C-reactive protein and vein graft disease: evidence for a direct effect on smooth muscle cell phenotype via modulation of PDGF receptor-β

Abstract

Plasma C-reactive protein (CRP) concentration is a biomarker of systemic atherosclerosis and may also be associated with vein graft disease. It remains unclear whether CRP is also an important modulator of biological events in the vessel wall. We hypothesized that CRP influences vein graft healing by stimulating smooth muscle cells (SMCs) to undergo a phenotypic switch. Distribution of CRP was examined by immunohistochemistry in prebypass human saphenous veins (HSVs, n = 21) and failing vein grafts (n = 18, 25–4,400 days postoperatively). Quiescent HSV SMCs were stimulated with human CRP (5–50 μg/ml). SMC migration was assessed in modified Boyden chambers with platelet-derived growth factor (PDGF)-BB (5–10 ng/ml) as the chemoattractant. SMC viability and proliferation were assessed by trypan blue exclusion and reduction of Alamar Blue substrate, respectively. Expression of PDGF ligand and receptor (PDGFR) genes was examined at RNA and protein levels after 24–72 h of CRP exposure. CRP staining was present in 13 of 18 diseased vein grafts, where it localized to the deep media and adventitia, but it was minimally detectable in most prebypass veins. SMCs pretreated with CRP demonstrated a dose-dependent increase in migration to PDGF-BB (P = 0.02), which was inhibited by a PDGF-neutralizing antibody. SMCs treated with CRP showed a dose-dependent increase in PDGFRβ expression and phosphorylation after 24–48 h. Exogenous CRP had no effect on SMC viability or proliferation. These data suggest that CRP is detectable within the wall of most diseased vein grafts, where it may exert local effects. Clinically relevant levels of CRP can stimulate SMC migration by a mechanism that may involve upregulation and activation of PDGFRβ.

vein graft failure remains a formidable challenge among patients with critical limb ischemia or coronary ischemia requiring revascularization. Even with autogenous vein conduit, the most durable small vessel substitute, 30–50% of lower extremity vein bypass grafts fail within 5 years (9). Vein graft failure is most commonly ascribed to myointimal hyperplasia, which involves a smooth muscle cell (SMC) switch from a quiescent/contractile phenotype to a migratory, proliferative, and synthetic phenotype (11).

Chronic baseline inflammation in humans, as measured by the high-sensitivity C-reactive protein (hsCRP) assay, has emerged as a novel biomarker that adds incremental predictive ability above traditional cardiovascular risk factors for primary and secondary cardiovascular events. Inflammation has been implicated not only in de novo atherosclerosis-related complications, i.e., myocardial infarction and stroke, but also in restenosis of arteries after balloon injury and in stenosis and failure of vein grafts used for arterial reconstructions (10, 21, 31, 44–46). We recently showed that plasma hsCRP concentrations are associated with primary vein graft failure in individuals undergoing lower extremity bypass procedures (40) and that inflammation, as assessed by hsCRP, influences the normal remodeling pattern of arterialized veins (41).

A central question arising from this work is whether CRP is not only a biomarker but, also, a direct biomodulator of biological events in vascular cells and tissues. CRP may contribute to vascular wall inflammation through multiple posited mechanisms: it is deposited in the wall at sites of endothelial injury, where it opsonizes lipids and activates the classical complement cascade; it induces expression of several cell adhesion molecules, as well as tissue factor; it mediates LDL uptake by macrophages; it induces monocyte recruitment into the arterial wall; it enhances production of monocyte chemoattractant protein-1; it promotes endothelial dysfunction and impairs endothelial progenitor cell survival and differentiation; and it enhances reactive oxygen species generation (7, 16, 18, 27, 48, 54, 55, 58, 65). Specific to the question of a potential role in vein graft disease, CRP has been immunolocalized to the media and adventitia in a canine vein graft model (19).

In the present study, we investigated the hypothesis that CRP plays a direct role in vein graft hyperplasia, and we demonstrate an induction of a promigratory SMC phenotype that may be related to alterations in platelet-derived growth factor (PDGF) signaling.

MATERIALS AND METHODS

Lower extremity vein and vein graft specimens.

