SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Microparticles (MPs) are small cell membrane vesicles that are released from cells during apoptosis or activation. Although circulating platelet MPs have been studied in some detail, the existence and functional role of T cell MPs remain elusive. We show that blood from patients with active hepatitis C (alanine aminotransferase [ALT] level >100 IU/mL) contains elevated numbers of T cell MPs compared with patients with mild hepatitis C (ALT <40 IU/mL) and healthy controls. T cell MPs fuse with cell membranes of hepatic stellate cells (HSCs), the major effector cells for excess matrix deposition in liver fibrosis and cirrhosis. MP uptake is partly intercellular adhesion molecule 1–dependent and leads to activation of nuclear factor kappa B and extracellular signal-regulated kinases 1 and 2 and subsequent up-regulation of fibrolytic genes in HSCs, down-regulation of procollagen α1(I) messenger RNA, and blunting of profibrogenic activities of transforming growth factor β1. Ex vivo, the induced fibrolytic activity is evident in MPs derived from activated CD4+ T cells and is highest in MPs derived from activated and apoptotic CD8+ T cells. Mass spectrometry, fluorescence-activated cell sorting analysis, and function blocking antibodies revealed CD147/Emmprin as a candidate transmembrane molecule in HSC fibrolytic activation by CD8+ T cell MPs. Conclusion: Circulating T cell MPs are a novel diagnostic marker for inflammatory liver diseases, and in vivo induction of T cell MPs may be a novel strategy to induce regression of liver fibrosis. (HEPATOLOGY 2011.)

Cirrhosis is a complication of many forms of chronic liver disease. Due to a shortage of donors, liver transplantation is available to only a fraction of patients. Consequently, there is an urgent need for antifibrotic treatments, which can prevent, halt, or even reverse advanced fibrosis.1 Significant progress has been made in our understanding of hepatic fibrosis, which is now viewed as a dynamic process characterized by an excess of extracellular matrix production (i.e., fibrogenesis) over its degradation (i.e., fibrolysis), which eventually leads to distortion of the hepatic architecture (i.e., cirrhosis) and loss of organ function.1, 2

In hepatic fibrosis, excessive extracellular matrix is produced by activated mesenchymal cells, which resemble myofibroblasts. Mesenchymal cells derive from quiescent hepatic stellate cells (HSCs) and periportal or perivenular fibroblasts, hereafter referred to collectively as HSCs. Activation of HSCs by several profibrogenic cytokines and growth factors, especially by transforming growth factor β1 (TGF-β1), is a general feature of fibrosis progression.2 These factors are mainly produced by activated macrophages or cholangiocytes, but also by liver infiltrating lymphocytes.3

Several studies have suggested that advanced experimental and possibly human liver fibrosis can regress once pathogenic triggers are eliminated and sufficient time for recovery is available.4, 5 Interestingly, the same cells that drive fibrogenesis (HSCs) can become major effectors of fibrolysis through the production and activation of certain matrix metalloproteinases (MMPs). This has been shown in vitro when dermal fibroblasts are plated from a two-dimensional cell culture dish into a three-dimensional collagen gel,6, 7 allowing them to contract, thereby up-regulating MMPs and down-regulating procollagen I production.

A recent report suggested that lymphocytes can modulate fibroblasts in a different, non–cytokine-mediated manner.8 Thus a crude microparticle (MP) preparation released from the membranes of Jurkat T cells (an immortal lymphoma T cell line) during activation and early apoptosis could induce synovial fibroblast fibrolytic MMP expression. However, the mechanisms by which these MPs exert their fibrolytic effects remain unclear. Moreover, the effect of T cell–derived MPs on the activation of fibrogenic effector cells of a major organ such as the liver, where lymphocyte-driven inflammation frequently occurs, has not been addressed.1 Finally, such MPs were not demonstrated in the circulation.

We report that T cell MPs circulate in blood and are elevated in patients with active chronic hepatitis C. Further, MPs derived both from CD8+ and CD4+ T cells can induce a fibrolytic phenotype in HSCs. This activity depends on fusion of the MPs with HSC membranes and transfer of T cell membrane molecules such as CD147 (Emmprin) to HSCs in a partly CD54 (intercellular adhesion molecule 1 [ICAM-1])-dependent manner. We conclude that T cell MPs may become a novel diagnostic tool and could be used therapeutically to mitigate (hepatic) inflammation and fibrosis.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Cell Lines.

Jurkat T cells (ATCC#: CRL-2570, Manassas, VA) were grown in 10% fetal bovine serum (FBS) in Roswell Park Memorial Institute 1640 medium, and LX-2 were grown in 2.5% FBS in Dulbecco's modified Eagle's medium (Cellgrow, Manassas, VA). THP-1 monocytes (American Type Culture Collection No. TIB-202) were grown in 10% FBS in Dulbecco's modified Eagle's medium (Cellgrow) and were differentiated into macrophages by way of incubation with 0.05 μg/mL phorbol myristate acetate for 24 hours.9

Lymphocyte Isolation.

Human peripheral blood was collected in heparinized tubes from healthy volunteers within a protocol approved by Children's Hospital (Boston, MA). Mononuclear cells were isolated by way of centrifugation over Ficoll-Paque Premium (GE Healthcare, Uppsala, Sweden). After three washes in Hank's balanced salt solution, CD4+ and CD8+ T cells were isolated by way of negative selection using magnetic beads (Miltenyi Biotec, Auburn, CA).

