Glucagon-like peptide-1(1-37) inhibits chemokine-induced migration of human CD4-positive lymphocytes

The present study examined the effect of GLP-1(1-37) on chemokine-induced CD4-positive lymphocyte migration as an early and critical step in atherogenesis. Pretreatment with GLP-1(1-37) reduced the SDF-induced migration of isolated human CD4-positive lymphocytes in a concentration-dependent manner. Similar effects were seen when RANTES was used as a chemokine. GLP-1(1-37)’s effect on CD4-positive lymphocyte migration was mediated through an early inhibition of chemokine-induced PI-3 kinase activity. Downstream, GLP-1(1-37) inhibited SDF-induced phosphorylation of MLC and cofilin and limited f-actin formation as well as ICAM3 translocation. Furthermore, exendin-4 inhibited SDF-induced migration of CD4-positive lymphocytes similarly to GLP-1(1-37), and transfection of these cells with GLP-1 receptor siRNA abolished GLP-1(1-37)’s action on chemokine-induced ICAM3 translocation, suggesting an effect mediated via the GLP-1 receptor. Thus, GLP-1(1-37) inhibits chemokine-induced CD4-positive lymphocyte migration by inhibition of the PI3-kinase pathway and via the GLP-1 receptor. This effect provides a potential novel mechanism for how GLP-1(1-37) may modulate vascular disease.

Whereas various effects of metabolically active GLP-1(7-36) on multiple organs have been suggested, only limited evidence exists about the actions of GLP-1(9-36) on the cardiovascular system (for review see [5]). Recently, a role of GLP-1  in preventing ischemia-reperfusion injury has been suggested in isolated perfused mouse hearts [6]. Work from another group showed that the degradation product GLP-1(9-36) improves left ventricular and systemic hemodynamics using a canine model of dilated cardiomyopathy [7]. We have previously demonstrated that GLP-1(1-37), a long and untruncated form of GLP-1, inhibits chemokineinduced CD4-positive lymphocytes migration, an early and critical step of atherogenesis [8,9]. This effect was shown to be dependent on the GLP-1 receptor protein and exendin-4, a specific GLP-1 receptor agonist, had similar effects on T-cell migration. However, while GLP-1  is the physiologic agonist of the GLP-1 receptor and increases plasma concentrations of insulin after food intake, the metabolite GLP-1(9-36) is an antagonist of the GLP-1 receptor and metabolically inactive [10,11]. The effects of the metabolite GLP-1(9-36) on human CD4-positive lymphocyte migration are unknown.
As patients with type 2 diabetes are a high-risk population for the development of a severe and diffuse atherosclerosis, the potential action of these drugs as modulators of vascular disease is of significant clinical interest. Given the different pharmacodynamics of GLP-1 analoga and DPP-IV inhibitors, it is crucial to better understand the effects of the degradation product GLP-1(9-36) on mechanisms involved in atherogenesis. Thus, we investigated the effects of GLP-1(9-36) on the chemokine-induced migration of human CD4-positive lymphocytes as an early and critical step in atherogenesis.

Cell Culture
Buffy coats were obtained completely anonymized from the Department of Transfusion Medicine, University Hospital of Aachen. As the buffy coats were obtained completely anonymized, Figure 1. GLP-1(9-36) inhibits chemokine-induced migration of isolated human CD4-positive lymphocytes. A. and B.: Cells were pretreated with GLP-1(9-36) for 30 minutes at concentrations indicated before migration experiments using SDF-1 (100 ng/mL) (A) or RANTES (100 ng/mL) (B) were performed for three hours in the modified Boyden chamber. The data show that GLP-1(9-36) inhibits both SDF-and RANTESinduced migration of human CD4-positive lymphocytes. C: Scrambled GLP-1(9-36) does not reduce SDF-1-induced T-cell migration. Cells were pretreated with 1 nM GLP-1(9-36) and 1 nM scrambled GLP-1(9-36) for 30 minutes before migration experiments using SDF-1 (100 ng/mL) were performed for three hours in the modified Boyden chamber. Data are expressed as fold induction of unstimulated cells. Bars represent mean6SEM (n = 7); ***p,0.001, **p,0.01, *p,0.05 compared with SDF-stimulated cells. doi:10.1371/journal.pone.0058445.g001 no ethical approval was necessary for this study in Germany. Human CD4-positive lymphocytes were isolated from buffy coats as has been previously described [9,12]. Briefly, a Ficoll-Histopaque (Sigma) gradient centrifugation was used to obtain mononuclear cells and subsequent negative selection of CD4positive T-cells was performed by magnetic bead separation (Miltenyi Biotech). The purity of CD4-positive lymphocytes was .97% as determined by flow cytometry.

