned. Images were acquired with a confocal microscope with a 63× objective. Cell morphology was followed using phase contrast microscopy (Bars = 7.5 µm). The percentage of T-cells displaying a CD95 cluster was assessed (∼300 cells counted for each condition). Among the activated T-lymphocytes showing CD95 cluster, the amount of CD95-CAP co-localized with lipid rafts was assessed. (B) H9 T-cells were pre-incubated for 30 min with DMSO (U), 10 µM of PP2, 5 µM of LY294002 (LY), 2.5 µM of wortmannin (Wort), or 40 µM of zVAD-fmk (zV) and then treated or untreated for 15 min with 100 ng/ml of cl-CD95L. Cells were lyzed and Akt phosphorylation (S⁴⁷³) and whole Akt were assessed by immunoblots. (C) The H9 T-cell line (left panels) and activated PBLs (right panels) were incubated for indicated times with 100 ng/ml of cl-CD95L or APO1-3. Then, cells were lyzed and CD95 was immunoprecipitated. The immune complex was resolved in a 10% SDS-PAGE, and indicated immunoblots were performed. Total lysates were depicted to confirm that the same amount of protein was present for each immunoprecipitation. The black stars represent heavy chains of Ig. (D) Boyden chamber assays were performed as described in Materials and Methods. PS120^CD95 were pre-incubated with or without a non-cytotoxic and non-cytostatic concentration of PP2 (10 µM) for 30 min and then treated (cl-CD95L) or untreated (−) for 24 h with 100 ng/ml of cl-CD95L. Data are representative of three independently performed experiments. (E) Activated PBLs were pre-incubated with 10 µM of the src inhibitor PP2 or 5 µM of the PI3K inhibitor LY294002 (LY) and then incubated in the presence or absence of cl-CD95L (100 ng/ml) for 24 h. Then, endothelial transmigration of PBLs was assessed as described in Materials and Methods. (F) The silencing effect of the c-yes-targeting shRNAmir-pGIPZ vectors was analyzed by immunoblot in lentiviral-transduced PS120^CD95 cells. 72 h after transduction, cells were lysed and 100 µg of lysate was loaded per line. β-actin serves as loading control. (G) The T-cell line H9 was transduced with scramble (scr) or c-yes (shYes#3)-targeting shRNAmir-containing lentivirus. 72 h after transduction, living cells were harvested and green cells were sorted by flow cytometry using FACSAria. Cells were immediately stimulated or unstimulated with cl-CD95L (100 ng/ml), and the amounts of Akt phosphorylation (S⁴⁷³) were assessed by immunoblot. (H) Cell migration of indicated shRNAmir-transduced PS120^CD95 was assessed using the Boyden chamber assay. Left panels: 72 h after transduction, cells were treated (cl-CD95L) or untreated (mock) with 100 ng/ml of cl-CD95L for 24 h. For each experiment, five pictures of the migrating cells were taken and a representative picture was depicted (Bars = 50 µm). Right panel: To quantitatively measure cell motility, Giemsa-stained migrating cells from the lower side of the membrane were lyzed and absorbance was measured at a wavelength of 560 nm. Values represent means and SD of three independently performed experiments. (I) H9 cells were transduced as mentioned in (G), and the green cells were incubated in the presence or absence of cl-CD95L (100 ng/ml) for 24 h. Then, the endothelial transmigration of the sh-RNA-expressing T-cells was assessed as described in Materials and Methods.

Figure 6. CD95L found in SLE patients promotes endothelial transmigration of activated PBLs.

(A) By ELISA, the amounts of soluble CD95L in sera of healthy donors (control) and SLE patients were quantified. n stands for the number of SLE patients and healthy subjects. (B) Correlation between aggravation markers of the disease and the amount of soluble CD95L. (C) Activated PBLs were incubated for the indicated times with sera of healthy donors or SLE patients and the distribution of CD95 was analyzed by confocal microscopy. (D) A 24 h wound healing assay was performed. The fibroblastic cell line PS120^control or its CD95-expressing counterpart PS120^CD95 was grown to confluence. then the monolayer was “wounded” by scratching cells using a pipette tip. Using microscopy, the migration was monitored for 24 h with the indicated concentrations of cl-CD95L as indicated in Materials and Methods. Pictures are representative of 3 independently performed experiments. (E) Activated PBLs were incubated in the presence of sera from SLE patients (n = 3) or healthy donors (n = 6), and then, adhesion to endothelial cells (left panel) and endothelial transmigration (right panel) of PBLs were assessed as described in Materials and Methods. Where indicated, activated PBLs were preincubated 30 min with the antagonist anti-CD95 mAb ZB4 (α-CD95). ZB4 was maintained in the culture at 10 µg/ml.