The extent of intrauterine growth restriction determines the severity of cerebral injury and neurobehavioural deficits in rodents

Abstract

Background: Cerebral Palsy (CP) is the most common physical pediatric neurodevelopmental disorder and spastic diplegic injury is its most frequent subtype. CP results in substantial neuromotor and cognitive impairments that have significant socioeconomic impact. Despite this, its underlying pathophysiological mechanisms and etiology remain incompletely understood. Furthermore, there is a need for clinically relevant injury models, which a) reflect the heterogeneity of the condition and b) can be used to evaluate new translational therapies. To address these key knowledge gaps, we characterized a chronic placental insufficiency (PI) model, using bilateral uterine artery ligation (BUAL) of dams. This injury model results in intrauterine growth restriction (IUGR) in pups, and animals recapitulate the human phenotype both in terms of neurobehavioural and anatomical deficits.

Methods: Effects of BUAL were studied using luxol fast blue (LFB)/hematoxylin & eosin (H&E) staining, immunohistochemistry, quantitative Magnetic Resonance Imaging (MRI), and Catwalk neurobehavioural tests.

Results: Neuroanatomical analysis revealed regional ventricular enlargement and corpus callosum thinning in IUGR animals, which was correlated with the extent of growth restriction. Olig2 staining revealed reductions in oligodendrocyte density in white and grey matter structures, including the corpus callosum, optic chiasm, and nucleus accumbens. The caudate nucleus, along with other brain structures such as the optic chiasm, internal capsule, septofimbrial and lateral septal nuclei, exhibited reduced size in animals with IUGR. The size of the pretectal nucleus was reduced only in moderately injured animals. MAG/NF200 staining demonstrated reduced myelination and axonal counts in the corpus callosum of IUGR animals. NeuN staining revealed changes in neuronal density in the hippocampus and in the thickness of hippocampal CA2 and CA3 regions. Diffusion weighted imaging (DWI) revealed regional white and grey matter changes at 3 weeks of age. Furthermore, neurobehavioural testing demonstrated neuromotor impairments in animals with IUGR in paw intensities, swing speed, relative print positions, and phase dispersions.

Conclusions: We have characterized a rodent model of IUGR and have demonstrated that the neuroanatomical and neurobehavioural deficits mirror the severity of the IUGR injury. This model has the potential to be applied to examine the pathobiology of and potential therapeutic strategies for IUGR-related brain injury. Thus, this work has potential translational relevance for the study of CP.

Fig 1. Whole brain, third ventricle and corpus callosum structural abnormalities in the brains of IUGR animals.

Animals were initially separated into different severities of IUGR based on birth weight; moderate IUGR animals had significantly lower birthweights than sham controls and mild IUGR animals, while mild IUGR only had significantly lower birthweight than sham controls (A). Although weight eventually normalized in adulthood (B), animals classified to have moderate IUGR at birth displayed persistent decreases in total brain area in adulthood, though this was not the case for mild IUGR animals (C). Representative whole-brain sections are depicted in (D-F) for sham, mild IUGR, and moderate IUGR, respectively. The CC is delineated in (G-I) for sham, mild IUGR, and moderate IUGR, respectively. Both groups of IUGR animals also displayed significant ventricular enlargement at the level of the diencephalon when compared to sham controls (J). The area of the CC was reduced in IUGR animals when compared to sham controls, and the CC of moderate IUGR animals was significantly smaller than mild IUGR animals (K). The thickness of the CC was also significantly reduced in both mild and moderate IUGR groups when compared to sham animals (L). Scale Bar = 1000 μm; *p<0.05; **p<0.01; ***p<0.001.

Fig 2. MRI reveals differences between moderate IUGR and sham animals.

At 3 weeks, abnormalities in both axial and radial diffusivities in white and grey matter were evident in moderate IUGR animals. (A) Representative T2-weighted images and fractional anisotropy (FA) maps, highlighting locations of ROIs in the corpus callosum (CC), external capsule (EC), and caudate (CA) at three anatomical levels (g: genu; b: body; and s: splenium). (B) Representative FA, axial diffusivity, and radial diffusivity maps at the level of the genu from sham and moderate IUGR rats. Note the higher radial diffusivities in the CC of the moderate IUGR animal when compared to sham, with similar FA and axial diffusivities. (C) Tabulated DTI metrics across the sham and moderate IUGR cohorts. Locations of significant differences are highlighted by * (p<0.05) and ** (p<0.01). (D) Graphical display of radial diffusivities in each ROI location for the moderate injury cohort (*p<0.05; **p<0.01).

