Mycolactone impairs T cell homing by suppressing microRNA control of L-selectin expression

Abstract

Mycolactone is a macrolide produced by Mycobacterium ulcerans with immunomodulatory properties. Here, we describe that in mouse, mycolactone injection led to a massive T-cell depletion in peripheral lymph nodes (PLNs) that was associated with defective expression of L-selectin (CD62-L). Importantly, preexposure to mycolactone impaired the capacity of T cells to reach PLNs after adoptive transfer, respond to chemotactic signals, and expand upon antigenic stimulation in vivo. We found that mycolactone-induced suppression of CD62-L expression by human primary T cells was induced rapidly at both the mRNA and protein levels and correlated with the reduced expression of one miRNA: let-7b. Notably, silencing of let-7b was sufficient to inhibit CD62-L gene expression. Conversely, its overexpression tended to up-regulate CD62-L and counteract the effects of mycolactone. Our results identify T-cell homing as a biological process targeted by mycolactone. Moreover, they reveal a mechanism of control of CD62-L expression involving the miRNA let-7b.

Mycolactone is a diffusible macrolide produced by Mycobacterium ulcerans, the causative agent of a chronic skin disease called Buruli ulcer (BU) (1). Mycolactone production is critical for bacterial virulence and underpins the pathogenesis of lesions marked by a striking lack of inflammatory infiltrates. The immune profiling of patients has recently revealed systemic defects in the production of multiple T-cell cytokines, suggesting that mycolactone impairs the generation of cellular responses (2). In line with this hypothesis, mycolactone was reported to suppress the production of various cytokines by monocytes, lymphocytes, and dendritic cells (DCs) in vitro (3–6). The mechanism by which mycolactone modulates protein expression in a gene- and cell-specific manner nevertheless remains mysterious, as differently from known immunosuppressants, mycolactone acts independently of the mammalian target of rapamycin (mTOR).

Studies in mice have shown that s.c.-delivered mycolactone gains access to the leukocytes of the peripheral blood and lymphoid organs and triggers immune defects comparable with those observed in BU patients (7). The mouse model therefore appears appropriate to study the impact of mycolactone on cellular responses in vivo. Mycolactone blocks the maturation and migration of DCs in the mouse (5); however, whether it impairs the trafficking properties of T cells, the key players in the adaptive immune response to pathogens, is still unknown. In this process, the capacity of naive T cells to gain access to peripheral lymph nodes (PLNs) and make contact with antigen-presenting cells is of critical importance. T-cell homing to PLNs is controlled by the expression of the receptors l-selectin (CD62-L), C-C chemokine receptor 7 (CCR7), and lymphocyte function-associated antigen 1 (LFA-1). High expression of CD62-L on the surface of naive T cells is essential for the initial steps of tethering and rolling of circulating lymphocytes in the high endothelial venules (HEV) of peripheral LNs, whereas CCR7 and LFA-1 play a crucial role in the subsequent adhesion of T cells to HEV (8). Within the LNs, T cells then migrate to specialized T-cell zones, where CCR7-driven motility is vital for scanning antigens presented by DCs (9).

Here, we have used a combination of transgenic mouse models, LN explants, and human primary T cells to evaluate the impact of mycolactone on T-cell dynamics. We found that s.c.-delivered mycolactone modified the expression of CD62-L, CCR7, and LFA-1 on circulating T cells. These alterations had major consequences on the capacity of T cells to home to PLNs and on the intensity of cellular responses driven by antigen. We identified the let-7b microRNA (miRNA) as a regulatory component of CD62-L expression in human T cells and provided a mechanistic explanation for how mycolactone regulates CD62-L expression by altering let-7b levels. Our results uncover a unique biological property of mycolactone through the regulation of T-cell trafficking and demonstrate that CD62-L expression is partially regulated by the let-7b miRNA.

Results

Mycolactone Induces T-Cell Depletion in PLNs.