In vitro Cell Migration Assay
For in vitro cell migration assays, human CD4-positive lymphocytes were cultured in serum-free media for 16 h. A 48well microchemotaxis chamber (Neuroprobe) was used to investigate T-cell chemotaxis [9]. Wells in the upper and lower chamber were separated by a polyvinylpyrrolidone-free polycarbonate membrane (pore size 5 mm; Whatman). CD4-positive cells were pre-treated for 30 min with GLP-1(9-36) (Phoenix) at the concentrations indicated before 3 hours of incubation with SDF-1 (Upstate) or RANTES (PeproTech). Migrated cells attached to the bottom face of the filter, which had been pre-incubated with collagen, and were subsequently stained and counted under the light microscope. Furthermore, a scrambled GLP-1(9-36), which consists of the same amino acids as GLP-1(9-36) but in a random order (Val-Arg-Asp-Ile-Ala-Leu-Glu-Lys-Gly-Gly-Thr-Ala-Glu-Leu-Phe-Phe-Val-Tyr-Ala-Gly-Lys-Thr-Gln-Ser-Ser-Ser-Trp-Glu-NH3, Thermo Fisher Scientific) was included in the migration experiments as a negative control.

Western Blot Analysis
Standard western blot analysis was performed using specific antibodies against p-cofilin, p-MLC (Cell Signaling) and human GLP-1 receptor (Upstate). Western blot images were analyzed by Carestream molecular imaging software (Carestream). For determination of protein levels, bands were normalized to ß-actin bands.

Quantitative RT-PCR
RNA was extracted using the RNeasy extraction kit (Quiagen). cDNA was generated by reverse transcription (SuperScript reverse transcriptase, Invitrogen). Quantitative RT-PCR was performed using a commercially available Taqman Gene Expression Assay for the full length GLP-1 receptor (Applied Biosystems). Samples were normalized to a control gene (GAPDH, forward primer: 59-GCC TCG TCC CGT AGA CAA AA-39; reverse primer: 59-TGG CAA CAA TCT CCA CTT TGC-39).

CXCR4 Expression and Vitality
Standard flow cytometry was performed using a CANTO II (BD) and a PE-labeled antibody against human CXCR4 (BD Pharmigen). For the determination of vitality, 1610 6 cells were treated with the same conditions as for the migration experiments and stained with trypan blue (Sigma Aldrich). Vital cells were counted in a Neubauer chamber.

Phosphatidylinositol Kinase Assay
A PI-3 kinase activity assay was performed as has been previously described [12]. Briefly, human CD4-positive T-cells were incubated after isolation for 16 hours in RPMI medium without serum. Cells were pretreated for 30 min with and without GLP-1(9-36) and were then stimulated with 100 ng/mL SDF-1 for 5 min. The PI 3-kinase activity assays were performed using an goat anti-human p85 antibody (Santa Cruz) [13].