Isolation of MPs from Activated or Apoptotic T Cells and Monocytes/Macrophages.

For induction of apoptosis, T cells or monocytes/macrophages were treated with 4 μmol/mL staurosporine (ST) (Cell Signaling Technology, Danvers, MA) for 4 hours. T cells were activated with 5 μg/mL phytohemagglutinin (PHA) (Roche, Mannheim, Germany) for 24 hours, and 3 days later restimulated. During stimulation with PHA, T cell cultures were supplemented with 5 ng/mL interleukin-2 (PeproTech, Rocky Hill, NJ). Three days after restimulation, cells were pelleted and cell-free supernatants were centrifuged at 10 × 103g for 20 minutes yielding S10-MPs, whereas the resultant supernatant was centrifuged at 100 × 103g for 90 minutes to yield purified, biologically active S100-MPs.

Characterization and Quantification of MPs.

The MP preparations were characterized by way of fluorescence-activated cell sorting (FACS) with an LSR2 sorter (Becton Dickinson, San Jose, CA) and cytometric data was analyzed with FlowJo 8.8.6 software (Tree Star, Inc., Ashland, Oregon). MP particles were gated on forward and sidescatter acquired from runs including 500 standard beads (Becton Dickinson, San Jose, CA). The number of CD3 (CD11a, CD14, CD147) and AnnexinV (eBioscience, San Diego, CA; GeneTex Inc., Irvine, CA for CD147) double-positive events were calculated relative to the number of beads added to the samples. To avoid unspecific antibody binding, Fc receptors on MPs and target cells were blocked with FcR Blocking Reagent (Miltenyi Biotec). Antibody solutions were centrifuged prior to FACS to avoid artifacts due to aggregation.

Isolation of T Cell MPs from Human Plasma and Liver Histology.

Human peripheral blood was collected in citrate-containing tubes (BD Vacutainer, Buff. Na. Citrate [9NC]; BD, Franklin Lakes, NJ) from patients and healthy controls (protocol approved by the Beth Israel Deaconess Medical Center, approval no. 2004-P-000318). MPs were isolated by way of differential centrifugation, and S100-MPs were characterized by way of FACS using staining for Annexin V, CD3, CD4, CD8, CD14, CD15, CD41, and CD25 (eBioscience) as detailed above. Levels of T cell MPs were correlated with liver histology as detailed in the Supporting Materials and Methods.

Incubation of HSCs with T Cell–Derived or Monocyte/Macrophage-Derived MPs and Quantitative Polymerase Chain Reaction.

HSCs (200 × 103/well) were seeded into six-well culture plates and serum-starved for 24 hours, followed by incubation with 1 × 103 or 50 × 103 S10-MPs or S100-MPs for 24 hours, and RNA extraction using TRIzol reagent (Invitrogen, Carlsbad, CA). ST (0.04 μM/mL) or plain medium served as controls. One microgram of RNA was reverse-transcribed using random primers and Superscript RNase H-reverse transcriptase (Invitrogen). Primers and probes are listed in Supporting Table 2. Relative transcript levels were quantified on a LightCycler 1.5 (Roche, Mannheim, Germany) using the TaqMan methodology.

Labeling of MPs.

MP membranes were labeled with the PKH26 lipid dye (Sigma-Aldrich, St. Louis, MO). Labeled S10-MPs and S100-MPs were coincubated with LX-2 cells for 0-1, 30, and 60 minutes, washed extensively, and fixed with 2% paraformaldehyde for 15 minutes at room temperature.

Quantification of CD3 Receptor Transfer.

HSCs (200 × 103/well) were seeded into six-well cell culture plates (BD Labware, Franklin Lakes, NJ) for 12 hours, serum-starved for 24 hours, followed by incubation with 100 × 103 S100-MPs for 1 minute up to 24 hours. Cells were then washed with phosphate-buffered saline and collected using trypsin/ethylene diamine tetraacetic acid (Cellgrow, Manassas, VA), and single-cell suspensions stained with anti–CD3-APC were followed by FACS analysis.

ICAM-1 Up-regulation of HSCs by TNFα.

Tumor necrosis factor α (TNFα) (PeproTech) was added to HSC cultures, and ICAM-1 expression assessed after 2, 4, and 24 hours by FACS using anti–ICAM-1 PE (eBioscience, San Diego, CA).

Comparative Proteomic Analysis.

A detailed description of the comparative proteomic analysis is provided in the Supporting Materials and Methods. A total of 20 μg of S-100 MP protein from ST-treated Jurkat T cells and Huh-7 hepatoma cells was extracted and denatured with 0.1% (vol/vol) sodium dodecyl sulfate in phosphate-buffered saline, reduced and alkylated, digested with trypsin, and labeled with isobaric tags (4-plex iTRAQ; Applied Biosystems, Foster City, CA). The two digested extracts were pooled and subjected to two-dimensional peptide fractionation and analyzed for their comparative proteomic signature by way of matrix-assisted laser desorption ionization/time of flight mass spectrometry.10

CD54 (ICAM-1) and CD147 (EMMPRIN) Blocking Studies.

Subconfluent, serum-starved HSCs were preincubated with monoclonal blocking anti-human CD54 or isotype-matched (immunoglobulin G1 [IgG1]) control antibody (50 μg/mL; GeneTex Inc., Irvine, CA) for 120 minutes, washed, and incubated with Jurkat T cell–derived S100-MPs. S100-MPs were incubated with monoclonal blocking anti-human CD147 (Abcam, Cambridge, MA) or IgG1 control antibody (50 μg/mL; GeneTex Inc.) for 60 minutes prior to their addition to HSCs.