In mice injected with mycolactone via the s.c. route, we observed a macroscopic and dose-dependent reduction in the size of all PLNs (as illustrated in Fig. 1A by inguinal LNs). This effect was maximal after 24 h and persisted for a few days (>5 d with 100 μg of mycolactone). Histological analysis of the PLNs showed a massive cellular depletion affecting both the B- and T-cell zones (Fig. 1B, a and b). In contrast, the HEV through which lymphocytes leave the blood to enter PLNs showed unaltered levels of peripheral node vascular adressin (PNAd), the ligand of CD62-L (Fig. 1B, c). Moreover, the overall structure of the PLNs, as visualized by staining of the stromal matrix with the ER-TR7 fibroblast marker, was preserved (Fig. 1B, d). Focusing on T cells, we found that CD4⁺ and CD8⁺ lymphocytes were comparably depleted in PLNs (Fig. 1C). A concomitant increase in CD4⁺ and CD8⁺ blood T cells was observed. Since mycolactone displays cytopathic effects in vitro (10), we assessed T-cell viability in mycolactone-injected mice. Confirming our previous studies showing that primary T cells are relatively resistant to mycolactone cytotoxicity (6), no evidence of enhanced apoptosis could be demonstrated by flow-cytometric analysis of T cells in peripheral blood, PLNs and spleen, or by staining of LN sections (Fig. S1). Together, these data suggested that mycolactone affects T-cell homing to PLNs.

Fig. 1.

Mycolactone induces lymphocyte depletion in PLNs. (A) Macroscopic view of inguinal LNs of mice 24 h after s.c. injection of 50 or 100 μg of mycolactone, or vehicle as control, at the base of the tail. (B) Hematoxilin/eosin (a), B220 + CD3 (b), PNAd + CD3 (c), and ER-TR7 + CD3 (d) staining of PLNs of mice injected with 50 μg of mycolactone (Lower) or vehicle as control (Upper). Photos in A and B are representative of at least two independent experiments. (C) CD4⁺ and CD8⁺ T cells counts in inguinal LNs and peripheral blood of mice 24 h after s.c. injection of 100 μg of mycolactone or vehicle (Ctrl). Data are mean cell counts ± SEM of two independent experiments using three mice per group.

Mycolactone Modulates the Expression of Homing Receptors by T Cells.

We thus examined the impact of mycolactone on T-cell expression of homing receptors. In mice injected with mycolactone, the expression of CD62-L by blood CD4⁺ and CD8⁺ T cells was markedly suppressed (Fig. 2A). Maximal decrease was reached after 14 h, after which T-cell concentration in the blood increased rapidly (Fig. 2A). Twenty-four hours after injection, the expression of CD62-L by LN T cells was also largely reduced (Fig. 2B and Fig. S2A). LFA-1 levels were also impaired, but to a lesser extent. In contrast, T-cell expression of CCR7 was up-regulated by mycolactone. To demonstrate that mycolactone was directly responsible for these phenotypic changes, T cells purified from PLNs were cultured in the presence of mycolactone in vitro. The same characteristic CD62-L^lo CCR7^hi LFA-1^lo T-cell phenotype was induced (Fig. 2C and Fig. S2B). Maximal effects were observed at mycolactone concentrations >50 ng/mL, in both the CD4⁺ and CD8⁺ T cells (Fig. S2C).

Fig. 2.

Mycolactone modulates the expression of T-cell homing receptors in vivo and in vitro. (A) Kinetic analysis of the CD62-L expression and concentration of blood CD4⁺ and CD8⁺ T cells after the s.c. injection of 100 μg of mycolactone. Data are mean fluorescence intensities (MFI) and mean cell counts ± SEM from three mice, expressed as percent of controls. They are representative of two independent experiments. (B) Expression of CD62-L, CCR7, and LFA-1 on CD4⁺ and CD8⁺ PLN T cells of mice injected with 100 μg of mycolactone or vehicle (Ctrl). Data are MFI ± SEM measured 16 h after s.c. injection from five mice, expressed as percent of controls. (C) Expression of CD62-L, CCR7, and LFA-1 on PLN T cells cultured for 16 h in the presence of 100 ng/mL mycolactone, or vehicle (Ctrl). Data are MFI ± SEM of triplicates, expressed as percent of controls. Experiments in B and C were repeated three times independently with comparable results. *P < 0.05; **P < 0.01; ***P < 0.001; NS, not significant.