ICAM3 Staining
To determine ICAM3 translocation, immunofluorescence staining for ICAM3 was performed on human CD4-positive lymphocytes as has been previously described [9,13]. 3610 5 CD4 positive lymphocytes were incubated in 4-well plates (Costar Corp.) in a final volume of 1 ml complete medium on coverslips coated with collagen. Prior to treatment with SDF-1 (100 ng/mL, 30 min), cells were pre-incubated with GLP-1(9-36) for 30 min. CD4-positive lymphocytes were then fixated with 3.7% (wt/vol) paraformaldehyde in PBS and rinsed in TBS. Cells were incubated with a specific mouse anti-human ICAM3 antibody (BD Pharmingen) and after washing with PBS, an carboxymethylindocyanine 3-Cy3-coupled goat anti-mouse antibody (Jackson) was added for 30 min. After washing slides were counterstained with DAPI. Images were recorded with a fluorescence microscope (Leica) and cells in 5 randomly selected fields (, 100 cells) were analysed. If ICAM-3 was located evenly on the outer side of the plasma membrane, T-cells were considered to be resting. When a clear clustering of ICAM3 at the uropod was visible, ICAM3 translocation was considered to be present.

cAMP Assay
For determination of intracellular cAMP concentrations 1610 6 isolated, human CD4-positive lymphocytes were treated with 10 nmol/L GLP-1(9-36) (Phoenix) or 100 ng/ml SDF-1 (Upstate) for times indicated. 100 ng/ml Forskolin (Calbiochem) was used as a positive control. The cAMP competitive ELISA (Thermo Scientific) was performed according to the manufacturers protocol. Briefly, cAMP was extracted by adding 0.1 mol/l HCl with 0.5% triton X-100 to the cells. After centrifugation at 600 g for 10 min, Table 1. GLP-1 does not decrease the expression of the chemokine receptor CXCR4 and does not influence cell viability; n = 4; the data is displayed as mean 6 SEM.

Transfection with siRNA of the GLP-1 Receptor
Transfection of human CD4-positive lymphocytes with siRNA of the GLP-1 receptor has been performed as has been previously described [9]. Briefly, specific and negative siRNAs were purchased from Dharmacon. CD4-positive lymphocytes were isolated from blood. After isolation, cells were incubated for one hour in RPMI with 5% human serum. The human T-cellnucleofector kit (Amaxa Biosystems) was used according to the manufacturer's protocol employing 200 nM GLP-1 receptor siRNA. After transfection, cells were incubated in 12-well plates with RPMI containing 5% human serum for 48 hours. Untreated cells as well as cells treated with the amaxa nucleofector solution and negative siRNA were used as negative controls.

Statistical Analysis
The data are reported as mean6standard error of the mean (SEM). Differences were analyzed by one-or two-way-ANOVA followed by the appropriate post-hoc test where applicable. A pvalue ,0.05 was regarded as statistically significant.
The Effect of GLP-1(9-36) on Chemokine-induced Migration of Human CD4-positive Lymphocytes is not Mediated by cAMP Signalling As cAMP has been suggested to be involved in GLP-1(9-36) signalling [18], we performed a quantitative cAMP assay of isolated CD4-postitive lymphocytes in order to investigate whether the effect of GLP-1(9-36) on chemokine-induced CD4-positive lymphocyte migration is caused by cAMP signalling. In this assay forskolin was used as a positive control and resulted in a 5.2-fold increase in cAMP levels (p,0.001). No change of cAMP was observed with GLP-1(9-36) in a time-response analysis (Fig. 4 B). Furthermore, no increase in cAMP was observed when cells were incubated with SDF-1 plus GLP-1(9-36) under the same conditions as in the migration experiments. Therefore, GLP-1(9-36) inhibits chemokine-induced CD4-positive lymphocyte migration independent of cAMP signalling.