P65 Nuclear Factor kappa B Translocation.

HSCs were serum-starved for 24 hours, then washed with phosphate-buffered saline and fixed in cold methanol for 10 minutes. Nuclear translocation of p65 nuclear factor kappa B (NFκB) was detected by incubating cells with polyclonal p65 antibody (1:100; Delta Biolabs) for 30 minutes followed by TRITC-conjugated anti-rabbit IgG (1:200, Dako, Germany). Representative images were documented using a scanning confocal microscope (Carl Zeiss, Germany).

Signaling Pathway Inhibition.

Serum-starved HSCs were incubated with the inhibitors SB203580 (p38 MAPK), U0126 (extracellular signal-regulated kinases 1 and 2 [ERK1/2]), and LY294002 (phosphatidyl-inositol-3 kinase) (LC Labs, Woburn, MA) as described.11 The proteasome inhibitor MG132 (Rockland Inc.) was used to block NFκB nuclear translocation and activity.

Statistical Analysis.

All data are presented as the mean ± SD. Differences between independent experimental groups were analyzed using a two-tailed Student t test. P < 0.05 was considered statistically significant. Correlations of MP levels with histological grade and stage were calculated by best-fit linear regression analysis based on a 95% confidence interval. All calculations were performed with Prism 4 (GraphPad Software, Inc.).

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

T Cell–Derived MPs Circulate in the Blood Plasma of Healthy Controls and Are Increased in Patients with Active Hepatitis C.

We searched for T cell–derived MPs in human plasma from normal controls and patients with chronic hepatitis. Pure S100-MPs that carried the MP marker Annexin V12, 13 and the T cell marker CD3 were present in human plasma (Fig. 1A). Their percentage increased significantly from 25% in healthy controls and patients with serologically mild hepatitis C (alanine aminotransferase [ALT] <40 IU/mL) to 31% in patients with serologically active hepatitis C (ALT >40 IU/mL and ALT >100 IU/mL) (Fig. 1B). The higher percentages were paralleled by a higher mean fluorescence intensity for CD3 (data not shown). Of note, looking at T cell subsets, patients with active hepatitis C had a significant increase in circulating MPs derived from CD4+ as well as CD8+ T cells (two- and 1.5-fold versus patients with mild hepatitis C and healthy controls, respectively). Furthermore, 80% of CD8+ MPs were additionally CD25+, a T cell activation marker.12 Levels of MPs derived from other cells,14 such as CD41+ MPs (from platelets) and CD15+ MPs (from neutrophils), were unchanged, whereas CD14+ MPs (from monocytes, macrophages, and dendritic cells) were reduced by nearly 50% in patients with active hepatitis C (P = 0.015) (Fig. 1C). When patients' liver histology was matched with MP plasma levels using linear regression analysis, both histological grade and stage showed a significant correlation with CD4+ and CD8+ MPs (Fig. 2).

thumbnail image

Figure 1. T cell–derived MPs are found in plasma and are elevated in patients with active hepatitis C. (A) Representative FACS analysis of CD3-APC and Annexin V–fluorescein isothiocyanate double-positive S100-MPs in a plasma sample from a healthy human donor. (B) Relative percentage of circulating CD3 and Annexin V double-positive S100-MPs from patients with hepatitis C and normal ALT levels (<40 IU/L; n = 4), elevated ALT levels (>40 IU/L; n = 10), or high ALT levels (>100 IU/mL; n = 7). (C) CD4/Annexin V and CD8/Annexin V double-positive, CD14/Annexin V, CD15/Annexin V, and CD41/Annexin V double-positive S100-MPs in the plasma of patients with ALT >100 IU/L (n > 9) compared with healthy controls and hepatitis C virus patients with ALT <40 IU/L (n > 9). (D) CD8+ S100-MPs are ≈80% positive for CD25. *P < 0.05. **P < 0.005.

Download figure to PowerPoint

thumbnail image

Figure 2. Levels of circulating T cell–derived S100-MPs correlate with histological grade and stage in patients with hepatitis C. Correlations of plasma CD4+ and CD8+ S100-MPs with patient biopsy specimens were performed as detailed in the Supporting Materials and Methods. Patients with normal and elevated ALT levels were included. CD8+ analysis did not work in one patient with biopsy (Bx) stage 4.

Download figure to PowerPoint

Isolation and Characterization of T Cell–Derived MPs.

Due to the low numbers of circulating MPs, initial characterization and functional analyses were performed with T cell MPs released from the human Jurkat T cell line and from peripheral blood of healthy human donors. We stimulated MP release either by activation with PHA,15, 16 or by induction of apoptosis using the tyrosine kinase inhibitor ST.8 Whereas the Jurkat S10-MP fraction was Annexin Vlow and CD3low, the Jurkat S100-MP fraction was Annexin Vhigh and CD3high (Supporting Fig. 1A), which was confirmed by analysis of mean fluorescence intensity (Supporting Fig. 1B). This difference between S100-MPs and S10-MPs was found regardless of the mode of generation (by way of PHA, ST, or PHA and ST combined) (Supporting Fig. 1B).