Mycolactone-Treated T Cells Have Impaired Homing Capacity and Weaker CCR7-Driven Motility.

The impact of mycolactone on the trafficking properties of T cells was further investigated in adoptive transfer experiments. Here, naive T cells were purified from donor mice and exposed to mycolactone, or vehicle as control, for 16 h. Mycolactone-treated cells and controls were then labeled with two distinct concentrations of 5-chloromethylfluorescein diacetate (CMFDA) to distinguish the two cell populations, mixed, and injected i.v. into recipient mice. After 4 h, cell suspensions were prepared from PLNs and spleen, and the CMFDA fluorescence was analyzed by FACS. Consistent with their lack of CD62-L expression, T cells exposed to mycolactone before transfer were recovered in the PLNs at lower frequencies than unexposed T cells (Fig. 3A). In comparison, the impact of mycolactone on T-cell recruitment to the spleen was of minor importance. Since CD62-L is not critical for T-cell homing to the spleen (11), this observation suggested that mycolactone primarily targets CD62-L–dependent homing processes.

Fig. 3.

Mycolactone-treated T cells have impaired homing capacity and reduced motility response to CCR7. (A) T cells purified from mouse PLNs were cultured in the presence of 100 ng/mL mycolactone (Myco) or vehicle (Ctrl) for 16 h, loaded with two different concentrations of CMFDA, then mixed in equal numbers and adoptively transferred into recipient mice. Total counts of mycolactone-treated T cells are compared with that of controls (100%) in PLNs and spleen of recipients 3.5 h after transfer. Data are mean cell counts ± SEM from three experiments using three recipient mice. (B) PLN T cells were incubated for 16 h with 100 ng/mL mycolactone (Myco) or vehicle (Ctrl), then labeled with two different fluorescent dyes and overlaid on LN slices. Data are mean motility and velocity coefficients ± SEM from three independent experiments, with >100 cells analyzed per experiment. *P < 0.05; **P < 0.01; NS, not significant.

Mobility of naive T cells within LNs largely depends on CCL19 and CCL21, the two ligands of CCR7 (12–14). Having shown that mycolactone up-regulates the expression of CCR7 on naive T cells, we investigated whether mycolactone-exposed T cells had an increased responsiveness to CCR7 stimulation. T cells purified from mouse PLNs were incubated in vitro with mycolactone, or vehicle as control. We then stained the cells with two distinct fluorescent dyes, loaded them onto LN slices kept alive in a warm oxygenated medium, and recorded their movements (Movie S1). We reported that in these conditions, naive T cells adopt the motile behavior of resident cells, which largely depends on CCR7 signaling (9). This motility has been shown to be a chemokinesis (random) more than a chemotaxis (gradient directed) walk mechanism, which can be assessed by measuring the motility (average speed) and the velocity (scanned surface) coefficients. Both parameters were reduced in mycolactone-exposed T cells compared with controls (Fig. 3B). Mycolactone also reduced the motility of T cells plated on an ICAM-1 layer and stimulated with CCL19 (Fig. S3A). However, it did not affect significantly T-cell migration toward CCL19 in transwell assays (Fig. S3B), suggesting that mycolactone impairs CCR7-driven T-cell motility by altering chemokinesis rather than chemotaxis.

Mycolactone Impairs T-Cell Response to Antigen in Vivo.