Specimens of human lower extremity veins and vein grafts were collected from discarded conduits at the time of coronary or lower extremity bypass grafting and lower extremity bypass graft revision, respectively. Specimens were fixed in 10% neutral buffered formalin (Sigma-Aldrich), embedded in paraffin, sectioned (6 μm), and collected on Superfrost/Plus slides (Fisher Scientific). A total of 21 human greater saphenous veins and 18 saphenous vein grafts were available for the study. Serial sections were cut and used for hematoxylin-and-eosin staining, elastin staining, and immunohistochemistry. Institutional Review Board approval was obtained for the collection of all human specimens and relevant clinical data as part of a tissue repository.

Immunohistochemistry, elastin staining, and image analysis.

The mouse monoclonal antibody (clone CRP-8) directed against denatured and native human CRP (1:100 and 1:250 dilution) was obtained from Sigma-Aldrich; it has no cross-reactivity with human serum amyloid P, haptoglobin, α₁-acid glycoprotein, IgG, or CRP from Limulus (54). A mouse monoclonal anti-human α-smooth muscle actin antibody (1:200 dilution) was obtained from Biogenex, a mouse monoclonal anti-human CD31 antibody (1:30 dilution) from Dako, and a rabbit polyclonal anti-human PDGF receptor (PDGFR)-β antibody (1:1,000 dilution) from Santa Cruz Biotechnology. α-Actin, CD31, PDGFRβ, or CRP immunolocalization was achieved with an indirect immunoperoxidase technique and enhanced diaminobenzidine (Dako) as the peroxidase substrate. Hematoxylin was the counterstain. The intensity of CRP staining was graded in a blinded manner as none, weak, or strong by two investigators on two separate occasions; if staining was detected, the vessel layer in which the staining was present was noted.

For morphometric analysis, elastin was stained using the Verhoeff-Van Gieson technique. All sections were dehydrated and mounted permanently with Cytoseal (Richard-Allan Scientific). Microscopy was performed with a Nikon Eclipse 80i microscope and Spot Flex digital camera (Micro Video Instruments). Neointima, media, total wall area, and thickness were measured from ×20 and ×40 images of sections using Image J software (National Institutes of Health).

Cell culture.

HSV SMCs were cultured from explants of veins obtained at the time of bypass surgery and maintained in high-glucose DMEM (GIBCO) containing 10% fetal calf serum, glutamine, and penicillin-streptomycin, as described elsewhere (60). Immunophenotyping after passage 3 with smooth muscle-specific α-actin staining revealed that cultured cells were completely SMCs. Cells were used between passages 4 and 8 in all experiments. All cell culture experiments were performed at least three times.

CRP preparation.

Commercial purified human CRP (Polysciences) was supplied in 50 mM Tris·HCl (pH 7.5), 250 mM NaCl, 5 mM CaCl₂, and 0.1% NaN₃. NaN₃ was removed by dialysis of CRP against two 1-liter changes of the same buffer without NaN₃ with use of a Slide-a-Lyzer dialysis cassette (model 10K MWCO, Pierce). This technique has been previously described to be sufficient for the removal of NaN₃ (26, 57). The concentration of dialyzed CRP was quantified using a modified Lowry assay (DC Assay, Bio-Rad), and purity was grossly confirmed by separation of the CRP preparation by SDS-PAGE and confirmation of the presence of a single band at ∼25 kDa on a Coomassie-stained gel (data not shown). The level of endotoxin in CRP preparations after dialysis was determined using Limulus amebocyte lysate Pyrogent (Cambrex; sensitivity 0.06 endotoxin unit/ml) following the manufacturer's instructions. All experiments were performed with two different lots of CRP.

Semiquantitative and quantitative real-time RT-PCR.

