C-Peptide Induces Chemotaxis of Human CD4-Positive Cells: Involvement of Pertussis Toxin–Sensitive G-Proteins and Phosphoinositide 3-Kinase

Human CD4⁺ cells were isolated from freshly drawn blood of healthy volunteers using gradient centrifugation with subsequent magnetic bead isolation, as previously described (12). After isolation, CD4⁺ cells were cultured in serum-free media for 16 h. T-cell chemotaxis was assayed under serum-free conditions in a 48-well microchemotaxis chamber (Neuroprobe). Wells in the upper and lower chambers were separated by a polyvinylpyrrolidone-free polycarbonate membrane (pore size 5 μm; Costar). CD4⁺ cells at a density of 5 × 10⁵/ml were incubated for 2.5 h with recombinant C-peptide or RANTES (Sigma) before migrated cells on the bottom face of the filter were stained and counted under the light microscope. Cells were counted in five random high-power fields per well. Checkerboard analysis was performed to differentiate chemotactic from chemokinetic activity. In some experiments, CD4⁺ cells were treated with pertussis toxin, wortmannin, or LY294002 (both PI 3-K inhibitors) or pretreated for 15 min with PPAR-γ–activating TZDs (rosiglitazone or pioglitazone).

Human T-cell membranes were prepared by incubating cells in a lysis buffer that contained 10 mmol/l HEPES (pH 7.5), 3 mmol/l MgCl₂, 40 μg/ml phenylmethylsulfonyl fluoride, 10 μg/ml leupeptin, 2 μg/ml pepstatin A, and 2 μg/ml aprotinin. After sonication, lysates were centrifuged at 15,000g for 10 min. The pellet was resuspended in radioimmunoprecipitation assay buffer (50 mmol/l Tris [pH 8.0], 150 mmol/l NaCl, 0.1% SDS, 0.5% DOC, 1% Nonidet P-40, and 1% Triton-X) and snap-frozen at −70°C.

Aliquots of the membrane fraction were boiled in Laemmli buffer before running on SDS-PAGE. Immunoblotting was performed by running the samples on SDS-PAGE and then electrotransferring onto nitrocellulose membranes (Amersham Pharmacia Biotech, Amersham, England), blocking with 5% skim milk in TBS buffer with 0.1% Tween 20 for 1 h, and incubating with 1:200 of the primary antibody (goat anti-human PI 3-Kγ [Santa Cruz] or rabbit anti-human PI 3-Kγ; rabbit anti-human p85 [Upstate]) and 1:10,000 dilution of the secondary antibody (anti-goat or anti-rabbit horseradish peroxidase [DAKO]). Development was done by using enhanced chemiluminescence reagents (Pearce, Rockford, IL) according to the manufacturer’s specifications.

After isolation, human T-cells were incubated for 16 h in RPMI medium without serum. Cells that were pretreated with or without wortmannin, pertussis toxin, or TZDs were stimulated with 10 nmol/l C-peptide, 10 nmol/l heat-inactivated C-peptide, or 100 pg/ml RANTES. Standard PI 3-K activity assays were performed (4) using goat anti-human PI 3-Kγ (Santa Cruz) or rabbit anti-human PI 3-Kγ (Santa Cruz) antibodies.

To examine the deposition of C-peptide in early arteriosclerotic lesions, we used postmortem artery specimen received from the multicenter cooperative project Pathobiological Determinants of Atherosclerosis in Youth, which collected material from 15- to 34-year-old trauma victims who were autopsied in forensic laboratories (13). Serial Cryostat sections of thoracic artery specimen were cut, air dried onto microscopical slides, and fixed in acetone at −20°C for 5 min. Staining for C-peptide was performed with a rabbit anti-human C-peptide antibody (Linco-Research). Human CD4⁺ cells were identified by staining with a mouse anti-human CD4 antibody (Dako). Negative controls used type- and class-matched IgG at similar concentrations. Sections were preincubated with PBS that contained 5% respective serum. Primary antibodies (1:50, anti–C-peptide; 1:20, anti-CD4) diluted in PBS that contained 3% serum were added for 1 h at room temperature. After washing with PBS, Alexa Fluor 488-coupled (Molecular Probes) and carboxymethylindocyanine 3-Cy3-coupled (Dianova) goat anti-mouse and anti-rabbit IgG were added as secondary antibodies (dilution 1:1,000) for 45 min. Images were recorded with a confocal laser-scanning microscope (Leica).

Results of the experimental studies are reported as mean ± SD. Differences were analyzed by one-way ANOVA followed by the appropriate post hoc test. P < 0.05 was regarded as significant.

