Report

Helicobacter pylori Vacuolating Cytotoxin Inhibits T Lymphocyte Activation

See allHide authors and affiliations

Science  22 Aug 2003:
Vol. 301, Issue 5636, pp. 1099-1102
DOI: 10.1126/science.1086871

Abstract

Helicobacter pylori (Hp) vacuolating cytotoxin VacA induces cellular vacuolation in epithelial cells. We found that VacA could efficiently block proliferation of T cells by inducing a G1/S cell cycle arrest. It interfered with the T cell receptor/interleukin-2 (IL-2) signaling pathway at the level of the Ca2+-calmodulin–dependent phosphatase calcineurin. Nuclear translocation of nuclear factor of activated T cells (NFAT), a transcription factor acting as a global regulator of immune response genes, was abrogated, resulting in down-regulation of IL-2 transcription. VacA partially mimicked the activity of the immunosuppressive drug FK506 by possibly inducing a local immune suppression, explaining the extraordinary chronicity of Hp infections.

Helicobacter pylori (Hp) causes chronic gastritis, peptic ulcer disease, and is an early risk factor for gastric cancer (1). To establish a chronic infection, Hp possibly evades host responses through the inhibition of antigen-specific T cell proliferation (2). Major histo-compatibility complex (MHC) class II–restricted, cell-mediated mechanisms control Hp infection (3), and adoptive transfer of CD4+ T cells demonstrated a protective role of T cells (4). Gastroduodenal disease depends on the expression of major bacterial virulence factors, such as the vacuolating cytotoxin (vacA) or the cag pathogenicity island (cag-PAI). The cag-PAI encodes a type IV secretion system, injecting the bacterial protein CagA into gastric epithelial cells, or professional phagocytes (5, 6). CagA eventually becomes tyrosine-phosphorylated and activates Ras signaling via the Raf/Mek/Erk stress kinase pathway to cause epithelial cell proliferation and scattering (7). VacA, a secreted 95 kD protein, varies in the signal sequence (s1a, s1b, s1c, s2) and/or its middle region (m1, m2) between different Hp strains (8). Among other functions, VacA selectively inhibits the invariant chain (Ii)–dependent pathway of antigen presentation mediated by the MHC class II (9) and might induce apoptosis in epithelial cells (10).

To investigate Hp–T cell interactions, Jurkat T cells were infected with different Hp wild-type (wt) strains and isogenic mutants lacking cagA, vacA, or both genes (table S1). A stimulation of bacteria-infected T cells [multiplicity of infection (MOI) of 30] with phytohemagglutinin/phorbol myristate acetate (PHA/PMA) resulted in a strong reduction of Jurkat T cell proliferation (40 to 60%), compared with the stimulated but uninfected control (Fig. 1A). Also Campylobacter jejuni C64, producing the cytolethal distending toxin (CDT) known to inhibit cell cycle progression in epithelial cells (11), caused a pronounced proliferation inhibition. But neither C. jejuni C63, with a defective cdt operon, nor Escherichia coli HB101 showed comparable effects (Fig. 1A). Fresh human peripheral blood lymphocytes (PBLCs) reacted even more dramatically. Activated Hp-infected PBLCs were reduced in their proliferation to the level of resting (nonstimulated) PBLCs (Fig. 1B). Use of CD4+ T cell subpopulations purified from PBLC resulted in similar data.

Fig. 1.

Hp blocks T cell proliferation by two activities. (A) Influence of different Hp strains (m1 and m2 genotype) and isogenic mutants and of C. jejuni or E. coli HB101 on the proliferation of Jurkat T cells. T cells were activated with PHA+PMA. Jurkat proliferation was measured by the CellTiter-Glo Luminescent cell Viability Assay (Promega, Mannheim, Germany). (B) Same experiment as (A), but with human peripheral blood lymphocytes (PBLC) instead of Jurkat cells. (C) Effect of CCS of Hp P12 and Tx30a, C. jejuni strains C63 and C64 and E. coli HB101 (CCS+), and P12ΔvacA strain (CCS) on Jurkat T cell proliferation. (D) Same as (C), but with human PBLCs as target cells. Bact., bacteria.

Preparation of supernatants [concentrated bacterial culture supernatant (CCS)] from genetically defined isogenic Hp strains producing (CCS+) or not producing VacA (CCS) (Fig. 1, C and D) demonstrated that CCS+, but not CCS, blocked the proliferation of Jurkat cells (Fig. 1C). Purified VacA reacted similarly to CCS+ (fig. S2). Hp P12 secreted a s1/m1 VacA, which was more efficiently produced, compared with the s2/m2 VacA of Tx30a (Fig. 1, C and D) (8). The possibility that the observed blockage in T cell proliferation by Hp was the result of T cell apoptosis was excluded by a caspase-3–based apoptosis assay (fig. S1, A and B). Thus, Hp produces two anti-proliferative activities for T cells: one bacteria-associated factor present in all Hp strains (Fig. 1, A and B) and secreted VacA (Fig. 1, C and D).