Electron microscopic images from both fractions demonstrated that S10-MPs were heterogeneous in size and contained electron dense material, indicating debris of intracellular organelles, whereas S100-MPs showed a more homogeneous structure, being surrounded by a double-layered cell membrane and being electron-lucent, with a variable diameter ranging from 30 nm to 700 nm (Fig. 3A). Fig. 3B shows a typical FACS scatter plot that characterizes the S-100 MPs along with 3-μm marker beads and intact T cells.

thumbnail image

Figure 3. Characteristics of T cell–derived S100-MPs and S10-MPs and demonstration of their fusion with HSC membranes. (A) Ultrastructural analysis of the two subfractions of MPs generated from apoptotic Jurkat T cells. Magnification ×51,000. (B) Representative forward and sidescatter profiles of events in blood-derived S100-MPs after addition of beads and intact T cells. (C) FACS analysis demonstrating CD3 receptor transfer from S100-MPs to HSCs. A total of 2 × 105 LX-2 cells were incubated with 105 Jurkat T cell–derived S100-MPs and CD3-positive LX-2 HSCs were quantified after 6 hours. Unstained HSCs and HSCs incubated with 0.04 μM/mL ST served as controls. (D) Time-dependent uptake of CD3 S100-MPs by HSCs assessed by way of FACS analysis, demonstrating maximal MP uptake (15%-17%) after 6 hours (n = 3 events; mean ± SD). *P = 0.003. **P = 0.01. (E) Fluorescence microscopy confirming S100-MP uptake and membrane fusion with HSCs. S100-MPs were labeled with PKH26 membrane dye and incubated with LX-2 HSCs.

Download figure to PowerPoint

CD3 T Cell Receptor Transfer from S100-MPs to Cell Membranes of Human HSCs.

The exclusive expression of transmembrane CD3 on T cells allowed us to monitor the transfer of CD3 from S100-MPs to human LX-2 HSCs. Six hours of incubation with S100-MPs, the transfer of CD3 from MPs to HSCs peaked, with 17% of the HSCs being positive for CD3 (Fig. 3C,D). In support of the FACS data, fluorescence microscopy demonstrated that S100-MPs labeled with the membrane-dye PKH26 began to attach to HSC membranes at 30 minutes, generating a punctate red-fluorescent membrane pattern, and a diffuse membrane staining, indicative of membrane fusion, from 60 minutes onward (Fig. 3E). Membrane fusion was not found with PKH26-labeled S10-MPs (Supporting Fig. 1C).

T Cell Derived S100-MP Do Not Induce Apoptosis of HSCs.

Because MMPs, especially MMP-3, are up-regulated in cells undergoing apoptosis,17 and because our data show that S100-MPs derived from apoptotic T cells prominently up-regulated MMP-3 in HSCs, we evaluated apoptosis induction by S100-MPs using Annexin V and 7-amino-actinomycin D staining as a readout (Supporting Fig. 1D,E). Jurkat T cell–derived S100-MPs did not induce enhanced apoptosis or necrosis in HSCs after 24 hours of incubation, which also ruled out a significant ST contamination in our MP preparations.

Effect of S100 T Cell MPs on Fibrosis-Related Gene Expression by HSCs.

Fibrosis related transcripts were measured in LX-2 HSCs 24 hours after addition of 1 × 103 or 50 × 103 S100-MP from Jurkat T cells using quantitative reverse-transcription polymerase chain reaction (RT-PCR). S10-MPs, plain medium, and ST alone served as controls. MPs were obtained from PHA-activated and/or apoptotic (ST-treated) Jurkat T cells. After induction of T cell apoptosis, significant changes in fibrosis-related transcripts were found with 50 × 103 S100-MP, whereas equivalent amounts of S10-MPs had no effect (Fig. 4A). S100-MPs induced a significant (2.05- to 4.9-fold) up-regulation of fibrolytic genes (MMP-1, MMP-3, MMP-9, MMP-13) in HSCs, whereas transcript levels of the profibrogenic genes tissue inhibitor of metalloproteinase 1 (TIMP-1) and procollagen α1(I) were unaffected (Fig. 4A). Similar results were obtained when S100-MPs were incubated with freshly isolated primary rat HSCs. Here, the human S100-MPs induced MMP-3 even nine-fold (Supporting Fig. 2). S100-MPs from apoptotic T cells that had been preactivated by PHA did not induce up-regulation of MMPs in human HSCs, but rather down-regulated MMP-3 (Supporting Fig. 4). A similar response was found with S100-MPs derived from merely PHA-activated T cells (data not shown). As non–T cell controls, MPs derived from THP-1 monocytes and macrophages did not induce significant changes in MMP, TIMP-1, or procollagen α1(I) transcript levels, except for induction of MMP-3 and TIMP-1 by macrophage-derived MPs (Supporting Fig. 4).

thumbnail image

Figure 4. S100-MPs from Jurkat T cells elicit antifibrogenic responses in HSCs. (A) MMP-1, MMP-3, MMP-9, MMP-13, TIMP-1, and procollagen α1(I) transcripts were determined by quantitative RT-PCR in LX-2 HSCs (2 × 105 cells per well in 12-well plates) that were incubated with 103 or 5 × 104 S10-MPs or S100-MPs from apoptotic Jurkat T cells suspended in 350 μL medium for 24 hours. ST (0.04 μM/mL) or plain medium served as controls. (B) Induction of fibrolytic and inhibition of fibrogenic genes in TGFβ1 (5 ng/mL)-activated HSCs when incubated with S100-MPs (2 × 105 MPs suspended in 350 μL medium) for 24 hours. All experiments were performed at least twice (n = 3-4/group). Results (mean ± SD) are expressed as arbitrary (arb.) units relative to β2-microglobulin mRNA. *P < 0.05 versus medium control. **P < 0.005.

