Plasma chemokine levels are associated with the presence and extent of angiographic coronary collaterals in chronic ischemic heart disease

Abstract

Background: In patients with chronic ischemic heart disease (IHD), the presence and extent of spontaneously visible coronary collaterals are powerful determinants of clinical outcome. There is marked heterogeneity in the recruitment of coronary collaterals amongst patients with similar degrees of coronary artery stenoses, but the biological basis of this heterogeneity is not known. Chemokines are potent mediators of vascular remodeling in diverse biological settings. Their role in coronary collateralization has not been investigated. We sought to determine whether plasma levels of angiogenic and angiostatic chemokines are associated with of the presence and extent of coronary collaterals in patients with chronic IHD.

Methodology/principal findings: We measured plasma concentrations of angiogenic and angiostatic chemokine ligands in 156 consecutive subjects undergoing coronary angiography with at least one ≥90% coronary stenosis and determined the presence and extent of spontaneously visible coronary collaterals using the Rentrop scoring system. Eighty-eight subjects (56%) had evidence of coronary collaterals. In a multivariable regression model, the concentration of the angiogenic ligands CXCL5, CXCL8 and CXCL12, hyperlipidemia, and an occluded artery were associated with the presence of collaterals; conversely, the concentration of the angiostatic ligand CXCL11, interferon-γ, hypertension and diabetes were associated with the absence of collaterals (ROC area 0.91). When analyzed according to extent of collateralization, higher Rentrop scores were significantly associated with increased concentration of the angiogenic ligand CXCL1 (p<0.0001), and decreased concentrations of angiostatic ligands CXCL9 (p<0.0001), CXCL10 (p = 0.002), and CXCL11 (p = 0.0002), and interferon-γ (p = 0.0004).

Conclusions/significance: Plasma chemokine concentrations are associated with the presence and extent of spontaneously visible coronary artery collaterals and may be mechanistically involved in their recruitment.

Figure 1. Box plots (median and interquartile range) of percentile rank of plasma concentrations of cytokines that promote vascular remodeling: CXCL1 (panel A), CXCL3 (panel B), CXCL5 (panel C), CXCL8 (panel D), CXCL12 (panel E), CCL2 (panel F), VEGF (panel G), and bFGF (panel H). CXCL1 (*p<0.0001), CXCL12 (*p = 0.0022), CCL2 (*p = 0.0002), and VEGF (*p = 0.008) significantly correlated across increasing Rentrop scores.

Figure 2. Box plots (median and interquartile range) of percentile rank of plasma concentrations of cytokines that inhibit vascular remodeling: CXCL9 (panel A), CXCL10 (panel B), CXCL11 (panel C), and IFN-γ (panel D).

All cytokines significantly correlated across increasing Rentrop scores: CXCL9 (*p<0.0001), CXCL10 (*p = 0.002), CXCL11 (*p = 0.0002), and IFN-γ (*p = 0.0004).