For determination of the effect of CRP on gene expression, HSV SMCs were serum starved for 48 h before the addition of increasing doses (0, 5, and 50 μg/ml) of dialyzed CRP for 1 and 24 h. Cells were lysed with TRIzol reagent (Invitrogen), and total RNA was purified according to the manufacturer's instructions and quantified by spectrophotometry. RT was performed for semiquantitative RT-PCR using 1 μg of total RNA and random hexamer primers (first-strand cDNA synthesis kit, Amersham Biosciences) and for real-time quantitative RT-PCR using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems) according to the manufacturers' instructions. Gene-specific primers for PDGFRα, PDGFRβ, PDGF-A, PDGF-B, and GAPDH and PCR amplification conditions for semiquantitative RT-PCR have been described previously (60). A 474-bp product from PDGF-D was amplified using Accuprime Taq polymerase (Invitrogen) with the forward primer 5′-CCATCCAGGTGAAAGGAAACG-3′ and the reverse primer 5′-TTTTTGTCCAGAGCATCCGC-3′ under the following conditions: one cycle at 94°C for 2 min; 35 cycles at 94°C for 30 s, 56°C for 30 s, and 68°C for 1 min; and 1 cycle at 68°C for 3 min. Inventoried Taqman gene expression assays (Applied Biosystems) were used for quantitative real-time PCR (Hs00182163_m1 for PDGFRα, Hs00183486_m1 for PDGFRβ, Hs00234994_m1 for PDGF-A, and Hs00234042_m1 for PDGF-B). Samples were run in triplicate in 25-μl reactions using 2× TaqMan Universal PCR Master Mix (Applied Biosystems) on a sequence detector (model 7900HT, Applied Biosystems). Target gene expression levels are expressed relative to GAPDH expression levels using the comparative cycle threshold (C_T) method (34).

Transwell migration assay.

HSV SMC migration was determined using gelatin-coated modified Boyden chambers (Costar) with 8-μm pores in 24-well plates. SMCs were serum starved for 48 h, treated with vehicle or dialyzed CRP (20 μg/ml) for an additional 48 h, trypsinized lightly with 0.05% trypsin, and seeded in Transwell filters at a density of 20,000 cells/well. PDGF-BB (20 ng/ml; R & D Systems) in serum-free medium was added as a chemoattractant in the bottom wells. In some cases, neutralizing antibody to PDGF-BB (5 μg/ml; R & D Systems) was added to the top and bottom wells. All treatments were done in duplicate or triplicate. After 6–9 h, unmigrated cells were scrubbed and rinsed twice from the top of the membranes with cotton swabs and PBS. Membranes were fixed in methanol, stained with 4′,6-diamidino-2-phenylindole (DAPI; Invitrogen), and carefully cut from the inserts. Membranes were mounted on slides in PBS and photographed under high power (×400) with a fluorescent microscope. Three to five high-power fields were photographed per membrane. Nuclei were counted, and migration is expressed relative to that of the control group (no chemoattractant).

Cell viability and proliferation.

HSV SMC viability and proliferation were assessed as functions of trypan blue exclusion, morphology of fluorescent-stained nuclei, and reduction of Alamar Blue substrate (60). For trypan blue exclusion, HSV SMCs were plated in 96-well plates and allowed to adhere overnight. On the following day, the wells were rinsed with PBS, and vehicle, 5 μg/ml CRP, or 50 μg/ml CRP was added in low-serum conditions (0.1%) in triplicate wells. After 48 h, cells were simultaneously trypsinized and stained with trypan blue (Sigma-Aldrich). Viable cells (which exclude dye) were counted, and viability was expressed as percent difference in the absolute number of viable cells from vehicle-treated wells. For detection of morphological changes in nuclei that accompany apoptosis (such as chromatin condensation), HSV SMCs were plated in chamber slides and treated with vehicle or CRP in triplicate as described above. After 48 h, cells were rinsed, permeabilized with Triton X-100 (Sigma-Aldrich), and stained with DAPI. Coverslips were applied, and slides were analyzed by fluorescent microscopy. Three high-powered fields (×400) of each well were photographed. Vehicle- and CRP-treated nuclei were compared qualitatively. For the Alamar Blue assay, HSV SMCs were plated in 24-well plates and allowed to adhere overnight in normal-serum conditions. On the following day (day 0), the wells were rinsed with PBS, and vehicle or CRP was added in low-serum conditions (0.1%) as described above. The positive control group consisted of cells that continued to grow in normal-serum conditions. All treatments were done in triplicate wells. Reduction of Alamar Blue was measured on days 0, 2, 4, and 8, converted to cell number based on a standard curve, and expressed as percent increase over vehicle-treated wells. Medium was changed and CRP was replenished every 48 h.

Immunoprecipitation and Western blot analysis.