Download figure to PowerPoint

S100-MPs Abrogate HSC Profibrogenic Responses to TGFβ1.

Human HSCs were exposed to 5 ng/mL TGFβ1, which elicits a strong fibrogenic response. Jurkat T cell-derived S100-MPs not only blunted the TGFβ1 response by reducing procollagen α1(I) expression, they induced fibrolytic MMP transcripts beyond the levels produced by unstimulated HSCs (Fig. 4B). Therefore, TGFβ1 enhanced HSC procollagen α1(I) expression 2.7-fold, which after MP addition was reduced by almost 40%, and MPs increased the expression of MMP-3 and MMP-13 almost 2.5- and 2.1-fold, respectively. In addition, both in TGFβ1-treated and TGFβ1-untreated HSCs the addition of S100-MPs significantly reduced profibrogenic TIMP-1 expression by 30%-35% (Fig. 4B).

Comparison of the Effect of S100-MPs Derived from CD4+ and CD8+ T Cells.

Overall, apoptotic CD4+ T cell–derived MPs induced MMP expression in HSCs much less efficiently than MPs from CD8+ T cells, irrespective of their mode of generation (with or without prior activation by PHA). Therefore, MPs from CD4+ T cells did not significantly affect MMP-1, MMP-3, MMP-9, MMP-13, TIMP-1, or procollagen α1(I) expression (data not shown). If MPs were induced only by CD4+ T cell activation with PHA, a significant induction was observed for MMP-1, MMP-3, and MMP-9 messenger RNA (mRNA) (between 1.7- and three-fold), whereas procollagen α1(I) and TIMP-1 transcript levels remained unchanged (Supporting Fig. 5). S100-MPs derived from apoptotic CD8+ T cells did not affect fibrosis-related gene expression (Supporting Fig. 6), whereas S100-MPs from CD8+ T cells that were only preactivated by PHA increased MMP-1 transcripts 1.9-fold and reduced procollagen α1(I) transcripts by 30% (data not shown). S100-MPs from apoptotic CD8+ T cells that were preactivated by PHA produced the strongest fibrolytic effects in HSCs, also reducing procollagen α1(I) mRNA significantly by 45% (Fig. 5A).

thumbnail image

Figure 5. S100-MPs from activated and apoptotic human CD8+ T cells increase MMP and reduce procollagen α1(I) gene expression in HSCs in a partly CD54-dependent manner. (A) Transcript levels were determined in LX-2 HSCs (2 × 105 cells/mL per well) incubated with S10-MPs or S100-MPs (103 or 5 × 104) from PHA-activated and apoptotic CD8+ T cells for 24 hours by way of quantitative RT-PCR. ST (0.04 μM/mL) or plain medium served as controls. *P < 0.05 versus medium control. (B) FACS analysis revealed that 64% of the S100-MPs were CD11a and Annexin V double-positive. (C) HSCs were stimulated with TNFα (10 ng/mL) for 0, 4, and 24 hours, resulting in a 40% up-regulation of CD54. *P < 0.001. (D) Up-regulation of CD54 on the surface of HSCs by TNFα facilitated MMP induction after addition of S100-MPs. *P < 0.05. **P = 0.04. ***P = 0.001. (E) HSCs were incubated with a CD54 blocking antibody (50 μg/mL) or an IgG-matched control antibody for 2 hours, followed by addition of S100-MPs for 24 hours. MMP-3 and MMP-13 transcripts were determined by way of quantitative RT-PCR. *P = 0.02. **P = 0.046. All experiments were performed at least twice (n = 3/group). Results (mean ± SD) are expressed as arbitrary (arb.) units relative to β2-microglobulin mRNA.

Download figure to PowerPoint

thumbnail image

Figure 6. T cell MPs engage CD147 (EMMPRIN) on HSCs and elicit MMP expression by way of ERK1/2. (A) FACS analyses of CD147 expression on S100-MPs and LX-2 HSCs (CD147 positivity 77% and 99%, respectively). (B) CD8+ T cell–derived S100-MPs (PHA+ST treatment) were incubated with CD147 blocking antibody (50 μg/mL) for 1 hour, followed by addition to LX-2 HSCs for 24 hours. CD147 blocking significantly decreased MMP-3 (by 35%, *P = 0.007) and MMP-9 (30%, **P = 0.03) induction as determined by way of quantitative RT-PCR. Experiments were performed twice (n = 3/group). Results (mean ± SD) are expressed as arbitrary (arb.) units relative to β2-microglobulin mRNA. (C) Induction of MMP-3 by T cell MPs in HSCs. There was a lack of inhibition by the phosphatidyl-inositol-3 kinase inhibitor LY294002 (LY, 5 μg/mL). Abrogation of MMP-3 induction by the ERK1/2 inhibitor U0126 (U, 5 μg/mL), and 50% inhibition by the p38 kinase inhibitor SB203580 (SB, 5 μg/mL) and the proteasome (NFκB) inhibitor MG132 (MG, 15 μg/mL). *P = 0.02. Comparison with untreated S100-MP–stimulated controls. (D) Nuclear translocation of NFκB p65 in LX-2 HSCs exposed to S100-MPs from Jurkat T cells for 60 minutes. Representative micrograph from three similar experiments. (E) Sketch illustrating the transfer of T cell–derived membrane-associated molecules including CD147 to HSC membranes by way of shredded MPs. These MPs fuse with the HSC membrane, which is facilitated by CD54. The transferred receptors can activate novel signaling pathways or autocrine/paracrine signaling loops in HSCs that favor a switch toward a fibrolytic phenotype by way of mitogen-activated protein kinase and/or NFκB pathway activation and subsequent induction of MMPs.

