Prolonged Blood-Brain Barrier Injury Occurs After Experimental Intracerebral Hemorrhage and Is Not Acutely Associated with Additional Bleeding

Abstract

Intracerebral hemorrhage (ICH) causes blood-brain barrier (BBB) damage along with altered element levels in the brain. BBB permeability was quantified at 3, 7, and 14 days with Evans Blue dye after collagenase-induced ICH in rat. At peak permeability (day 3), a gadolinium (Gd)-based contrast agent was injected to further characterize BBB disruption, and X-ray fluorescence imaging (XFI) was used to map Gd, Fe, Cl, and other elements. XFI revealed that Ca, Cl, Gd, and Fe concentrations were significantly elevated, whereas K was significantly decreased. Therefore, using Gd-XFI, we co-determined BBB dysfunction with alterations in the metallome, including those that contribute to cell death and functional outcome. Warfarin was administered 3 days post-ICH to investigate whether additional or new bleeding occurs during peak BBB dysfunction, and hematoma volume was assessed on day 4. Warfarin administration prolonged bleeding time after a peripheral cut-induced bleed, but warfarin did not worsen hematoma volume. Accordingly, extensive BBB leakage occurred after ICH, but did not appear to affect total hematoma size.

Keywords: Blood-brain barrier; Gadolinium extravasation; Intracerebral hemorrhage; Ion dyshomeostasis; X-ray fluorescence imaging.

Fig. 1

Evans Blue extravasation at 3D was significantly higher than at 7D and 14D in IPSI tissue. Ipsilateral EB extravasation at 3D was significantly higher than SHAM. Significant elevations persisted to 7D. There were no differences in EB extravasation in the CONTRA hemisphere. There was a significant relationship between time and proportion of samples with BBB dysfunction in the IPSI, but not CONTRA, hemisphere (mean ± SD; *p < 0.05; **p < 0.01; ***p < 0.001)

Fig. 2

ICP-MS measurement of element concentrations after ICH. Concentrations of Ca (a) within the HEM did not differ from CONTRA tissue. Fe (b) and Gd (c) concentrations were significantly increased in the HEM as compared to CONTRA striatum. Concentrations of K (d) and Na (e) within the HEM did not differ from CONTRA tissue (mean ± SD; *p < 0.05; **p < 0.01; ***p < 0.001)

Fig. 3

Representative x-ray fluorescent images of Ca (a) Cl (b), Fe (c), Gd (d), and K (e). The solid black line marks the HEM boundary, which was determined in the Fe channel by locating a sharp decrease in Fe levels. The HEM boundary was propagated to all other channels. Cresyl violet staining (f) confirms the HEM boundaries determined with XFI. A dotted black line illustrates a sample distance ROI (f). Three separate arrays of data collection were used to quantify PHZ element levels. The tissue fold at the bottom is an artifact from sectioning

Fig. 4

Percent of sampled tissue with greater than 25% change (arbitrary threshold likely to denote biological significance) in ion concentration as compared to within-subjects CONTRA striatum. Over 50% of IPSI tissue sampled outside of the HEM showed Ca, Cl, Fe, and K dyshomeostasis, while ~ 40% of tissue showed Gd dyshomeostasis. Figure 3 shows representative x-ray fluorescent images

Fig. 5

XFI measurement of element concentrations after ICH. Ca (a) and Cl (b) concentrations were significantly increased in the HEM center compared to CONTRA striatum. Fe (c) concentrations were significantly increased in the HEM edge compared to CONTRA striatum. Gd (d) concentrations were significantly increased in the HEM center compared to CONTRA striatum. K (e) concentrations were significantly decreased in the HEM center and edge compared to CONTRA striatum (mean ± SD; *p < 0.05; **p < 0.01; ***p < 0.001)

Fig. 6

There was a distance main effect for Ca, Cl, Fe, and Gd, such that values were highest closest to the HEM and decreased with distance into the PHZ at 3D (a–d). There was a distance main effect for K, where values were lowest in the HEM and normalized with distance into the PHZ (e). See Fig. 3f for an illustration of sampling method (mean ± SD; *p < 0.05 as compared to CONTRA. ^‡p < 0.05 as compared to SHAM)

Fig. 7

Warfarin did not significantly increase the cerebral HEM volume (a) although it did significantly increase the time it took peripheral blood to clot (b) (mean ± SD; *p < 0.05; **p < 0.01; ***p < 0.001)