For determination of the effect of CRP on PDGFRβ protein levels, HSV SMCs were serum starved for 48 h and treated with increasing doses (0, 5, and 50 μg/ml) of dialyzed CRP for an additional 24 or 48 h. Cells were lysed in buffer [25 mM HEPES (pH 7.5), 150 mM NaCl, 1% Triton X-100, 0.1% β-mercaptoethanol, 10% glycerol, 5 mM EDTA, and 5 mM EGTA] containing protease inhibitor (Roche) and phosphatase inhibitor (Pierce) cocktails. Lysates were cleared by centrifugation, and protein concentration was determined using a modified Lowry assay (RC DC assay, Bio-Rad). Total protein (30 μg) was boiled briefly, separated on 7.5% SDS-PAGE gels, and transferred to nitrocellulose membranes using a semidry technique. Membranes were blocked in 5% milk and incubated with a rabbit polyclonal antibody to PDGFRβ (1:200 dilution; Santa Cruz Biotechnology) overnight. Bound antibody was detected using a horseradish peroxidase-conjugated goat anti-rabbit antibody (1:10,000 dilution; Jackson ImmunoResearch) with enhanced chemiluminescent substrate (West Pico, Pierce). β-Actin (1:10,000 dilution; Sigma-Aldrich) served as the loading control.

For determination of the effect of CRP on phosphorylation levels of PDGFRβ, total receptor was immunoprecipitated from 250 μg of cleared total protein lysates with 1 μg of anti-PDGFRβ overnight at 4°C. Protein A/G beads (Pierce) were added, and, after 1 h, protein A/G-antibody conjugates were pelleted, washed, and disrupted by boiling. Western blotting was performed as described above using mouse monoclonal antibody to phosphorylated tyrosine (1:200 dilution; Santa Cruz Biotechnology). Total PDGFRβ served as the loading control. All blots were scanned. Bands were quantified by densitometry (GelDoq, Bio-Rad) and normalized relative to bands in control (non-CRP-treated) groups.

Statistical analysis.

Values are means ± SE unless otherwise specified. Student's t-test and ANOVA were used to assess differences between two and more than two groups, respectively. Differences were considered significant at P < 0.05.

RESULTS

Immunolocalization of CRP in HSVs and vein graft lesions.

If CRP is to be considered a potential direct modulator of biological processes within the vein graft, its presence should be identifiable in the preoperative substrate (saphenous vein) or in the wall of mature grafts. Selected clinical data on the subjects from whom preimplantation veins (n = 21) or vein grafts (n = 18) were analyzed are shown in Table 1. CRP was minimally detectable in preimplantation veins (Fig. 1, A–D). In 72% (13) of failed lower extremity vein grafts, CRP immunolocalized to the adventitia and deep media (Fig. 1, E–H). CRP was weakly present or not detected in 28% of these grafts (5). There was no observable correlation between intensity of staining and age of graft or clinical characteristics, including patient age, sex, history of diabetes mellitus, renal insufficiency, or indication for bypass (claudication vs. limb salvage).

Table 1.

Selected clinical data on preimplantation vein and failing lower extremity bypass graft specimens

Parameter	Value
Preimplantation veins (n = 21)
Patient age (mean ± SD)	65.1±10.9 yr
Male patients	11 (53%)
Patients with diabetes mellitus	7 (35%)
Patients on statin therapy preoperatively	6 (21%)
Indication for bypass
Critical limb ischemia	13 (61.9%)
Unstable angina	5 (23.8%)
3-vessel coronary artery disease	3 (14.3%)
Failing lower extremity bypass grafts (n =18)
Patient age, yr (mean ± SD)	67±9.63 yr
Male patients	13 (72.2%)
Age of graft (median, interquartile range)	220 (106–561.6) days
Patients with diabetes mellitus	15 (83.3%)
Indication for vein graft revision
Critical stenosis	9 (50%)
Nonhealing ulcers	6 (33.3%)
Graft occlusion	2 (11.1%)
Other	1 (5.6%)

Fig. 1.