Mycolactone-exposed T cells had an impaired capacity to home to PLNs and a reduced motility within these tissues, both processes being essential for efficient T-cell responses. To measure the impact of mycolactone treatment on the ability of T cells to mount primary responses, we used CD45.1⁺ CD45.2⁺ or CD45.2⁺ OT-I CD8⁺ T cells, which both express a TCR specific for the SIINFEKL peptide of ovalbumin (OVA). Once purified from PLNs, these cells were labeled with carboxyfluorescein succinimidyl ester (CFSE) and incubated for 16 h with mycolactone or vehicle, respectively. After this treatment, CD45.1⁺ CD45.2⁺ OT-I T cells showed the characteristic CD62-L^lo CCR7^hi phenotype observed for mycolactone-treated wild-type T lymphocytes. We then mixed the two OT-I T-cell populations, using larger inputs of mycolactone-treated cells so that comparable numbers of control and mycolactone-exposed cells reached the draining LNs, and injected them i.v. into normal recipient mice. Concomitantly, mice were given a s.c. injection of OVA in the footpad. The proliferation of adoptively transferred T cells was then analyzed in draining LNs 48 h after immunization. Fig. 4A shows that mycolactone-treated OT-I T cells reaching the LNs divided actively in response to antigen. However, their CFSE dilution profile showed a delay in proliferation, as evidenced by the lower number of mycolactone-exposed cells in the last division peak compared with control cells (Fig. 4B). Therefore, mycolactone limits the intensity of host T-cell responses to antigen primarily by inhibiting their homing properties.

Fig. 4.

Mycolactone impairs T-cell response to antigen in vivo. Purified CD45.1⁺ CD45.2⁺ OT-I T cells or CD45.2⁺ OT-I T cells were stained with CFSE then exposed in vitro to mycolactone (Myco) or vehicle (Ctrl) for 16 h, respectively. Cells were mixed at a ratio of 5:1 (Materials and Methods) and adoptively transferred into CD45.1-recipient mice, which were injected concomitantly into the hind footpad with OVA. (A) A representative CFSE profile is shown for CD45.1⁺ CD45.2⁺ (Myco, black) and CD45.2⁺ (Ctrl, gray) positive LN T cells 48 h after immunization, with peak labels indicating the successive generations. (B) Mean percent dividing cells ± SEM in each cycle, calculated from four mice. Independent experiments performed with Myco and Ctrl cells delivered at a ratio of 1:1 gave comparable results. **P < 0.01.

Mechanism of Mycolactone-Induced Suppression of CD62-L.

Having characterized the consequences of mycolactone-induced suppression of CD62-L on the functions of T cells in the mouse system, we examined whether the same effects applied to human cells and explored the underlying molecular mechanism. The surface expression of CD62-L by CD4⁺ T cells purified from human peripheral blood was reduced after 20 h of mycolactone treatment (Fig. 5A). This reduction was not due to intracellular accumulation of the receptor, because CD62-L suppression was detected in permeabilized cells after only 4 h (Fig. 5A). Since CD62-L expression is negatively regulated by proteolytic cleavage from the cell surface, we measured the concentration of soluble CD62-L (sCD62-L) in culture supernatants. Compared with controls, mycolactone-treated cells released lower amounts of sCD62-L, indicating that mycolactone does not suppress CD62-L expression by promoting extracellular shedding of the receptor (Fig. S4A). To determine whether mycolactone instead promoted the intracellular elimination of CD62-L, we used MG132, an inhibitor of proteosome and lysosome functions. MG132 did not restore higher levels of the receptor in cells pretreated with mycolactone for 20 h, ruling out this possibility (Fig. S4B). These results suggested that mycolactone modulates the rate of CD62-L production rather than its elimination. A transcriptional analysis revealed that CD62-L mRNA were slightly less abundant in cells treated with mycolactone than in controls (Fig. 5B). This inhibition was detectable after 4 h and persisted until 20 h. Because CD62-L gene expression is controlled in human T cells by the Forkhead box O 1 (FOXO1)/Kruppel-like factor 2 (KLF2) transcriptional axis, the expression of KLF2 was also examined (15–17). KLF2 was not modulated in mycolactone-exposed T cells (Fig. 5B), suggesting that mycolactone down-regulates CD62-L transcription by a FOXO1/KLF2-independent mechanism.

Fig. 5.