For efficient lymphocyte activation and proliferation, secretion of IL-2 and surface localization of the high-affinity IL-2 receptor are necessary. Hp m1-VacA–producing strains [P12, P12ΔPAI, 26695; American Type Culture Collection (ATCC) 43526] or purified VacA (fig. S2, A and B), but not a P12ΔvacA strain, severely reduced IL-2 secretion in Jurkat T cells (60 to 80%) (Fig. 2A). Furthermore, m2-VacA–producing strains (Tx30a, SS1, P49) as well as C. jejuni controls did not dramatically affect IL-2 secretion (Fig. 2A). Use of an α-CD3/CD28 antibody, instead of PMA/ionomycin for stimulation, produced similar results. A single bacterium per Jurkat T cell (MOI = 1) was already sufficient to cause a 25% reduction in IL-2 secretion, indicating that physiological amounts of VacA could have an efficient suppressive effect on T cell proliferation in vivo (Fig. 2B). Furthermore, treatment of PBLCs with CCS+ also resulted in down-regulation of IL-2Rα surface localization, similar to the treatment of PBLCs with FK506, a drug that blocks the activity of the cellular phosphatase calcineurin (12) (Fig. 2C).

Fig. 2.

Hp VacA controls the secretion of IL-2 and the surface location of IL-2Rα (CD25). (A) Sandwich enzyme-linked immunosorbent assay (ELISA) to measure secreted IL-2. Jurkat T cells were infected with Hp strains as indicated, stimulated with PMA/ionomycin, and compared with stimulated noninfected control (100%). The data represent the means of three independent experiments. (B) Dose-response curve to determine the minimal number of Hp bacteria (MOI) to measure a response on IL-2 secretion of Jurkat T cells. (C) Isolated human PBLCs were treated with CCS+ (250 μg/ml), CCS (250 μg/ml), or FK506 (100 nM). Cells were activated with PHA (24 hours), stained with fluorescein isothiocyanate (FITC)-conjugated α-CD25 (Pharmingen), and analyzed by fluorescence-activated cell sorting (FACS). (D) Human PBLCs incubated with or without Hp for 24 or 48 hours were analyzed for cyclin D3 expression and phosphorylation of retinoblastoma protein (Rb) by immunoblot analysis (α-cyclin D3 or α-Phospho-Rb; actin used as loading control).

For gastric adenocarcinoma (AGS) epithelial cells, co-culture with Hp caused cell cycle inhibition from G1 to S phase (13). Thus, we analyzed the expression of cyclins from human PBLCs treated with Hp CCS+ or CCS. Cyclins D3 and E were efficiently down-regulated after 24 and 48 hours by Hp P12 (CCS+), but not by P12ΔvacA (CCS) (Fig. 2D). Consequently, the phosphorylation, but not the production of retinoblastoma protein (Rb) (which is a major downstream target of these cyclins controlling G1/S phase progression), was strongly reduced by VacA treatment of PBLCs after 48 hours (Fig. 2D). Thus, secreted Hp VacA plays a pivotal role in blocking T cell activation and proliferation at the stage of IL-2 signaling and by inducing cell cycle arrest.

To better understand the effect of VacA on IL-2 secretion, luciferase (luc) reporter constructs were used for transfection of Jurkat T cells, demonstrating that VacA efficiently blocked IL-2 production at the transcriptional level (Fig. 3A). Transcription factors AP-1, nuclear factor kappa B (NF-κB), NFAT, and RE/AP have been identified as essential for activation of the IL-2 gene promoter (14). Neither the AP-1–luc (Fig. 3B), nor the NFκB-luc (Fig. 3C) or the RE/AP-luc reporter (Fig. 3D), was down-regulated by VacA. However, an NFAT-luc reporter was strongly silenced after treatment of transfected Jurkat cells with Hp P12, but not P12ΔvacA (Fig. 3E), indicating a transcriptional silencing of the IL-2 gene by targeting the NFAT signaling pathway. Down-regulation of the high-affinity IL-2Rα, the gene of which is also under control of NFAT (15), supports this observation (Fig. 2C). CDT seems to use a different strategy (Fig. 3, F and G) by acting as a deoxyribonuclease (DNase) inducing cell cycle arrest by DNA strand breakage (16).

Fig. 3.

VacA inhibits NFAT, but not NFκB, AP-1, or RE/AP, which are involved in IL-2 gene expression. Reporter plasmids carrying the complete IL-2 promoter (A and F), an AP-1 binding element (B), an NFκB-binding element (C), a RE/AP-binding sequence (D), and an NFAT-binding sequence (E and G) coupled to the luc reporter gene were transfected into Jurkat T cells. Cells were treated with Hp or C. jejuni strains, as indicated, and activated with PMA/ionomycin. The Luc activity of the activated, nontreated control was set to 100%. Data points represent the mean ± SEM from three experiments.