Representative immunolocalization of C-reactive protein (CRP) in failing lower extremity bypass graft section and preimplantation greater saphenous vein. A–D: sequential sections of a representative greater saphenous vein. E–H: sequential sections of a failing lower extremity bypass graft (72-day-old graft). L, lumen. A and E: negative (IgG) staining control. B and F: CRP staining. Note CRP deposition in deep media and adventitia in the graft. C and G: α-smooth muscle actin staining. Area of most intense CRP staining in the graft is localized in deep media and adventitia. D and H: CD31 staining. There is no colocalization of CRP with endothelial cells in the vein or the graft. Images are representative of 21 preimplantation veins and 18 failing grafts. Original magnification ×40.

To determine the cellular localization of CRP staining, adjacent sections were stained with CD31, an endothelial cell marker, and with α-smooth muscle actin, a marker of SMC differentiation and myofibroblasts. We observed no colocalization of CRP and CD31 staining at the lumen surface or around adventitial microvessels (Fig. 1). Colocalization of CRP and α-smooth muscle actin was observed only in the deep media of vein grafts. In the adventitial layer of vein graft specimens, CRP immunoreactivity appears most intense in areas of extracellular matrix deposition.

Morphological measurements of vein graft lesions were performed on elastin-stained sections (data not shown). The mean compartment areas of the neointima, media, and total wall were 5.1 ± 1.5, 4.0 ± 0.91, and 9.1 ± 2.3 mm², respectively. The mean neointimal thickness was 0.71 ± 0.16 mm, and mean total wall thickness was 1.1 ± 0.19 mm. No correlation was observed between intensity of CRP staining and any of these morphological parameters. We observed intimal hyperplasia in most preimplantation veins, as others have also reported (29). Mean compartment areas of the intima, media, and total wall of preimplantation veins were 0.33 ± 0.10, 1.41 ± 0.47, and 2.3 ± 0.05 mm², respectively. Mean intimal thickness was 0.08 ± 0.03 mm, and mean total wall thickness was 0.65 ± 0.03 mm.

Exogenous CRP does not affect HSV SMC viability or proliferation.

Migration and proliferation of SMCs and myofibroblasts are considered important components of vein graft hyperplasia. The doubling time of our SMCs is ∼48 h. Cells treated with vehicle or CRP for this period demonstrated no loss of viability by trypan blue exclusion: 96.3 ± 5.7% difference in absolute number of viable cells between vehicle-treated wells and wells treated with 5 μg/ml CRP and 91.8 ± 4.3% difference between vehicle-treated wells and wells treated with 50 μg/ml CRP. Furthermore, there was no gross evidence of apoptosis in any treatment group, as evidenced by DAPI-stained nuclei (Fig. 2). We used relative reduction of Alamar Blue substrate to compare the effect of exogenous CRP on SMC proliferation. We observed no difference in proliferation between SMCs cultured (in low serum for 8 days) in vehicle and SMCs cultured in CRP (Fig. 2).

Fig. 2.

Exogenous CRP has no effect on human saphenous vein (HSV) smooth muscle cell (SMC) viability or proliferation. A: HSV SMCs were grown in low-serum conditions (0.1%) in the presence of vehicle (control group), 5 μg/ml CRP, or 50 μg/ml CRP or in normal-serum conditions (10% FCS) for 8 days. Alamar Blue reduction was checked every 2–4 days. Proliferation is expressed as percent increase in reduction of substrate relative to vehicle-treated cells (horizontal line). Data represent results from 3 independent experiments in which each treatment was applied in triplicate. *Significant divergence in proliferation in 10% FCS groups at 4 and 8 days (P = 0.016 and P = 0.011, respectively, by ANOVA) but no difference between vehicle- and CRP-treated groups at any time point. B: HSV SMCs treated for 48 h with vehicle or CRP (5 and 50 μg/ml) demonstrated neither loss in cell viability by trypan blue exclusion nor evidence of nuclear crescents or condensation by nucleus-specific 4′,6-diamidino-2-phenylindole (DAPI) staining.

Exogenous CRP enhances SMC chemotaxis to PDGF-BB.

To examine the influence of exogenous CRP on the migratory phenotype of HSV SMCs, we used a modified Boyden chamber assay. As shown in Fig. 3, we observed a maximal threefold enhancement of chemotaxis to PDGF-BB (P < 0.05) in cells pretreated with 20 μg/ml CRP for 48 h vs. control; this enhancement was effectively blocked by a PDGF-neutralizing antibody.

