Figure 1.
Schematic representation of TGF-β1 processing.
Sites of proteolytic cleavage are indicated by arrow heads. Thin black bars indicate disulfide bonds. Small black circles indicate cysteine 33 (Cys33). This position is relative to the starting methionine of the TGFB1 translation product. It must be noted that some authors number this cysteine as Cys4, referring to the position in pro-TGF-β1 after cleavage of the signal peptide. LAP: Latency Associated Peptide. TGFBR: TGF-β receptors.
Figure 2.
GARP increases cleavage of the pro-TGF-β1 precursor and secretion of latent TGF-β1 in T lymphocytes.
Cell lysates were analyzed by WB after SDS-PAGE under reducing conditions with antibodies against GARP, β-actin and a C-terminal epitope of the TGF-β1 peptide (top panels). Supernatants were treated or not with acid and analyzed by ELISA to measure concentrations of total (latent + active) and active TGF-β1, respectively (bottom panels). Total TGF-β1 detected in the acid-treated samples corresponds to latent TGF-β1 because no active TGF-β1 was detected in the non-treated samples. Values represent means of duplicates + SD. A. Analysis of human T cell lines transduced or not with lentiviruses coding GARP or GFP. T cells were left resting (Rest) or stimulated for 24 hours with anti-CD3/CD28 antibodies (Stim) in serum-free medium. B. Analysis of stable clones of murine BW5147 T cells and 293 cells transiently transfected with GARP and WT or C33S mutant TGFB1. Untransfected BW5147 and 293 cells express low levels of endogenous TGF-β1 that are not detectable by WB in these conditions (not shown). By comparison to WT, transfection of mutant C33S results in increased production of total TGF-β1 (pro- + mature), as previously described [49].
Figure 3.
GARP does not increase FURIN expression or activity, and does not co-immunoprecipitate with FURIN.
A. Expression of FURIN mRNA and protein were analyzed by RT-qPCR and WB in the human cells described in Figure 2. B. FURIN activity was measured 24 hours after transfection of 293 cells. Lysates of transfected cells were incubated with a FURIN fluorogenic substrate directly (top panel), or after capture on plastic-coated anti-FURIN antibody (bottom panel), to measure FURIN-like or FURIN specific activity, respectively. Graphs show mean fluorescence intensity at the indicated time (min) after addition of the substrate. The FURIN inhibitor Dec-RVKR-CMK was added to one condition to verify the specificity of the assay. C. Lysates of cells described in Figure 2 were immunoprecipitated with anti-GARP (IP GARP) or anti-FURIN (IP FURIN) antibodies. Immunoprecipitation products or total cell lysates (25% of input used for IPs) were analyzed by WB with anti-GARP, anti-TGF-β or anti-FURIN antibodies, as indicated.
Figure 4.
Disulfide-linked GARP/TGF-β1 complexes are released in the supernatant of T cells, but not 293 cells.
A. Cells described in Figure 2 were lysed and immunoprecipitated (IP) with anti-GARP or anti-LAP antibodies. IP products were submitted to SDS-PAGE under non-reducing or reducing conditions, followed by WB with anti-LAP antibodies (top and middle panels), or anti-GARP antibodies (bottom panels). Pro-TGF-β1 and LAP homodimers in the top panels are not clearly resolved, but can be distinguished better with longer migrations or higher concentrations of polyacrylamide. The +/- 85-90 kDa bands that appear in the middle panel correspond to non-specific bands, or to incompletely reduced pro-TGF-β1. B. Cells (2x106/ml for murine and human T cells, 2.5x105/ml for transfected 293 cells) were incubated in serum free medium during 24 hours. Different cell concentrations were used to adjust for the different amounts of secreted TGF-β1 (see Figure 2). Human Th A2 and Jurkat cells were stimulated with anti-CD3/CD28 antibodies to increase secretion. Supernatants (0.5-10 µl) were analyzed by WB under non-reducing conditions with anti-GARP and anti-LAP antibodies. * Band that also appears when the secondary anti-IgG2b-HRP antibody is used alone (without anti-GARP antibody), due to cross reactivity against the anti-CD3/CD28 antibodies used for T cell stimulation.
Figure 5.