NFAT represents a family of related transcription factors involved in the coordinate induction of expression of a number of cytokines upon T cell activation. The calcium-activated phosphatase calcineurin unmasks nuclear localization signals on NFAT by dephosphorylation of serine-threonine motifs, resulting in its translocation into the nucleus. A HeLa cell line constitutively expressing a green fluorescent protein (GFP)-NFAT fusion protein (17) showed nuclear GFP-fluorescence upon activation (Fig. 4A). Hp P12, but not P12ΔvacA, clearly abrogated the nuclear translocation of NFAT in 80% of cells (Fig. 4, A and B; P12/P12ΔvacA). Similar data were obtained with Jurkat T cells transiently transfected with the GFP-NFAT–expressing plasmid pAD4 (17). Thus, VacA inhibits nuclear translocation of NFAT, which explains the failure of T cells to express and secrete IL-2. To determine whether VacA targeted NFAT directly or whether it prevented dephosphorylation of NFAT by blocking calcineurin phosphatase activity, we engineered Jurkat T cells to express a constitutively active form of calcineurin (pSRα-ΔCaM-AI). This abrogated the activity of VacA completely (Fig. 4C), suggesting that VacA did not interact with NFAT directly but blocked the signaling cascade at the level of the phosphatase calcineurin, or further upstream.

Fig. 4.

VacA blocks NFAT nuclear translocation and suppresses a subset of genes in Jurkat T cells also inhibited by FK506. (A) HeLa cells stably expressing GFP-NFAT were infected as indicated or left uninfected. Cells were treated with PMA/ionomycin (+) or left untreated (–), were fixed, and were examined by confocal microscopy. (B) Summary of GFP-NFAT localization, scored for each cell in five randomly chosen microscopic fields. (C) Jurkat T cells were transfected with pIL-2-Luc or pIL-2-Luc as well as a constitutively active version of calcineurin A (pSRαΔCaMAI). Activation of cells was with PMA/ionomycin. (D) Expression patterns of 136 genes with statistically significant (false discovery rate ≤ 0.82%) differential expression by a factor of two or higher are shown. Red indicates expression level above, green indicates expression level below, and black near to the row-wise mean, respectively. Group 1, genes induced by both FK506 and VacA; group 2, genes suppressed by FK506 and VacA (2A, suppression by FK506 > VacA; 2B, suppression by FK506 = VacA); group 3, genes suppressed by FK506 only. See table S3 for details.

FK506 (Tacrolimus) and Cyclosporin A (CsA) are medically important immunosuppressive drugs that inhibit calcineurin activity. Using microarray analysis, a set of 136 genes composed of four groups was determined as differentially regulated (Fig. 4D) (table S3). Group 1 genes were up-regulated by both FK506 and VacA, whereas genes in group 2 were down-regulated by both compounds. Group 2A consisted of genes strongly suppressed by FK506 but only moderately by VacA, whereas group 2B contained genes equally suppressed by FK506 and VacA. Finally, group 3 comprised genes down-regulated by FK506 but not significantly changed in their expression by VacA. All genes identified as being differently expressed by VacA represented a subgroup also modulated by FK506, supporting our hypothesis that both factors use a common mechanism of gene regulation by blocking calcineurin (fig. S3). Whereas VacA neither influenced NFκB nor AP-1 transcription factors negatively (Fig. 3, B and C), FK506 could suppress NFκB and AP-1 (18), which might explain its repression of group 3 genes. Among the genes most strongly down-regulated (group 2B) were those encoding IL-2 and NFAT, but also genes encoding C- or CC-chemokines, such as single C motif-1 (SCM-1β), SCM-1α (lymphotactin, ATAC), macrophage inflammatory protein (MIP)-1α, or MIP-1β. Most of these carry experimentally verified NFAT binding sites (19, 20).

The approach of using bacterial pathogens or viruses to block T lymphocyte activation and/or proliferation and to induce immune suppression has recently been recognized as a successful strategy in the interaction of microbes with its host (21). A vacA-negative Hp SS1 mutant is compromised in its ability to initially establish an infection when competing with the corresponding wt strain (22). However, how might Hp or VacA meet T cells in the gastric mucosa? The number of lamina propria and intraepithelial T cells of the CD4+ and CD8+ subtype are significantly increased in Hp-infected versus noninfected patients (23), and Hp could directly contact such intraepithelial T cells. Because tight junctions can be opened by Hp (24), secreted bacterial products such as VacA can be found deep in the lamina propria (25). Thus, VacA might act as a “long distance weapon” to efficiently block proliferation of T cells in the local gastric environment. For Hp, classified as a type I carcinogen, a mechanism of local immune suppression might also be an important instrument for induction of malignant mucosa-associated lymphoid tissue (MALT) lymphoma and adenocarcinoma of the stomach.

Supporting Online Material

www.sciencemag.org/cgi/content/full/301/5636/1099/DC1

Materials and Methods

Figs. S1 to S3

Tables S1 to S3

References

References and Notes

View Abstract

Navigate This Article