Discussion
This study demonstrates that the degradation product GLP-1(9-36) inhibits chemokine-induced migration of human CD4-positive lymphocytes, which is mediated through an early inhibition of PI-3 kinase activity with subsequent inhibition of MLC and cofilin phosphorylation as well as f-actin formation and ICAM3 translocation. The action of GLP-1(9-36) on CD4-positive lymphocyte migration and the PI3-kinase signaling cascade is not mediated by cAMP signaling or the GLP-1 receptor protein.
Current new treatment options for patients with type 2 diabetes include DPP-IV inhibitors, which decrease the rapid degradation of GLP-1  to the metabolite GLP-1 . Given that patients with type 2 diabetes are a high-risk population for cardiovascular events [3], beneficial vascular effects of the metabolite GLP-1(9-36) on the cardiovascular system are therefore of potential clinical interest. Recently, a role of GLP-1(9-36) in preventing ischemiareperfusion injury has been suggested in isolated perfused mouse hearts [6]. Nikolaidis et al. showed that the degradation product GLP-1(9-36) improves left ventricular and systemic hemodynamics using a canine model of dilated cardiomyopathy [7]. However, nothing is known about the effects of GLP-1(9-36) on CD4positive lymphocyte migration, an early and critical step in atherogenesis [8]. Our study extends the current knowledge on GLP-1(9-36) by demonstrating an inhibition of chemokineinduced CD4-positive lymphocyte migration, an effect which could potentially contribute to the modulation of vascular inflammation in diabetic patients.
The effect of GLP-1(9-36) on human CD4-positive lymphocyte migration is mediated by an early inhibition of PI-3 kinase activity, an important step in chemokine-induced human CD4-positive lymphocyte migration. PI-3 kinase activity results in the activation of small Rho GTPases. Downstream of these GTPases, the phosphorylation of cofilin is important to allow for actin polymerization of migrating cells, while the phosphorylation of MLC contributes to cell contraction at the uropod of migrating Tcells. In this study, GLP-1(9-36) reduces chemokine-induced phosphorylation of cofilin with subsequent f-actin formation. Moreover, GLP-1(9-36) limits the phosphorylation of MLC and inhibits SDF-1-mediated ICAM3 translocation to the uropod of migrating cells, thus inhibiting two crucial steps in T-cell migration.
GLP-1(9-36) diminishes both SDF-1-as well as RANTESinduced lymphocyte migration and thus the effect of GLP-1(9-36) on cell migration does not appear to be depended on the chemotactic stimulus employed. However, the concentrations of GLP-1(9-36) necessary to inhibit T-cell migration were physiological for RANTES, but supraphysiological for SDF. This may be caused by the fact that SDF was observed to be a much stronger inducer for T-cell migration (3.43-fold for SDF vs 1.98-fold for RANTES), with higher GLP-1(9-36) concentrations necessary to inhibit SDF compared to RANTES -induced T-cell migration.
Moreover, the GLP-1(9-36) concentrations employed in our study correspond to the concentrations used in previous experiments [7,19] and the effects observed are not caused by alterations in cell viability or the chemokine receptor expression. Furthermore, the phenomenon demonstrated is not due to unspecific effects of the peptide nor caused by endotoxins as shown by experiments with scrambled GLP-1(9-36) and heat-inactivated GLP-1(9-36). The effect of GLP-1(9-36) on chemokine-induced CD4-positive lymphocyte migration and the PI-3 kinase signaling cascade was not mediated by intracellular levels of cyclic AMP, which has previously been suggested to be involved in GLP-1(9-36) signaling [18]. Thus, signaling in isolated human CD4-positive lymphocytes is therefore different from rodent neonatal cardiomyocytes, in which GLP-1(9-36) results in increased levels of cAMP and increased phosphorylation of the PI-3 kinase target Akt [18]. Furthermore, we could demonstrate that the inhibitory action of the degradation product GLP-1(9-36) on CD4-positive lymphocyte migration is independent of the GLP-1 receptor protein, which has previously been shown to mediate the inhibitory effect of untruncated GLP-1 on chemokine-induced CD4-positive lymphocyte migration [9].
Potential mechanisms by which GLP-1(9-36) may inhibit PI3kinase activity may be at the level of the chemokine receptor via interaction with a so far unknown GLP-1(9-36) receptor, mediated by G-proteins, caused by dimerization of and with PI3-kinase or via modification of the regulatory p85 or catalytic p110 subunit.
Thus, the present study shows that GLP-1(9-36) inhibits chemokine-induced CD4-positive lymphocyte migration by inhibition of the PI3-kinase pathway independent of cAMP and GLP-1 receptor signaling. Further studies are needed to assess whether such effects may be clinically relevant for type 2 diabetes patients treated with DPP-IV inhibitors.