Mechanism of mycolactone-induced suppression of CD62-L. (A) Time-dependent modulation of CD62-L total and surface expression, as measured by flow cytometry on permeabilized or nonpermeabilized cells, by CD4⁺ T cells exposed to 100 ng/mL mycolactone or vehicle. Data are MFI ± SEM of triplicates, expressed as percent of controls, and are representative of >3 donors. (B) Effect of mycolactone on CD62-L and KLF2 transcript abundance in CD4⁺ T cells. Data are mean fold changes ± SEM from three donors, after cell treatment with 100 ng/mL mycolactone, compared with controls by random effect models (Materials and Methods). (C) Effect of mycolactone on the expression of let-7b by T cells. Data are mean fold changes ± SEM from Jurkat cells (n = 4) or PBLs (three donors) treated with 100 ng/mL mycolactone for 30 min or 4 h, respectively, compared with controls. (D) Expression of CD62-L mRNA in PBLs transfected with an anti-let-7b or an irrelevant anti-miR-Ctrl for 24 h. Data are mean fold changes ± SEM from three donors. (E) Expression of CD62-L mRNA in PBLs transfected with pre-Mir-let-7b for 24 h (let-7b), then incubated with 100 ng/mL mycolactone for 4 h (Myco + let-7b). Controls are cells transfected with an irrelevant premiR for 24 h, then incubated with mycolactone (Myco) or vehicle (Ctrl). Data are fold changes from >5 donors, compared with premiR-transfected controls by Wilcoxon matched pairs signed rank sum test, presented as box and whiskers. *P < 0.05; **P < 0.01; ***P < 0.001; NS, not significant.

To determine whether mycolactone modified the expression of other potential regulators of CD62-L transcription, T cells were subjected to a global transcriptomic analysis. However, no significant alterations in mRNA expression could be identified. Because microRNA (miRNAs) can regulate gene transcription and translation, we tested the possibility that mycolactone uses this mechanism to repress CD62-L expression, by a microarray approach. miRNAs from Jurkat T cells, a CD4⁺ T-cell line, were analyzed after 30 min of exposure to mycolactone, because this time point was optimal for CD62-L mRNA suppression in this model. Of the 377 miRNAs tested, 149 could be detected. Among the 18 miRNAs whose levels changed upon mycolactone treatment by more than twofold, the greatest alteration was in let-7b expression (Fig. S5). Mycolactone-induced down-regulation of let-7b was observed consistently in Jurkat cells (Fig. 5C). Importantly, let-7b was also suppressed by mycolactone in human primary CD4⁺ T cells exposed to mycolactone for 4 h (Fig. 5C), and in mouse LN T cells treated for 1 h, establishing the inhibitory action of mycolactone on this miRNA. We then asked whether let-7b interfered with the expression of CD62-L by transfecting T cells with either a let-7b inhibitor or a premiR-let-7b precursor. Here, experiments were performed in Jurkat cells or PBLs, because mRNA recovery from primary CD4⁺ T cells after transfection was limiting. Notably, silencing of let-7b resulted in the down-regulation of CD62-L mRNAs (Fig. 5D). Although statistical significance was not reached because of marked differences in CD62-L mRNA values among donors, overexpression of let-7b induced a trend toward up-regulation of CD62-L gene expression (Fig. 5E), counteracting the inhibitory effect of mycolactone. From these results, we concluded that mycolactone suppresses CD62-L expression in T cells by inhibiting let-7b–mediated positive regulation of CD62-L.

Discussion

Mycolactone was known to block the capacity of primary T cells to produce multiple cytokines upon activation (2, 6). Here, we show that mycolactone also impairs their migratory properties. Mycolactone imprinted a characteristic CD62-L^lo CCR7^hi signature on T cells in vitro and in vivo, leading to defective homing to PLNs. The same phenotypic alterations were observed in B cells. Consistent with the fact that CD62-L expression controls B-cell entry in LNs (18), these cells were also depleted from PLNs in mycolactone-injected mice (Fig. 1B).