Fig. 3.

Stimulation of platelet-derived growth factor (PDGF)-directed chemotaxis of HSV SMCs by CRP. A: representative DAPI-stained membranes from each treatment group at the end of the migration period. B: effect of CRP pretreatment on HSV SMC chemotaxis to PDGF-BB in the presence or absence of neutralizing antibody to PDGF. Results represent ≥5 independent experiments and are expressed as fold increase in cells migrated per high-power field relative to control (no CRP) group. *P < 0.01.

Modulation of PDGFRβ gene expression in HSV SMCs by exogenous CRP.

Activation of PDGFR tyrosine kinase function stimulates a series of cell signaling events leading to multiple cellular activities, including membrane ruffling and chemotaxis. Furthermore, given the established relevance of the autocrine/paracrine PDGF signaling pathway in vein graft biology and in the regulation of SMC phenotype, we next investigated the influence of CRP on gene expression of the relevant ligands and receptors. Primary cultures of HSV SMCs were treated with purified, dialyzed human CRP, and the expression of PDGF-A, PDGF-B, PDGF-D, PDGFRα, and PDGFRβ was measured by semiquantitative RT-PCR. As shown in Fig. 4A, there was a dose-dependent increase in PDGFRβ gene expression as determined by semiquantitative RT-PCR but no significant change in PDGFRα, PDGF-A, or PDGF-D gene expression. Under these culture conditions, PDGF-B expression was consistently minimal or nonexistent in these assays (not shown). These findings were then quantified by real-time PCR assays (Fig. 4B). Using the comparative C_T method, we calculated a 1.4- and a 1.6-fold relative increase in PDGFR-β gene expression after 24 h of treatment with 5 and 50 μg/ml CRP, respectively (P < 0.05).

Fig. 4.

A: representative semiquantitative RT-PCR analysis of PDGF receptor (PDGFR)-α, PDGFRβ, PDGF-A, and PDGF-D gene expression in HSV SMCs treated with 5 and 50 μg/ml CRP for 24 h. No reproducible evidence of PDGF-B gene expression was obtained (not shown). B: real-time PCR amplification of PDGFRβ in SMCs treated with CRP for 1 and 24 h. Data, which represent results from 3 independent experiments, are expressed relative to expression of an endogenous control (GAPDH). *P < 0.05 vs. no CRP treatment.

PDGFRβ protein level and phosphorylation state in HSV SMC are modulated by exposure to CRP.

Consistent with the findings of altered gene expression, exogenous CRP treatment also induced a dose-dependent increase in PDGFRβ protein levels, as determined by Western blotting of HSV SMC lysates. As shown in Fig. 5A, PDGFRβ protein level increased ∼1.3- and 2.5-fold (P < 0.05) after 48 h of treatment with 5 and 50 μg/ml CRP, respectively, an effect that was not seen after shorter durations (24 h) of exposure. Furthermore, cells that were exposed to the higher dose of CRP for 48 h demonstrated increased tyrosine phosphorylation of immunoprecipitated receptor (Fig. 5B), consistent with increased receptor activation. Also, immunohistochemistry showed the presence of PDGFRβ in the vessel wall of failing vein grafts, most prominently in the adventitia, as well as in the endothelial layer of preimplantation veins (Fig. 5C).

Fig. 5.

Representative Western blot analysis of PDGFRβ protein expression (A) and PDGFRβ tyrosine phosphorylation (B) after treatment with 5 and 50 μg/ml CRP for 24 and 48 h. Blots were scanned with a densitometer, and relative expression of total receptor (A) or relative phosphorylation of PDGFRβ (B) is shown as fold increase over control (no CRP) group. Densitometric analysis of bands in Western blots from 48-h time points represents results from 3 independent experiments. IP, immunoprecipitation; WB, Western blot; p-tyr, phosphorylated tyrosine. *P < 0.05 vs. no CRP. C: preimplantation veins and failing vein grafts shown in Fig. 1 were analyzed for expression of PDGFRβ by immunohistochemistry. Representative staining is shown. Top, preimplantation vein; bottom, failing vein graft; left, negative (IgG) control staining; right, anti-PDGFRβ. L, lumen. Original magnification ×40.