Disulfide-linked GARP/TGF-β1 complexes are released by stimulated human Tregs, which naturally express GARP.
The indicated Treg and Th cell populations were left resting or stimulated with anti-CD3/CD28 antibodies in serum-free medium. A. Supernatants were collected after 48 hours. B. Cell lysates were collected after 24 hours and IP with an anti-GARP antibody. Supernatants (A) and immunoprecipitated lysates (B) were submitted to SDS-PAGE under non-reducing conditions, then analyzed by WB with anti-LAP antibodies. Similar results were obtained with freshly isolated CD4+ T cells from 2 other donors and with expanded CD4+ T cells from 5 others donors.
Figure 6.
Identification of 6 miRNAs targeting the GARP 3’ UTR in 293 cells.
A. 293 cells were cotransfected with a reporter plasmid and the indicated miRNA mimics (black regular: 20 miRNAs predicted to bind the GARP 3’ UTR and expressed in T cells; grey italic: negative controls). The reporter plasmid contains the GARP 3’ UTR cloned downstream of the Renilla luciferase gene, and a Firefly luciferase gene to control for transfection efficiency. Graphs indicate the ratio of Renilla to Firefly activities in cotransfected cells, normalized to the ratio in cells transfected with the plasmid alone (no miRNA). Data presented are the mean normalized ratios + SD measured in 3 to 12 independent experiments. * p < 0,0001 by comparison to control random miRNA for normalized ratios <1 (unpaired two-tailed Student’s t test). B. Schematic representation of the GARP 3’ UTR region, with predicted miRNA binding sites indicated by black boxes. The end of the truncated 1.7 kb 3’ UTR region cloned in a lentivirus used in Figure 9 is indicated by an arrow. C. Nucleotide sequences in black correspond to regions of the GARP 3’ UTR where the indicated miRNAs are predicted to bind (subscript numbers indicate positions relative to the first nucleotide after the STOP codon). miRNA sequences are shown in green, with seed regions underlined. Optimal alignments between the miRNA and the partial GARP 3’ UTR sequences were calculated with the mfold software [50]. Red letters and red boxes indicate nucleotides in the GARP 3’ UTR that were substituted or deleted by targeted mutagenesis, respectively. D and E. 293 cells were cotransfected as in A, except that the reporter plasmid contained wild type (WT) or mutated (mut) forms of the GARP 3’ UTR, as indicated. Data presented are the normalized ratios of Renilla to Firefly activities (means of triplicates + SD) and are representative from 2 to 4 independent experiments.
Figure 7.
Endogenous GARP levels in Tregs are reduced after transfection of miR-181a, miR-142-3p and miR-185 mimics.
Polyclonal CD4+CD25+CD127lo populations were purified from human PBMCs and amplified in vitro. Amplified cells were electroporated with the indicated miRNA mimics and stimulated 6 hours later with anti-CD3/CD28 antibodies. Cell lysates were collected 24 hours later and analyzed by WB with anti-GARP and anti-β-ACTIN antibodies.
Figure 8.
Expression of miRNAs in human CD4+ Th and Treg clones.
Expression levels of the indicated miRNAs and of control small nuclear RNA RNU44 were measured by RT-qPCR in 4th clones (dashed lines) and 3 Treg clones (plain lines), at rest or 24 hours after stimulation with anti-CD3 and anti-CD28 antibodies.
Figure 9.
A short GARP 3’ UTR region controls GARP levels and production of TGF-β1 in T cells.
Clone Th A2 was transduced with lentiviruses containing the GARP coding sequence (GARP ORF) alone, or followed by a full length (FL) or a truncated form (1.7 kb) of the GARP 3’ UTR. Transduced cells were restimulated for 24 hours with anti-CD3/CD28 antibodies in serum-free medium.
A to C. Cell lysates were analysed by WB after SDS-PAGE in reducing conditions with antibodies against GARP, β-ACTIN (A) and a C-terminal epitope of TGF-β1 (C), and by RT-qPCR for GARP and housekeeping gene EF-1 (B). Values in B are means of duplicates + SD. D. Concentration of latent TGF-β1 in the supernatants was measured by ELISA as indicated in Figure 2. Values are means of duplicates + SD, and are representative of 3 independent experiments. Similar results were obtained with unstimulated cells.