Fig 3. IUGR is accompanied by a loss of white matter in the corpus callosum.

Olig2+ OL counts in the corpus callosum revealed a decrease in OLs in mild IUGR (B), and moderate IUGR (C) animals, but not in sham controls (A). This decrease was significantly different in both IUGR groups when compared to sham controls; moderate IUGR animals also displayed significantly reduced counts when compared to mild IUGR animals (D). Both mild and moderate IUGR animals had decreased myelination (as assessed through MAG+ staining) when compared to sham controls (F-H). The number of NF200+ axons was also reduced in both mild and moderate IUGR groups when compared to sham controls (I-K). Merged images of MAG+/NF200+ staining are also shown (L-N). There were fewer axons in IUGR animals regardless of the severity of injury (O). A significantly greater proportion of the remaining axons were unmyelinated (MAG-/NF200+) in only moderate IUGR animals, while mild IUGR animals exhibited a phenotype similar to sham controls (P). When we investigated the results in (P) as a percentage of total axons in the CC, only moderately injured animals had a significantly higher proportion of unmyelinated axons when compared to other groups (Q). Scale Bar = 100 μm;*p<0.05; **p<0.01; ***p<0.001.

Fig 4. Hippocampal structure is differentially affected according to the severity of IUGR.

NeuN+ staining in the hippocampus (A-C) revealed severity-specific deficits in CA2 (D-E), CA3 (F-G) and the dentate gyrus (H). Quantification of hippocampal area revealed a significantly reduced size only in moderate IUGR animals when compared to sham animals (I). Scale Bar = 1000 μm;*p<0.05; **p<0.01; ***p<0.001.

Fig 5. Reduced area and number of tracts in the caudate nucleus.

The caudate nucleus was affected by IUGR, as depicted in representative LFB/H&E stains (A-C). This structure showed reduced size in IUGR animals, regardless of severity (D). IUGR animals displayed decreased numbers of axonal tracts within the caudate nucleus (E), though mean axonal size was unaffected (F). Scale Bar = 1000 μm;*p<0.05; **p<0.01; ***p<0.001.

Fig 6. Structural abnormalities in sub-cortical structures of IUGR animals.

Animals with IUGR displayed differences in sub-cortical structures when compared to Sham animals. Internal capsule size was reduced in both mild and moderate IUGR groups (A). The number of Olig2+ cells in the internal capsule remained unchanged (B). The optic chiasm showed similar results as the internal capsule (C), in addition to fewer OLs in the mildly injured group (D). There was no significant difference in the size of the nucleus accumbens between IUGR and sham animals (E). There were, however, fewer Olig2+ OLs in the nucleus accumbens of both mild and moderate IUGR animals when compared to sham animals (F). The septofimbrial and lateral septal nuclei were decreased in size in IUGR animals relative to sham controls (G). The pretectal nucleus was also significantly decreased in size in moderate IUGR animals only when compared to sham animals (H). *p<0.05; **p<0.01; ***p<0.001.

Fig 7. Motor gait deficits are present in IUGR rats.

Animals with IUGR showed clear and consistent gait deficits at 3 weeks, some of which persisted into young adulthood. Paw intensity (0–255) for fore- and hindpaws was significantly increased in moderate IUGR animals (p<0.0001 for both) (A). Swing speed (m/s) for fore- and hindpaws did not show overall significance between IUGR and sham animals. There were, however, significant differences at early timepoints between moderate IUGR and sham animals (B). Relative print positions (mm) for right and left pairs of paws showed overall significant differences in the left paws of both mild and moderate IUGR animals when compared to sham animals (p = 0.0089 and p = 0.0019, respectively). There was a similar trend toward significance in the right paws of both mild and moderate IUGR groups (p = 0.0503 and p = 0.0603, respectively) (C). Phase dispersions (AU) of both fore and hind girdle paw pairs were significantly decreased in moderate IUGR animals when compared to sham animals (p = 0.0114 and p = 0.0002, respectively) (D). *p<0.05, **p<0.01, ***p<0.001.