DISCUSSION

CRP is a classical acute-phase reactant protein that is nonspecifically elevated in diverse settings of stress, trauma, or disease that are characterized by inflammation. In apparently healthy individuals, increased levels of plasma CRP have been correlated with cardiovascular risk, with incremental predictive value over traditional Framingham risk factors. After implantation, vein grafts are the target of an acute inflammatory response that may be related to several inciting factors, including mechanical injury, ischemia-reperfusion, endothelial dysfunction, and acute hemodynamic stresses. We have hypothesized that the magnitude of this inflammatory response may be a critical factor in the variability in remodeling and clinical outcomes following peripheral vein bypass grafting. In recent prospective clinical studies, we demonstrated that the baseline level of hsCRP was predictive of clinical outcome (40) and lumen remodeling changes in patients undergoing lower extremity bypass (41). In the present study, we have begun to address whether CRP may be of direct biological relevance to vein graft disease. The data demonstrate that CRP frequently immunolocalizes to the deep media and adventitia of failing HSV lower extremity bypass grafts but is not usually identified in normal (preimplantation) veins. We have also shown that CRP directly stimulates HSV SMC migration and increases the expression and activation of PDGFRβ in a dose-dependent fashion. Taken together, these results suggest a potential mechanistic link between inflammation and the early injury response in venous SMCs.

Our immunohistochemistry studies demonstrated preferential staining of diseased vein grafts explanted at the time of graft revision over preimplantation saphenous veins. A similar pattern was reported in a canine model of arterialized vein grafts (19) but is in contrast to previous reports of CRP staining in the wall of normal vascular tissue, including the pulmonary, reproductive, and cardiac systems (13). Furthermore, prior studies reported CRP staining in the neointima of diseased coronary and lower extremity bypass grafts (24), and not a predominance of staining in the media and adventitia, as we and others observed (19).

Although the production of CRP within inflamed or diseased vessel walls is controversial (19, 24, 64), the hypothesis that CRP, by virtue of its binding sites for phosphocholine, binds preferentially to damaged cell membranes and apoptotic cells over healthy membranes is relevant to vascular injury. Since the CRP staining that we observed did not exclusively colocalize with cell type-specific markers (i.e., α-smooth muscle actin or CD31) or follow discrete boundaries in the vessel wall (endothelium vs. neointima vs. media vs. adventitia), we speculate that CRP is deposited in the graft wall, rather than produced in situ or bound to cell-specific receptors; this may explain the differences in the localization of CRP immunoreactivity across studies (24). Interestingly, a small-molecule inhibitor of CRP that occludes the phosphocholine-binding face of CRP effectively reduces myocardial infarct size and dysfunction after coronary ligation in a rat model (43). An alternative explanation for the discrepancy in staining patterns between diseased and preimplantation tissue is that the subclinical inflammatory state (as measured by plasma CRP) of patients is highly variable. We recently reported in a prospective, multicenter study that the baseline plasma hsCRP levels in patients undergoing lower extremity bypass surgery are quite heterogenous, although they can be dichotomized into groups that correlate with discrete clinical outcomes (40, 41); the comparatively small number of clinical specimens in our study may not have allowed us to discern this pattern.

We focused our in vitro studies on the effects of exogenous CRP on SMCs, since we found the deposition of CRP in failing human lower extremity bypass grafts to be concentrated in the deep media and adventitia. Midterm lower extremity bypass graft failure is generally due to neointimal hyperplasia. Neointimal lesions, composed of extracellular matrix and SMCs, are thought to occur through dedifferentiation of medial and intimal SMCs to a synthetic and reparative phenotype, which migrate toward the lumen and proliferate in response to increased shear stress, wall tension, and endothelial denudation at the time of surgery (8, 50). Previous reports demonstrated the activation of SMC proliferation (7, 59) and apoptosis (5) and increased matrix metalloproteinase-2 gene and protein expression and activity in arterial SMCs (14, 15) in response to exogenous CRP activity. However, because of histological and physiological differences between artery and vein (37, 56), these results may not be generalizable to venous SMCs. Therefore, our experiments were conducted using primary cultures of HSV SMCs, and not coronary artery or animal cell lines. In addition, we used only purified human CRP (not the recombinant form) that had been dialyzed to remove sodium azide and was tested for endotoxin (26, 33, 39, 42, 57). Another important distinction of the present report is that we applied CRP within the range found at baseline in human lower extremity bypass graft patients (41), thereby increasing the clinical relevance of the findings.