To examine whether C-peptide induces CD4⁺ cell migration, we subjected human CD4⁺ cells to an in vitro chemotaxis assay using a modified Boyden chamber. Stimulation of CD4⁺ cells with C-peptide induced cell migration in a concentration-dependent manner with a maximal induction of 2.1 ± 0.5 at 10 nmol/l (P < 0.05 vs. unstimulated cells; n = 7). Heat inactivation of C-peptide (10 nmol/l) abolished its migratory effect (P < 0.05; n = 5), excluding endotoxin contamination as the responsible mechanism (Fig. 1A). The extent of C-peptide–induced CD4⁺ cell migration was similar to the effect of an established T-cell chemokine, RANTES, which led to a 2.1 ± 0.6-fold increase (P < 0.05; n = 6) in cell migration. It is interesting that combined stimulation of CD4⁺ cells with C-peptide and RANTES did not have an additive migratory effect (Fig. 1B).

To specify the true nature of C-peptide’s chemotactic activity, we conducted a checkerboard analysis using serial dilutions of C-peptide above and below the filter. As shown in Fig. 1C, CD4⁺ cell migration depended on the presence of a C-peptide gradient between the upper and lower face of the filter, suggesting that C-peptide induces CD4⁺ cell chemotaxis rather than chemokinesis.

To examine the effect of antidiabetic PPAR-γ–activating TZDs on CD4⁺ cell migration, we pretreated cells for 15 min with different concentrations of rosiglitazone or pioglitazone before performing stimulation with C-peptide. Both rosiglitazone and pioglitazone reduced C-peptide–induced CD4⁺ cell chemotaxis in a concentration-dependent manner, with a maximal reduction to baseline levels at 5 μmol/l (Fig. 2A). None of the TZDs affected cell viability, as determined by trypan blue staining (data not shown).

To examine whether the effects of rosiglitazone and pioglitazone on C-peptide–induced T-cell migration are PPAR-γ–mediated, we performed similar experiments with the established non-TZD PPAR-γ activator GW1929. Pretreatment of human CD4⁺ cells with GW1929 at 10 μmol/l abolished C-peptide–induced cell migration (Fig. 2B), suggesting that the TZD effects observed are PPAR-γ dependent.

To investigate further intracellular signaling pathways involved in C-peptide–induced CD4⁺ cell chemotaxis, we performed inhibition migration experiments. Treatment of human CD4⁺ cells with pertussis toxin inhibited C-peptide–induced CD4⁺ cell migration in a concentration-dependent manner, with a maximal reduction to baseline levels at 10 nmol/l pertussis toxin (Fig. 3A), suggesting that pertussis toxin–sensitive G-proteins are involved. Because some G-protein–coupled receptors activate PI 3-K and given the involvement of PI 3-K on T-cell motility (14–16), we next examined the role of PI 3-K. Treatment of CD4⁺ cells with wortmannin or LY294002, two PI 3-K inhibitors, completely inhibited C-peptide–induced CD4⁺ cell chemotaxis (Fig. 3B). Pertussis toxin, wortmannin, or LY294002 did not affect cell viability as examined by trypan blue staining (data not shown). These data suggest that C-peptide–induced CD4⁺ cell migration involves pertussis toxin–sensitive G-proteins as well as PI 3-K.

To gain further insight into C-peptide–induced PI 3-K activation, we examined the recruitment of PI 3-Kγ into the cell membrane. Incubation of CD4⁺ cells with C-peptide resulted in an association of PI 3-Kγ with CD4⁺ cell membranes. Two different anti–PI 3-Kγ antibodies showed similar results. Heat-inactivated C-peptide had no such effect. Consistent with former data on RANTES-induced recruitment of PI 3-K to the membrane, pretreatment of CD4⁺ cells with wortmannin did not inhibit C-peptide–mediated association of PI 3-Kγ with cell membranes (Fig. 4A). Given the involvement of class IA PI 3-Ks in insulin signaling (17), we investigated the role of this class of PI 3-Ks in C-peptide–induced cell activation. Western blot analysis of activated cells did not show an association of the regulatory subunit p85 with CD4⁺ cell membranes (data not shown), suggesting that C-peptide–induced PI 3-K activation involves only the p110γ isoform of PI 3-Ks.

Next, we examined the ability of PI 3-Kγ to generate phospholipids after C-peptide stimulation. C-peptide increased PI 3-Kγ activity in human CD4⁺ cells, whereas heat inactivation of C-peptide as well as pretreatment of cells with wortmannin abolished this effect. It is interesting that pretreatment of CD4⁺ cells with pertussis toxin also inhibited C-peptide–induced PI 3-Kγ activation, suggesting the involvement of pertussis toxin–sensitive proteins upstream of PI 3-K (Fig. 4B).

Because antidiabetic TZDs limited C-peptide–induced CD4⁺ cell chemotaxis, we investigated the effect of rosiglitazone and pioglitazone on C-peptide–mediated PI 3-K activation. Pretreatment (15 min) of human CD4⁺ cells with rosiglitazone or pioglitazone inhibited C-peptide’s stimulatory effect on PI 3-K activity (Fig. 4C).