Optimal clonal expansion of T cells requires that they migrate inside secondary lymphoid organs and travel within the T-cell zone to encounter the antigens presented by DCs. In agreement with our demonstration that both mechanisms, namely homing and intranodal motility, were inhibited by mycolactone, a reduced number of T-cell divisions was observed in our OT-I transfer experiments (Fig. 4). Within LNs, T-cell motility is mostly triggered by CCR7 and its ligands CCL19 and CCL21 (9, 19, 20). Our observations that mycolactone affected CCL19-induced motility in T cells loaded onto an ICAM-1 layer but not in transwell migration assays (Fig. S3) suggest that it may a have a selective impact on random motility.

Mycolactone modulated the expression of CD62-L at the transcriptional level. Recent studies have identified the transcription factor FOXO1 as a master regulator of CD62-L transcription, likely via the expression of KLF2. In quiescent T cells, active FOXO1 in the nucleus maintains expression of KLF2, which regulates the expression of CCR7 and CD62-L (15–17, 21–26). However, mycolactone reduced the transcription of CD62-L without modulating the expression of KLF2 (Fig. 5B). Moreover, it modulated CD62-L and CCR7 in opposite directions, whereas both receptors are positively regulated by the FOXO1/KLF2 axis. Because mycolactone triggers the activation of Lck in primary T cells (6), we also examined whether Lck somehow participated to mycolactone-induced suppression of CD62-L. However, mycolactone suppressed CD62-L expression in Lck-deficient Jurkat T cells, ruling out this possibility. Our study shows that other pathways exist that regulate the expression of homing receptors in T cells and identifies let-7b as a factor contributing to the control of CD62-L expression.

Let-7b belongs to a highly conserved miRNA family, which in humans contains let-7a-g/I and the related miR-98 and miR-202 (27). The LIN28 protein in Caenorhabditis elegans and its homologs in mammalian cells LIN28 and LIN28B have been shown to control the temporal expression of all let-7 family members by binding to their terminal loop, thereby blocking the processing of primary transcripts into mature miRNAs (28). Our microarray analysis showed that mycolactone reduces the expression of let-7b without affecting the other members of the family, suggesting that mycolactone-mediated repression does not proceed via LIN28/LIN28B. miRNAs generally function as gene silencers repressing translation and/or directing the sequence-specific degradation of complementary mRNA. They were found in some instances to enhance gene expression by interacting with promoter sequences (29). We found that let-7b regulated the expression of CD62-L positively. Since we failed to identify a sequence complementary to let-7b in the CD62-L gene and flanking regions, it seems unlikely that let-7b activates CD62-L transcription directly and rather modulates the expression of transcriptional and/or translational regulators of CD62-L. Proteomic studies have shown that variations in let-7b levels finely tune the production of thousands of proteins (30). Hence, down-regulation of let-7b by mycolactone is likely to trigger a cascade of events leading to the modulation of CD62-L expression and, potentially, other factors not studied here.

Through the study of CD62-L, our work elucidates a molecular mechanism by which mycolactone modulates mammalian gene expression. Mycolactone has been shown to block the expression of cytokines and chemokines by human leukocytes without major impact on gene transcription (3, 6). Overexpression of let-7b in T cells did not restore their capacity to produce IL-2 upon activation, suggesting that mycolactone might suppress the expression of target genes that are induced during cell activation by modulating other miRNAs. Keeping in mind that mycolactone does not induce a deep alteration of gene transcription, as assessed by mRNA microarray, a detailed analysis of the miRNA transcriptome of cells activated in the presence of mycolactone should allow one to better understand the molecular pathways targeted by this toxin. Understanding the impact of mycolactone-mediated modulation of miRNA expression will not only help understand the pathogenesis of BU disease, but also provide a paradigm for how miRNAs modulate immune responses.

Materials and Methods

Animals.

Six-week-old female C57Bl6/J (B6) mice were purchased from The Jackson Laboratory. CD45.2⁺ OT-I and CD45.1⁺ mice were obtained from Charles River and bred under SPF conditions. F₁ mice (CD45.1⁺ CD45.2⁺ OT-I) transgenic mice were produced by crossing CD45.1⁺ B6 with CD45.2⁺ OT-I. All experiments were performed in accordance with the guidelines of the National French Veterinary Department.