Both PDGF, a potent SMC mitogen released at the site of vascular injury by platelets, endothelial cells, and SMCs (20, 32, 53), and its receptors are upregulated in the vessel wall in atherosclerotic plaques and in animal models of vascular injury (25, 35, 53, 63). PDGF-BB is the most potent isoform in terms of mitogenicity, and stimulus for migration and PDGFRβ appears to mediate these responses. In a baboon model of arterial injury, administration of a neutralizing antibody to PDGFRβ reduced neointimal size after 1 mo (17). Furthermore, antisense oligonucleotide inhibition of PDGFRβ expression in injured rat carotid arteries decreased neointima formation by 80% (52). We observed overall higher expression of PDGFRβ by immunohistochemistry in our failing vein graft specimens than in the preimplantation veins, regardless of preexising intimal hyperplasia, consistent with prior reports of PDGFRβ immunohistochemistry in injured vessels (35, 53). Although we established a dose-dependent increase in PDGFRβ expression and phosphorylation and PDGF-BB-mediated chemotaxis with CRP treatment of HSV SMCs, neither CRP blocking antibodies (16, 28) nor the specific small-molecule inhibitor of CRP (43) was available to us to confirm the specificity of our interaction.

Our studies are also limited by the lack of identification of a definite CRP receptor. Ongoing, but conflicting, work in this area has suggested the Fcγ receptors FcγRI and FcγRIIα in various cell types (2, 4, 6, 12, 36, 38, 47–49); a recent report describes colocalization of FcγRIIα with α-actin-positive SMCs in atherosclerotic human arteries, and this receptor may mediate the proinflammatory effects of CRP in arterial SMCs (48). Whether this receptor also mediates the in vitro effects of CRP in venous SMCs is not known.

Furthermore, we cannot preclude definitively that our preparations of CRP are devoid of PDGF ligands, which could activate PDGFRβ. However, we did not detect bands corresponding to the size of PDGF-A, -B, or -D in Coomassie-stained gels of CRP separated by SDS-PAGE.

In contrast to other reports, we did not observe stimulation of HSV SMC proliferation by exogenous CRP (59). However, different assays for proliferation were used in those studies ([³H]thymidine incorporation), as was a recombinant CRP preparation. The downstream signals following ligand binding, dimerization, and autophosphorylation of PDGFRβ control proliferation, matrix deposition, immediate early gene transcription, and migration (for review see Ref. 3, 23). Autophosphorylation occurs on multiple tyrosine residues that are binding sites for ≥10 distinct Src homology 2 (SH2) domain-containing proteins (22). Since chemotaxis, but not proliferation, is controlled by an interplay of SH2 domain-containing proteins phospholipase Cγ, phosphatidylinositol-3-OH kinase, and Ras-GTPase-activating protein binding to specific tyrosine residues (30, 61, 62), we likely have observed only the net effect of CRP exposure on HSV SMCs, and not accompanying effects on the activity of multiple regulatory proteins that mediate the biological effects of PDGFRβ signaling.

In conclusion, we have detected CRP in failing human lower extremity bypass grafts as opposed to preimplantation veins by immunohistochemistry and demonstrated that treatment of HSV SMCs with purified, dialyzed human CRP in vitro leads to a phenotypic change that may be mediated by increased expression and phosphorylation of PDGFRβ, and not PDGF ligands. Future investigations using specific CRP inhibitors, inhibitors of putative receptors for CRP, and PDGFRβ promoter assays may further elucidate this relationship. Furthermore, additional mechanistic details on the effects of CRP downstream of PDGFRβ, such as Akt and Rho-ROCK (1, 51), need to be addressed.

GRANTS

This work was supported by grants from the National Heart, Lung, and Blood Institute (to K. J. Ho), the American Vascular Association/Lifeline Foundation (to T. Longo and C. Ifantides), and the Harvard-MIT Clinical Investigator Training Program (to C. D. Owens).