Download figure to PowerPoint

CD54 (ICAM-1)–Dependent Uptake of S100-MPs.

It remains to be shown what cell membrane molecules or receptors mediated attachment and uptake of S100-MPs by HSCs. Our FACS analysis revealed that >60% of S100-MPs were highly positive for the CD54 ligand CD11a (Fig. 5B). Assuming that ICAM-1 expressed by the recipient HSCs is engaged by CD11a/CD18 on the S100-MPs, an increased HSC CD54 expression should enhance MP uptake. We therefore incubated HSCs with 10 ng/mL TNFα, a strong inducer of CD54,18 which induced a robust (>10-fold) up-regulation (Fig. 5C). This pretreatment led to a further significant MP-induced increase of MMP-3, MMP-9, and MMP-13 expression in HSCs (Fig. 5D). A direct fibrolytic effect of TNFα on HSCs was largely ruled out, because TNFα alone did not enhance HSC MMP-3 mRNA, and alone modestly induced HSC MMP-9 and MMP-13 expression (Fig. 5D).

To corroborate that the observed effects were indeed due to an engagement of CD54 on HSCs, HSCs were incubated with CD54 blocking or an isotype-matched control antibody 2 hours prior to addition of S100-MPs. CD54 blocking resulted in a significant down-regulation of MMP-3 and MMP-13 transcripts induced by MPs from Jurkat T cells (40% and 45%, respectively) (Fig. 5E).

Emmprin (CD147) Is Involved in MP-Induced MMP Induction in HSCs.

Comparative quantitative proteomics of T cell versus control (Huh7 hepatoma) cell S100-MPs using iTRAQ isobaric tagging yielded three candidate cell (membrane)-associated molecules, other than growth factor or cytokine receptors, namely nodal modulator 1 and 2 (molecules involved in the inhibition of TGFβ signaling and Emmprin/Basigin (CD147) (Supporting Table 1). FACS analysis showed that T cell–derived S100-MPs as well as HSCs were highly positive for CD147 (>70% and 99%, respectively), a molecule that requires homodimeric interaction for MMP induction (Fig. 6A). Blocking of CD147 on S100-MPs (CD8+ T cell–derived after induction with PHA and ST) resulted in a significant reduction of MMP-3 and MMP-9 transcripts (35% and 30%, respectively) (Fig. 6B), confirming the functional involvement of CD147.

Table 1. Summary of Observed Fibrolytic Effects on HSCs Induced by S100-MPs Derived from Activated and/or Apoptotic Human T Cells
 JurkatCD4+CD8+
STPHA+STPHASTPHA+STPHASTPHA+STPHA
  1. MMP-1, MMP-3, MMP-9, MMP-13, TIMP-1, and procollagen α1(I) transcript levels were determined by way of quantitative RT-PCR in LX-2 HSCs (2 × 105 cells per well) incubated with (active) S-100 or (inactive) S-10 MPs for 24 hours. Only effects ≥50% were considered relevant. Up-regulation was categorized as follows: +++, more than four-fold; ++, more than two-fold; +, less than two-fold compared with plain medium without MPs or ST; down-regulation was categorized as follows: - - -, more than 75%; - -, more than 50%; -, less than 50% compared with plain medium without MPs or ST; ∼, not significant toward ST control.

MMP-1(++)-+++++
MMP-3++-----++++
MMP-9+++(+++)+++(++)
MMP-13+++++
TIMP-1
Pro-collagen α1(I)

Fibrolytic Activation of HSCs by S100-MPs Depends on NFκB and ERK1/2 Pathways.

HSC MMP-3 induction by T cell MPs was completely abrogated by inhibition of p42/p44 mitogen-activated protein kinase (ERK1/2), to a modest degree by inhibition of p38 or NFκB, and remained unaffected by inhibition of phosphatidyl-inositol-3 kinase/Akt (Fig. 6C). >10% of HSC showed NFκB relocation to the nucleus after incubation with S100-MP, confirming modest activation of the NFκB pathway (Fig. 6D).

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

We have shown that CD4+ and CD8+ T cell–derived MPs can be detected in human plasma, and that their percentages were significantly elevated in patients with active hepatitis C, as reflected by high ALT levels. In vitro, S100-MPs are released from human T cells after activation (and apoptosis) and fuse with the cell membranes of HSCs and transfer membrane molecules (CD147, Emmprin), which triggers up-regulation of fibrolytic MMP-1, MMP-3, MMP-9, and MMP-13. Of note, the circulating CD4+ and CD8+ S100-MPs found in patients' plasma mainly derive from activated T cells, and their equivalent generated ex vivo by PHA stimulation of donor CD4+ and CD8+ T cells most strongly up-regulated putatively fibrolytic MMPs in HSCs (Table 1). This finding will likely have relevance in vivo, because activated HSCs are the principal driving force of liver fibrogenesis.

