MMP-9-Dependent Serum-Borne Bioactivity Caused by Multiwalled Carbon Nanotube Exposure Induces Vascular Dysfunction via the CD36 Scavenger Receptor

Abstract

Inhalation of multiwalled carbon nanotubes (MWCNT) causes systemic effects including vascular inflammation, endothelial dysfunction, and acute phase protein expression. MWCNTs translocate only minimally beyond the lungs, thus cardiovascular effects thereof may be caused by generation of secondary biomolecular factors from MWCNT-pulmonary interactions that spill over into the systemic circulation. Therefore, we hypothesized that induced matrix metalloproteinase-9 (MMP-9) is a generator of factors that, in turn, drive vascular effects through ligand-receptor interactions with the multiligand pattern recognition receptor, CD36. To test this, wildtype (WT; C57BL/6) and MMP-9(-/-)mice were exposed to varying doses (10 or 40 µg) of MWCNTs via oropharyngeal aspiration and serum was collected at 4 and 24 h postexposure. Endothelial cells treated with serum from MWCNT-exposed WT mice exhibited significantly reduced nitric oxide (NO) generation, as measured by electron paramagnetic resonance, an effect that was independent of NO scavenging. Serum from MWCNT-exposed WT mice inhibited acetylcholine (ACh)-mediated relaxation of aortic rings at both time points. Absence of CD36 on the aortic rings (obtained from CD36-deficient mice) abolished the serum-induced impairment of vasorelaxation. MWCNT exposure induced MMP-9 protein levels in both bronchoalveolar lavage and whole lung lysates. Serum from MMP-9(-/-)mice exposed to MWCNT did not diminish the magnitude of vasorelaxation in naïve WT aortic rings, although a modest right shift of the ACh dose-response curve was observed in both MWCNT dose groups relative to controls. In conclusion, pulmonary exposure to MWCNT leads to elevated MMP-9 levels and MMP-9-dependent generation of circulating bioactive factors that promote endothelial dysfunction and decreased NO bioavailability via interaction with vascular CD36.

Keywords: CD36; MMP-9.; carbon nanoparticle; cardiovascular; serum; toxicity; vascular.

FIG. 1.

A and B, MWCNT-7 imaging by electron microscopy demonstrates the relative size and adequacy of dispersion. C and D, Markers of injury in the lung lavage, LDH and albumin levels, were increased following exposure (*P < .05 vs DM by ANOVA, N = 4–8 per group). MWCNT: multiwalled carbon nanotubes.

FIG. 2.

Effects of serum from MWCNT-7-exposed mice on NO generation and bioavailabilty. A, Representative EPR spectra from endothelial cells incubated with control serum and serum from MWCNT-7-exposed mice. B, Unstimulated endothelial cells exhibited minimal baseline levels of detectable NO, which was not different when incubated with serum from DM (control) or MWCNT-7-treated mice. However, when stimulated by ATP, endothelial cells incubated with control serum demonstrated a significantly greater capacity to generate NO than cells treated with serum from MWCNT-7-exposed mice (*P < .05, N = 3 per group). C, In an acellular assay, levels of NO in iron-free media containing a known concentration of the NO donor, spermine NONOate, were not different in the presence of serum from control or MWCNT-7-treated mice. NO: nitric oxide; EPR: electron paramagnetic resonance.

FIG. 3.

A, Mouse serum obtained 4 h following MWCNT-7 treatment inhibited ACh-mediated vasorelaxation in aortic rings from untreated (naïve) mice (left) and maximum vessel constriction induced by serum (normalized to a KPSS response; right). Interestingly, the serum from low dose-treated mice was more potent than serum from high dose-treated mice. Asterisks indicates significant difference from control by 2-way ANOVA with Tukey’s multiple comparison test (*P < .05, **P < .001), dagger represents significant difference between the 10 and 40 μg doses (^†P < .01; N = 10 per group). B, Mouse serum obtained 24 h following MWCNT-7 treatment inhibited ACh-mediated vasorelaxation in aortic rings from untreated (naïve) mice (left), however the increased potency of the low dose exposure at 4 h was no longer observed. Maximum vessel constriction induced by serum (normalized to a KPSS response) is also shown (right of each relaxation curve). *Indicates significant difference from DM control by 2-way ANOVA (*P < .05, **P < .01; N = 6 − 8 per group). Ach: acetylcholine; KPSS: potassium containing physiological salt solution.

FIG. 4.

A, Mouse serum obtained 4 h following MWCNT-7 treatment was filtered to remove all larger proteins, leaving only biomolecules <10 kDa. This filtered serum still inhibited ACh-mediated vasorelaxation in aortic rings from untreated (naïve) mice. Asterisks indicate significant difference from control by 2-way ANOVA with Tukey’s multiple comparison test (*P < 0.05, **P < .01; N = 5 per group). B, Filtered mouse serum obtained 24 h following MWCNT-7 treatment did not affect ACh-mediated vasorelaxation in aortic rings from untreated mice (N = 5 per group). C, Mass spectroscopic analysis of numerous biomolecules ranging from 0.5 to 10 kDa remaining after filtration confirms the removal of large molecules in this bioactive fraction. Individual peaks are collected into bins of the histogram and separated by dose group for the 4 h posttreatment serum.

FIG. 5.

A, Mouse serum obtained 4 h following MWCNT-7 treatment had no effect on ACh-mediated vasorelaxation in aortic rings from untreated CD36-null mice (N = 5 per group). Significant differences in the contractile response to the serum addition were noted for serum from MWCNT-7-treated mice. Asterisks indicate significant difference from control by ANOVA with Dunnett’s multiple comparison test (*P <.05; n = 5 per group). B, Filtered mouse serum obtained 24 h following MWCNT-7 treatment had no effect on ACh-mediated vasorelaxation in aortic rings from untreated CD36-null mice (N = 4–5 per group). Significant differences in the contractile response to the serum addition were noted for serum from the 40 µg MWCNT-7-treated mice. Asterisks indicate significant difference from control by ANOVA with Dunnett’s multiple comparison test (*P < .05; n = 5 per group).

FIG. 6.

MMP-9 protein levels in bronchoalveolar lavage (A and B) and whole lung lysates (C and D) from MWCNT-7-exposed mice. Asterisks indicate significant difference from control by ANOVA with Dunnett’s multiple comparison test (*P < .05; **P < .01).

FIG. 7.

A, MMP-9^−/− mice exhibit similar degree of lung injury at 4 h post MWCNT-7 aspiration, as measured by LDH from the BALF. Asterisks indicate significant difference from control by 2-way ANOVA with Tukey’s multiple comparison test (*P < .01; **P < .001). B, Serum from MMP-9^−/− mice treated with 10 or 40 μg MWCNT-7 caused a modest right-shift effect on the concentration response to ACh-mediated vasorelaxation in WT vessels, compared with serum from the DM control mice, but did not reduce the overall magnitude of relaxation. Serum from the various groups exhibited a consistent contractile effect on all aortas. Asterisks indicate significant difference from control by 2-way ANOVA with Tukey’s multiple comparison test (*P < .05; n = 5–6 per group).WT: wild type; MMP-9: matrix metalloproteinase-9; DM. dosing media.