Magnetic Resonance Imaging Correlates of White Matter Gliosis and Injury in Preterm Fetal Sheep Exposed to Progressive Systemic Inflammation

Abstract

Progressive fetal infection/inflammation is strongly associated with neural injury after preterm birth. We aimed to test the hypotheses that progressively developing fetal inflammation leads to neuroinflammation and impaired white matter development and that the histopathological changes can be detected using high-field diffusion tensor magnetic resonance imaging (MRI). Chronically instrumented preterm fetal sheep at 0.7 of gestation were randomly assigned to receive intravenous saline (control; n = 6) or a progressive infusion of lipopolysaccharide (LPS, 200 ng intravenous over 24 h then doubled every 24 h for 5 days to induce fetal inflammation, n = 7). Sheep were killed 10 days after starting the infusions, for histology and high-field diffusion tensor MRI. Progressive LPS infusion was associated with increased circulating interleukin (IL)-6 concentrations and moderate increases in carotid artery perfusion and the frequency of electroencephalogram (EEG) activity (p < 0.05 vs. control). In the periventricular white matter, fractional anisotropy (FA) was increased, and orientation dispersion index (ODI) was reduced (p < 0.05 vs. control for both). Histologically, in the same brain region, LPS infusion increased microglial activation and astrocyte numbers and reduced the total number of oligodendrocytes with no change in myelination or numbers of immature/mature oligodendrocytes. Numbers of astrocytes in the periventricular white matter were correlated with increased FA and reduced ODI signal intensities. Astrocyte coherence was associated with increased FA. Moderate astrogliosis, but not loss of total oligodendrocytes, after progressive fetal inflammation can be detected with high-field diffusion tensor MRI.

Keywords: MRI; brain; infection; inflammation; preterm infant.

Figure 1

Time-course of changes in circulating interleukins. Time-course of changes in circulating interleukin (IL)-6 and IL-8 in the control (open circles, n = 6) and lipopolysaccharide (LPS) (closed circles, n = 7) groups. The dark shaded area represents the period of increasing LPS infusion. Data are hourly means ± SEM. * p < 0.05, LPS vs. control.

Figure 2

White matter immunohistochemistry. Ionized calcium-binding adaptor molecule 1 (Iba-1), oligodendrocyte transcriptase factor 2 (Olig-2), amoeboid (Iba-1 positive) microglia, and 2′,3′-cyclic-nucleotide 3′-phosphodiesterase (CNPase)-positive cell counts, the proportion of mature CNPase oligodendrocytes (CNPase/Olig-2 positive cells) and the percentage area of myelin basic protein (MBP)-positive staining in the intragyral and periventricular white matter tracts (IGWM and PVWM, respectively) and corpus callosum (CC) in control (open bars, n = 6) and LPS groups (closed bars, n = 7). Data are means ± SEM. LPS, lipopolysaccharide, * p < 0.05, LPS vs. control.

Figure 3

Photomicrographs of white matter tracts. Representative photomicrographs showing Iba-1-, glial fibrillary acidic protein (GFAP)-, Olig-2-, CNPase-, and MBP-positive staining in the periventricular white matter tracts. LPS, lipopolysaccharide. Scale bar is 200 µm. Arrows in the Iba-1 photomicrographs indicate microglia displaying a resting ramified phenotype, characterized by a small cell body with >1 branching process. Arrowheads in the Iba-1 photomicrographs from the LPS group indicate microglia displaying an amoeboid morphology, characterized by a large cell body with ≤1 branching process. Arrows in the GFAP photomicrographs indicate astrocytes with an isotropic appearance. Arrowheads in the GFAP photomicrographs from the LPS group indicate astrocytes displaying greater coherence and more anisotropy compared to controls.

Figure 4

Diffusion tensor magnetic resonance imaging (MRI). (Left) Fractional anisotropy (FA) and orientation dispersion index (ODI) in the intragyral and periventricular white matter (IGWM and PVWM, respectively) and corpus callosum (CC) in control (open bars, n = 6) and LPS groups (closed bars, n = 7). (Right) Correlative analysis between FA, ODI, and periventricular GFAP-positive astrocytes in control (open circles, n = 6) and LPS groups (closed circles, n = 7). * p < 0.05, LPS vs. control.

Figure 5

Astrocyte coherence. (Left) Astrocyte coherence in the periventricular white matter quantified from GFAP-stained histological sections. Global coherence was derived from whole periventricular white matter tissue sections. Cellular coherence represents coherence values derived from 15 randomly selected astrocytes per section. (Right) Correlative analyses between global and cellular coherence in the periventricular white matter from GFAP-stained tissue sections and fractional anisotropy. * p < 0.05, LPS vs. control.

Figure 6

Schematic outlining the study design. The study consisted of two groups: control (n = 6), LPS (n = 7). The dark shaded area represents the period of progressive LPS infusion (200 ng LPS for the first 24 h, then doubled every 24 h for the next 96 h). Controls received an equivalent volume of vehicle (saline) during the infusion period. Continuous physiological recordings were performed throughout the experimental period. Fetal preductal arterial blood was collected every morning starting from 30 min before increasing the LPS or saline infusion and at 6 h after increasing the infusion for measurement of cytokine levels. Representative direction-encoded color map outlining the regions of interest sampled for MRI and histological analyses. Brain regions used for analysis were collected from the forebrain at the level of the mid-striatum from sections taken approximately 23 mm anterior to stereotaxic zero. White squares indicate regions of interest (ROIs) for assessment of the corpus callosum (1, 2), intragyral white matter within the first (3, 4) and the periventricular white matter (5, 6). The diffusion-weighted images and ROIs chosen for MRI analysis corresponded with the same regions used for histological assessment.

Figure 7

(Top) Representative examples of coherence analysis of GFAP-stained sections. Coherence analysis was performed using the OrientationJ plug-in for Fiji. Areas for coherence analysis were selected with the rectangular selection tool to measure coherence within the whole periventricular white matter section (global coherence, No. 1). The round selection tool was used to measure coherence for 15 individual astrocytes that were randomly selected from each periventricular white matter image (cellular coherence, no. 2–16). (Bottom) Mean diffusivity (MD), axial diffusivity (AD), radial diffusivity (RD), fractional anisotropy (FA); direction-encoded color (DEC) maps; intra-neurite volume fraction (fin); isotropic volume fraction (fiso); and orientation dispersion index (ODI). LPS, lipopolysaccharide.