MPs were described as a product of various kinds of cell types, including T cells, as a product of activation or early apoptosis. However, characterization of the biological effects of these MPs has been limited. A prior study implicated MPs from the Jurkat T cell line in fibrolytic activation of synovial fibroblasts.8 Questions relevant to liver disease or diseases of other epithelial-mesenchymal organs have not been addressed.

We demonstrated that increased T cell activation (and apoptosis) in active hepatitis C19 is paralleled by excess release of T cell–derived MPs, which can be detected in the circulation. Using T cell subpopulations and HSCs, both of which are key players in liver inflammation and fibrogenesis, we demonstrated the functional relevance of these MPs in vitro. Therefore, T cell MPs ameliorated or even blunted the fibrogenic response that is usually prevalent in chronic hepatitis,1 including the neutralization of fibrogenic activation of HSCs by TGFβ1, the strongest profibrogenic cytokine in hepatic fibrosis and other fibrotic diseases.2

Of note, not all T cell–derived MPs were equally potent inducers of fibrolytic MMP expression in HSCs. Therefore, MPs derived from apoptotic and activated CD8+ T cells were the strongest inducers compared with MPs from activated CD4+ T cells or from the CD4-expressing Jurkat T cell line (Table 1). In this regard, it is noteworthy that CD8+ cells predominate in livers with hepatitis C, and the presence of CD8+ rather than CD4+ T cells has been correlated with the progression of liver fibrosis.20-22 These contrast with circulating MPs in inflammatory intestinal diseases where CD4+ T cell–derived MPs predominate (unpublished data). Therefore, MPs derived from activated (and apoptotic) CD8+ and CD4+ T cells may represent a negative feedback loop that counteracts the yet ill-defined profibrogenic activity of T cells once they become highly stimulated (as reproduced in vitro with PHA) with or without subsequent deletion by apoptosis. Human T cell–derived MPs could also potently induce MMP expression in primary HSCs from rats, suggesting a conserved mechanism, which is working beyond species boundaries. MPs from THP-1 monocytes and macrophages did not significantly induce fibrosis or fibrolysis-related transcripts, except for an induction of MMP-3 and TIMP-1 in macrophages, underlining the unique properties of T cell–derived MPs.

As a prominent adhesion molecule in T cell interactions, we evaluated the role of CD54 (ICAM-1). CD54 is needed for the adhesion of lymphocytes to antigen-presenting cells for immune priming and for the interaction between T cells and HSCs.3, 23 We showed that both the fusion of and biological effects elicited by T cell–derived MPs were at least partly mediated through CD54. In addition, proteomic analysis revealed several membrane and intracellular molecules in the S100-MP preparation from Jurkat T cells that were absent in the S100-MP fraction from inactive control cells (Supporting Table 1). A primary candidate molecule in this search was the transmembrane MMP inducer Emmprin/Basigin (CD147). CD147 is expressed on monocytes, stromal fibroblasts, platelets, cardiac myocytes, and on tumor epithelia including hepatocellular cancer cells.24-27 Homodimerization of CD147 by interaction of neighboring cells elicits signaling pathways that lead to expression of MMP-1, MMP-2, MMP-3, MMP-9, and MMP-11.28-30 Of note, CD147–CD147 interactions were found between tumor cells31 and suggested between tumor cells and surrounding fibroblasts.32 CD147 activation on monocytes was reported to activate the NFκB pathway and induce MMP-9 expression,33 and to stimulate the ERK1/234 and p38 MAPK pathways.35 By using a CD147 blocking antibody, we confirmed the functional involvement of this molecule (MMP down-regulation by 30%-35%). Additional fibrolytic mechanisms may be engaged in HSCs by S100-MPs, which involve mainly ERK1/2 and NFκB activation. Furthermore, although not the focus of the present study, transfer of bioactive soluble molecules within MPs (e.g., cytokines, microRNAs, or effectors of hedgehog signaling) may occur.36

To date, the generation of MPs in general and of T cell–derived MPs in particular for in vivo therapeutic use remains elusive. So far, only one group infused tumor cell–derived MPs under well-defined conditions in vivo to accelerate arteriolar occlusion.13 The reasons are several, including potential difficulties to induce MPs specifically in CD8+ T cells as the major fibrolytically active T cell subset, or to prevent undesired side effects when using cytokines, biological agents, or proapoptotic agents. Alternatively, MPs could be generated ex vivo to be infused or even injected into the target organ.

In conclusion, we demonstrated a novel mechanism by which activated (and apoptotic) T cells induce fibrolytic activation of HSCs, the most relevant fibrogenic effector cells in the liver. The proposed mechanisms are schematically illustrated in Fig. 6E. We assume that similar mechanisms likely apply to cells of other organs once T cell infiltration dominates the inflammation. We identified CD54 and CD147 as important mediators of MP fusion with HSC membranes and MMP induction, but we anticipate that additional molecular players will be discovered. T cell–derived MPs may give rise to their exploitation as novel diagnostic markers and potential antifibrotic agents.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