Given the importance of CD4⁺ cell recruitment in early atherogenesis and C-peptide’s chemotactic effect on CD4⁺ cells shown here, we examined C-peptide deposition and CD4⁺ cell infiltration in early arteriosclerotic lesions of diabetic subjects. We used 21 postmortem thoracic aorta specimens from the Pathobiological Determinants of Atherosclerosis in Youth study in which subjects with HbA_1c levels ≥8% were classified as diabetic (13) and compared them with 21 age-, sex-, and risk factor–matched nondiabetic control subjects. Recently, we showed that the extent of C-peptide deposition in these diabetic subjects exceeds that of nondiabetic control individuals (4). Immunofluorescence staining of artery sections from both groups revealed that intimal CD4⁺ cells were detectable in 12 of 21 (57%) diabetic patients, but only in 2 of the 21 (9.5%) nondiabetic subjects. In these 12 diabetic subjects, immunofluorescence staining demonstrated intimal CD4⁺ cells in areas with C-peptide deposition (Fig. 5A–D and F). Staining of parallel sections with isomatched IgGs at similar concentrations showed no immunoreactivity (Fig. 5E and G), thus affirming the specificity of the detected signals.

The present study demonstrates that C-peptide, a cleavage product of proinsulin, exhibits chemotactic activity on human CD4⁺ cells by involving pertussis toxin–sensitive G-proteins and PI 3-K activation. In addition, the study reports intimal CD4⁺ cells in areas with C-peptide deposition in early atherosclerotic lesions of diabetic subjects as well as inhibition of C-peptide–induced CD4⁺ cell migration by antidiabetic TZDs. These data raise for the novel hypothesis that C-peptide may promote early atherogenesis in patients with type 2 diabetes by initiating or promoting CD4⁺ cell migration into developing lesions.

The effect of C-peptide on CD4⁺ cell migration is similar to the chemotactic action of the well-established T-cell chemokine RANTES. It is interesting that C-peptide and RANTES had no additive effect on CD4⁺ cell chemotaxis, suggesting that both agents share similar signaling pathways in these cells. Previous work has shown that C-peptide binds to a cell surface receptor, but this receptor remains unidentified (3). Work in renal cells implicates a pertussis toxin–sensitive G-protein–coupled receptor in mediating C-peptide’s effects (18), and our data showing an inhibitory effect of pertussis toxin on C-peptide–induced CD4⁺ cell migration suggest that similar pathways are involved here. Downstream of these pertussis toxin–sensitive G-proteins, C-peptide stimulation results in translocation of PI 3-Kγ to the cell membrane as well as an increase in PI 3-Kγ activity. These results are in accordance with data in Swiss 3T3 fibroblasts (19), where C-peptide has been shown to activate PI 3-K. Previous work has reported that insulin signaling involves class IA PI 3-Ks (17), but our data, demonstrating a lack of an association of p85 with cell membranes, suggest that C-peptide–induced activation of CD4⁺ cells involves only PI 3-Kγ (the class IB enzyme) but not class IA PI 3-Ks.

Two TZDs, antidiabetic drugs used to treat patients with type 2 diabetes, completely inhibited C-peptide’s chemotactic effect. Previous work (9,11) has shown that TZDs modulate proliferation as well as TH1-cytokine expression of human CD4⁺ cells. Our study extends the knowledge about TZDs’ effects in these cells by demonstrating an inhibition of T-cell migration. TZDs are activators of the nuclear transcription factor PPAR-γ, thus regulating the expression of various target genes. It is interesting that a 15-min pretreatment of CD4⁺ cells with TZDs already prevented the rapid effect of C-peptide on PI 3-K activation, raising the question of a nontranscriptional effect of these agents in this context. Recent work by Chawla et al. (20) found that some effects of TZDs could occur independent of the presence of PPAR-γ, at least in cells of the monocytic lineage, and such mechanisms could potentially explain the rapid inhibition of PI 3-K activation observed here. However, our data showing similar inhibitory action on CD4⁺ cell migration by the non-TZD PPAR-γ activator GW1929 suggest that the effects of TZDs are PPAR-γ mediated. Further work should focus on the molecular mechanism of the effects of TZDs on PI 3-K activation.

The present study suggests an active role of C-peptide in the development of atherosclerotic lesions in patients with type 2 diabetes by showing intimal CD4⁺ cells in areas with C-peptide deposition in early atherosclerotic lesions of diabetic subjects. During endothelial dysfunction with increased endothelial permeability, C-peptide might deposit in the arterial intima in diabetic patients and subsequently attract CD4⁺ cells to migrate into the vessel wall. Still, future studies should examine C-peptide deposition and CD4⁺ cell infiltration in arteriosclerotic lesions from type 1 diabetic patients to link our observational study to possible mechanisms of disease.

This work was supported by grants from the Deutsche Forschungsgemeinschaft (SFB 451, Project B9), the Else-Groener-Fresenius-Stiftung, and the IZKF Ulm (Project G6) (to N.M.).

The authors thank Miriam Grüb, Helga Buch, and Renate Durst for excellent technical assistance.