Fucosyltransferase 1 mediates angiogenesis in rheumatoid arthritis

Abstract

Objective: To determine the role of α(1,2)-linked fucosylation of proteins by fucosyltransferase 1 (FUT1) in rheumatoid arthritis (RA) angiogenesis.

Methods: Analysis of α(1,2)-linked fucosylated proteins in synovial tissue (ST) samples was performed by immunohistologic staining. Expression of α(1,2)-linked fucosylated angiogenic chemokine in synovial fluid (SF) was determined by immunoprecipitation and lectin blotting. To determine the angiogenic role of α(1,2)-linked fucosylated proteins in RA, we performed human dermal microvascular endothelial cell (HMVEC) chemotaxis and Matrigel assays using sham-depleted and α(1,2)-linked fucosylated protein-depleted RA SF samples. To examine the production of proangiogenic chemokines by FUT1 in HMVECs, cells were transfected with FUT1 sense or antisense oligonucleotides, and enzyme-linked immunosorbent assay was performed. We then studied mouse lung endothelial cell (EC) chemotaxis using wild-type and FUT1 gene-deficient mouse lung ECs.

Results: RA ST endothelial cells showed high expression of α(1,2)-linked fucosylated proteins compared to normal ST. The expression of α(1,2)-linked fucosylated monocyte chemoattractant protein 1 (MCP-1)/CCL2 was significantly elevated in RA SF compared with osteoarthritis SF. Depletion of α(1,2)-linked fucosylated proteins in RA SF induced less HMVEC migration and tube formation than occurred in sham-depleted RA SF. We found that blocking FUT1 expression in ECs resulted in decreased MCP-1/CCL2 and RANTES/CCL5 production. Finally, we showed that FUT1 regulates EC migration in response to vascular endothelial cell growth factor.

Conclusion: Our findings indicate that α(1,2)-linked fucosylation by FUT1 may be an important new target for angiogenic diseases such as RA.

Figure 1

Expression of α(1,2)-linked fucosylated proteins in ST ECs. A) The left panel shows ST staining with UEA-1 lectin and goat anti-UEA-1. The middle panel shows staining with rabbit anti-vWF. The right panel shows merging of the left panel and middle panel. The yellow color indicates α(1,2)-linked fucosylated proteins associated with ST ECs (original magnification 200×). Arrows indicate α(1,2)-linked fucosylated protein positive ECs. B) The percentages of α(1,2)-linked fucosylated protein positive vessels in each of the tissues were calculated and graphed. We determined the percentage of positive vessels in RA, OA and NL ST by counting the total number of yellow vessels, and divided this value by the total number of vWF positive vessels (green) in each tissue section. α(1,2)-linked fucosylated protein containing vessels were significantly higher in RA compared to NL ST (n=number of patients). IgG control staining for all three groups did not show staining (data not shown).

Figure 2

Expression of α(1,2)-linked fucosylated MCP-1/CCL2 in SFs. A) α(1,2)-linked fucosylated MCP-1/CCL2 was detected by immunoprecipitation and lectin blotting. Six RA and 6 OA patients SFs were analyzed. B) α(1,2)-linked fucosylated MCP-1/CCL2 in RA SFs was significantly elevated compared to OA SFs, normalized to total MCP-1/CCL2. The data is graphed as the ratio of the densitometry of α(1,2)-linked fucosylated MCP-1/CCL2 to total MCP-1/CCL2 shown in 2A for RA and OA SFs. C) The left panel shows that α(1,2)-linked fucosylated proteins in RA SF are depleted using UEA-1 conjugated agarose beads by lectin blotting. The right panel shows Coomassie Brilliant Blue staining. α(1,2)-linked fucosylated proteins are decreased after depletion. D) Sham and α(1,2)-linked fucosylated protein depleted RA SFs were measured in an MCP-1/CCL2 ELISA. Percent α(1,2)-linked fucosylated MCP-1/CCL2 of total MCP-1/CCL2 was 76 ± 9 % (mean ± SEM, n=10; n=number of patients).

Figure 3

Chemotactic and tube formation activity of α(1,2)-linked fucosylated proteins in RA SFs. A) HMVEC migration was measured in a chemotaxis assay using sham or α(1,2)-linked fucosylated protein depleted RA SFs. α(1,2)-linked fucosylated protein depletion resulted in a 54 ± 2% (n=4) reduction in RA SF-induced HMVEC migration. B) α(1,2)-linked fucosylated protein depleted RA SFs show less HMVEC tube forming activity compared to sham depleted RA SFs (n=number of patients). C) The left panel shows HMVEC tubes formed by sham depleted RA SFs. The right panel shows HMVEC tubes formed by α(1,2)-linked fucosylated protein depleted RA SFs. Arrows indicate tubes (original magnification 100x). D and E) Blocking antibody against MCP-1/CCL2 inhibited TNF-α-mediated HMVEC tube formation on Matrigel. Arrows indicate the number of tubes formed. Results are expressed as mean ± SEM and *p<0.05 was considered significant (n=number of replicates).

Figure 4

A) Purity of wild type MLECs via flow cytometry. Purity of C57BL/6 wild type MLECs was 76.3%, as determined by CD146 staining via flow cytometric analysis. The green graph represents CD146 positive cells while the black graph is control IgG. This graph is representative of two independent assays. Blocking fut1 in ECs reduces EC chemotaxis. B) A complete lack of fut1 expression in ECs reduces cell migration. Fut1 gene deficient MLECs had significantly less migration compared to wild type MLECs in response to mouse VEGF (10 and 100 nM). C) TNF-α (25 ng/ml) induced Flt1/VEGFR1 was markedly increased in wild type MLECs compared to fut1 gene deficient MLECs at 8, 16, and 24 hours (representative of two independent assays). D) Fut1 antisense ODN transfected HMVEC conditioned medium reduces HMVEC migration compared to fut1 sense ODN transfected HMVEC conditioned medium. The negative and positive controls were PBS and VEGF respectively (number of HMVECs migrating towards PBS was 5 ± 1, and VEGF was 40 ± 6; n=number of experiments).

Figure 5

Blocking fut1 expression in ECs reduces expression of proangiogenic chemokines. Cells were treated with IL-1β for 72 hours, and conditioned medium assayed. A) MCP-1/CCL2 in IL-1β stimulated fut1 antisense ODN transfected HMVEC conditioned medium is 21 ± 3 % (n=4) decreased compared to IL-1β stimulated fut1 sense ODN transfected HMVEC conditioned medium. B) RANTES/CCL5 in IL-1β stimulated fut1 antisense ODN transfected HMVEC conditioned medium is 41 ± 0 % (n=4) decreased compared to IL-1β stimulated fut1 sense ODN transfected HMVEC conditioned medium. C) MCP-1/CCL2 in IL-1β stimulated fut1 gene deficient MLEC conditioned medium is 15 ± 5 % (n=4) decreased compared to IL-1β stimulated wild type MLEC conditioned medium. D) RANTES/CCL5 in IL-1β stimulated fut1 gene deficient MLEC conditioned medium is 59 ± 1 % (n=4) decreased compared to IL-1β stimulated wild type MLEC conditioned medium (n=number of replicates). E) TNF-α (25 ng/ml) induced MCP-1/CCL2 expression was significantly less in fut1 gene deficient MLEC conditioned medium compared to wild type MLEC conditioned medium at 16 and 24 hours (n=number of samples).