LX-2 HSCs were kindly donated by Scott L. Friedman. We thank Gunda Millonig (Beth Israel Deaconess Medical Center) and Veronika Lukacs-Kornek (Dana-Farber Cancer Institute) for help with initial FACS experiments and Franck Grall (Beth Israel Deaconess Medical Center) for help with mass spectroscopy.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
HEP_23999_sm_SuppInfo.doc34KSupporting Information
HEP_23999_sm_SuppInfoFig1.tif2013KSupplementary Figure 1: Characteristics of S100-MP and S10-MP T cell-derived microparticles. (A) S100-MP are Annexin V FITChigh /CD3 APChigh, whereas the S10-MP fraction is Annexin V FITClow/CD3 APClow. (B) Mean fluorescence intensity (MFI) of samples in (A); analysis of n?4 events; means±SD; *p<0.0001 and **p=0.004. MFI values from S100-MP and S10-MP obtained from apoptotic (ST), PHA-activated and apoptotic (ST & PHA), or PHA-activated Jurkat T-cells (PHA). (C) S10-MP were labeled with the PKH26 membrane dye and added to HSC. They remained a particulate fraction that was only loosely associated with the HSC, in contrast to S100-MP which merged with HSC membranes (see Fig.3G). (D) Lack of significant apoptosis induction in HSC 24h after incubation with S100-MP. FACS analysis using Annexin V and against 7-AAD staining. In contrast, ST, a dose that reflects maximal possible ST contamination in MP preparations, induced apoptosis. (E) Quantitative analysis of the 7-AAD/Annexin FACS data for late stage apoptosis, showing a 7-fold higher percentage of late apoptotic HSC after exposure to ST (0.04μM/mL) compared to S100-MP (*; p=0.047).
HEP_23999_sm_SuppInfoFig2.tif92KSupplementary Figure 2: Effects of human Jurkat T cell-derived S100-MP on primary rat HSC. MMP-3 transcript levels were determined by quantitative RT-PCR in primary rat HSC (2×105 cells per well in 12-well plates) that were incubated with S10-MP or S100-MP (2×103 or 5×104 MP per well) generated from apoptotic Jurkat T cells for 24h. Medium only served as control. Results (means±SD) are expressed as arbitrary units relative to beta2-microglobulin mRNA. *; p<0.03 vs. medium control.
HEP_23999_sm_SuppInfoFig3.tif231KSupplementary Figure 3: S100-MP from activated and apoptotic human Jurkat T cells do not induce MMP expression, but suppress MMP-3 RNA in HSC. Transcript levels were determined by quantitative RT-PCR in L×-2 cells (2×105 cells each well in 12-well plates) that were incubated with S10-MP or S100-MP (103 or 5×104) from activated and apoptotic Jurkat T cells for 24h. ST (0.04μM/mL) or plain medium served as controls. All experiments were at least performed twice with n=3-4 per group. Results (means±SD) are expressed as arbitrary units relative to beta2-microglobulin mRNA; * p<0.05 vs. medium control.
HEP_23999_sm_SuppInfoFig4.tif156KSupplementary Figure 4: Effects of human macrophage- or monocyte-derived S100-MP on L×-2 HSC fibrolyitic and fibrogenic gene expression. Transcript levels were determined by quantitative RT-PCR in L×-2 HSC (2×105 cells in ml in 12-well plates) that were incubated with 5×104 S100-MP from PHA-activated THP-1 monocytes, which were either left untreated or which were differentiated into macrophages by treatment with PMA for 24h (0.05μg/mL). Plain medium served as control. Experiments were performed twice with n=3 per group. Results (means±SD) are expressed as arbitrary units relative to beta2-microglobulin mRNA; * p<0.05 vs. medium control.
HEP_23999_sm_SuppInfoFig5.tif234KSupplementary Figure 5: S100-MP from activated human CD4+ T cells induce fibrolytic MMP-1, MMP-3, and MMP-9 in HSC. Transcript levels were determined by quantitative RT-PCR in L×-2 HSC (2×105 cells in ml in 12-well plates) that were incubated with 103 or 5×104 S10-MP or S100-MP from PHA-activated human CD4+ T cells for 24h. PHA (0.05μg/mL) or plain medium served as controls. Experiments were performed twice with n=3 per group. Results (means±SD) are expressed as arbitrary units relative to beta2-microglobulin mRNA; * p<0.05 vs. medium control.
HEP_23999_sm_SuppInfoFig6.tif236KSupplementary Figure 6: S100-MP from apoptotic human CD8+ T cells do not affect fibrosis-related gene expression in HSC. Transcript levels were determined by quantitative RT-PCR in L×-2 cells (2×105 cells each well in 12-well plates) that were incubated with S10-MP or S100-MP (103 or 5×104) from apoptotic CD8+ T cells for 24hrs. ST (0.04μM/mL) or plain medium served as controls. All experiments were at least performed 2-3 times with n=3-4 per group. Results (means±SD) are expressed as arbitrary units relative to beta2-microglobulin mRNA; * p<0.05 vs. medium control.
HEP_23999_SuppInfoTable1.tif232KTable 1: Selection of proteins identified in purified T cell derived MP by proteome analysis. S100-MP proteins were extracted from apoptotic Jurkat T cells and Huh-7 hepatoma cells as negative controls, digested with trypsin and labeled with isobaric tags. Tagged tryptic digests were pooled, peptides fractionated by ion exchange and HPLC analysis, and differential protein expression analyzed by MALDI-TOF mass spectroscopy as described. Shown is a selection of most abundant proteins specifically expressed on S100-P from T cells.
HEP_23999_SuppInfoTable2.tif223KTable 2: Primers and probes used for quantitative RT-PCR.

Please note: Wiley Blackwell 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.