Mycolactone.

Mycolactone was purified from M. ulcerans 1615 (ATCC 35840) cultures as described and stock solutions kept in ethanol (7). For injections, these stocks were diluted 10× in PBS and a 150-μL volume injected s.c. at the base of the tail. For cellular assays, they were diluted at least 1,000× in culture medium. In all cases, mycolactone-treated samples were compared with vehicle-treated controls.

T-Cell Isolation and Culture.

Whole PLN T cells from B6 mice or CD8⁺ T cells from OT-I mice were negatively purified by using the Pan mouse T-cell isolation kit (Miltenyi Biotec) or the mouse CD8 isolation kit (Invitrogen), respectively. Human primary T cells were isolated from blood donors by Ficoll density gradient centrifugation. When required, CD4⁺ T-cell purification was performed by negative depletion (Dynal Biotech). A Jurkat cell variant JCaM1 (referred to as Jurkat in this study) was used in miRNA studies. T cells were cultured in RPMI medium 1640 (Seromed) + 10% FCS (Dominique Dutscher), supplemented with 10 ng/mL rhIL-7 in the case of mouse T cells (R&D Systems).

Flow Cytometry and Histological Analyses.

FITC or PE-conjugated CD62-L, PerCP-conjugated CD4, APC-conjugated CD8 were from Becton Dickinson. PE-conjugated CCR7 and biotin-conjugated CD127 were from eBioscience, and Alexa-Fluor 488-conjugated streptavidin from Invitrogen. Cells were analyzed with a FACScalibur flow cytometer (Becton Dickinson). For hematoxylin and eosin staining, LNs were fixed in 4% paraformaldehyde in PBS, paraffin-embedded, and cut on 4-μm-thick sections. After deparaffinization, sections were incubated in hematoxylin then differentiated by using a lithium carbonate solution. Next, slides were incubated in eosin, rinsed, and finally dehydrated with ethanol. For immunostaining, PLNs were embedded in Optimal Cutting Temperature (OCT) compound (Tissue-Tek) and frozen into liquid nitrogen-chilled isopentane. Blocks were sliced into 7-μm-thick sections by using a Leica cryostat (Leica CM 3050S) and sections attached to Superfrost PLUS glass coverslips (Thermo Scientific). They were then washed in PBS + 3% BSA and stained overnight at 4 °C with unconjugated anti-CD3 and biotinylated anti-B220 or biotinylated anti-PNAd or unconjugated anti-ER-TR7 antibodies. Alexa Fluor 488-conjugated streptavidin (Invitrogen), a DyLight 594 Goat Anti-Armenian Hamster antibody (Jackson ImmunoResearch), and an Alexa Fluor 488 anti-Rat IgG were used as secondary reagents. In situ cell death detection was performed on PLN sections as recommended (in situ cell death detection kit; Roche).

Adoptive Transfer and in Vivo Proliferation Assays.

For adoptive transfer experiments, control or mycolactone-treated T cells (10 × 10⁶/mL) were loaded with 2 or 200 nM CMFDA (Molecular Probes) in PBS at 37 °C for 5 min, washed twice in PBS, mixed (2 × 10⁶ of each), and injected i.v. into recipients. For in vivo proliferation assays, T cells (20 × 10⁶/mL) purified from PLNs of CD45.1⁺ CD45.2⁺ OT-I mice and CD45.2⁺ OT-I mice were labeled with 2 μM of CFSE (Molecular Probes) 7 min at 37 °C in PBS, washed twice in complete medium and cultured overnight with or without mycolactone, respectively. The two cell populations were then mixed (2.5 × 10⁶ CD45.1⁺ CD45.2⁺ and 0.5 × 10⁶ CD45.2⁺ OT-I cells) and injected i.v. into CD45.1 recipients. Four hundred micrograms of OVA in saline was injected at the same time in the hind footpad. Mice were killed 48 h later, and draining LNs were analyzed for T-cell proliferation.

T-Cell Motility Analyses.

For video imaging of T cells in lymphoid tissues, LN slices were prepared from PLNs as described (9). Purified PLN T cells (2 × 10⁶/mL) were incubated overnight with 100 ng/mL mycolactone or vehicle as control, then stained with 0.5 μM Cell Trace calcein green or red (Molecular Probes). Labeled-T cells (10⁵ cells in 20 μL RPMI medium 1640 + 10% FCS) were then deposited onto the cut surface of each slice. Imaging was performed with an inverted microscope at 37 °C. Stacks of four sections (z step = 13.3 μm) were acquired with the MetaView software (Universal Imaging) every 20 s for 20 min. Cell motility was analyzed with the Imaris software (version 5.7; Bitplane). For in vitro analyses of T-cell motility, glass coverslips were coated overnight at 4 °C with 3 μg/mL human ICAM-1–Fc (R&D Systems), washed with PBS, and blocked with PBS containing 1% BSA for 30 min at 37 °C. T cells prepared as above were plated on the ICAM-1 layer and stimulated with 100 ng/mL recombinant CCL19 (PeproTech). Images were acquired every 4 s by using an inverted microscope equipped with a 60× objective and the MetaView software and analyzed with Imaris. For transwell assays, PLN T cells were cultured for 18–24 h at 2 × 10⁶ cells per mL in the presence of 100 ng/mL mycolactone. T cells (3 × 10⁵) in RPMI medium 1640/0.5% BSA were then placed in the upper chambers of 24-well transwell plates with 5-μm pore membrane insert (Corning), with 100 ng/mL of CCL19 in the lower chamber, and allowed to migrate for 3 h at 37 °C. Cells in the lower chamber were then mixed with an equal amount of flow check fluorospheres (Beckman Coulter) and quantified by flow cytometry.

Quantitative Real-Time PCR (Q-PCR).

Total RNA was extracted from human CD4⁺ T cells with the Qiagen RNeasy Mini Kit and digested with TURBO DNase (Ambion) for 15 min at 37 °C. First-strand cDNA was synthesized from 2 μg of total RNA with the high capacity cDNA reverse transcription kit (Applied Biosystems). Expression was quantified with the TaqMan Gene Expression Assay (Applied Biosystems) and specific primers (Table S1). Amplification was performed in triplicate, from 30 ng of cDNA template in a final volume of 20 μL in a 96-well PCR plate. Amplification conditions were 2 min at 50 °C, 10 min at 95 °C, followed by 40 cycles of 15 s at 95 °C and 1 min at 60 °C with the ABI 7300 Sequence Detection System (Applied Biosystems). Results were normalized with the 2-ΔΔCt method by using L32 as a reference.

miRNA Analysis.

Total RNA was extracted from T cells with the mirVana miRNA Isolation Kit (Applied Biosystems) and analyzed for miRNA expression with the TaqMan Human MicroRNA A Array (Applied Biosystems). In brief, 900 ng of RNA was reverse-transcribed into cDNA by using Megaplex RT primers (Applied Biosystems) and loaded onto a 384-well TaqMan MicroRNA Array according to the manufacturer’s protocol. Quantitative analysis was performed with the ABI Prism 7900HT Sequence Detection System (Applied Biosystems). The let-7b miRNA was assessed with the TaqMan miRNA assay (Applied Biosystems), using the RNU48 endogenous control for normalization. Induction of let-7b miRNA was performed by electroporation at 300 V and 500 μF (Gene Pulser Xcell Electroporation system, Bio-Rad) in Opti-MEM medium (Invitrogen), with 50 nM of premiR let-7b precursor or an irrelevant premiR as control (Ambion). Silencing of let-7b was performed by electroporation with 50 nM anti-let-7b inhibitor or an irrelevant anti-miR as control (Ambion).

Statistics.

Unless otherwise specified, comparison between mean expression values of markers were assessed by Student t test: *P < 0.05; **P < 0.01; ***P < 0.001; NS, not significant. When the donor effect was found to be important (i.e., there were marked differences in expression values of markers among human donors), random effect models allocating common intercepts for observations using the same donor cells were used (Fig. 5B).