A modified controlled cortical impact technique to model mild traumatic brain injury mechanics in mice

Abstract

For the past 25 years, controlled cortical impact (CCI) has been a useful tool in traumatic brain injury (TBI) research, creating injury patterns that includes primary contusion, neuronal loss, and traumatic axonal damage. However, when CCI was first developed, very little was known on the underlying biomechanics of mild TBI. This paper uses information generated from recent computational models of mild TBI in humans to alter CCI and better reflect the biomechanical conditions of mild TBI. Using a finite element model of CCI in the mouse, we adjusted three primary features of CCI: the speed of the impact to achieve strain rates within the range associated with mild TBI, the shape, and material of the impounder to minimize strain concentrations in the brain, and the impact depth to control the peak deformation that occurred in the cortex and hippocampus. For these modified cortical impact conditions, we observed peak strains and strain rates throughout the brain were significantly reduced and consistent with estimated strains and strain rates observed in human mild TBI. We saw breakdown of the blood-brain barrier but no primary hemorrhage. Moreover, neuronal degeneration, axonal injury, and both astrocytic and microglia reactivity were observed up to 8 days after injury. Significant deficits in rotarod performance appeared early after injury, but we observed no impairment in spatial object recognition or contextual fear conditioning response 5 and 8 days after injury, respectively. Together, these data show that simulating the biomechanical conditions of mild TBI with a modified cortical impact technique produces regions of cellular reactivity and neuronal loss that coincide with only a transient behavioral impairment.

Keywords: biomechanics; controlled cortical impact; glia reactivity; mild traumatic brain injury; strain rate.

Figure 1

Comparison of estimated strains and strain rates occurring in the brain from human-based and animal-based studies of brain biomechanics. (A) The range of controlled cortical impact (CCI) brain injury extends over a much broader range of strain rates compared to those found in human and other animal TBI models. (B) A range of studies for the tissue loading conditions associated with mild TBI in humans provides a design corridor for modifying the cortical impact model to reproduce these mild TBI conditions in the mouse. (C) A finite element model of the mouse cortical impact model was used to study the tissue strain and strain rates that appear in the cortex and hippocampus when both impact depth and velocity are varied. The range of conditions for 0.43 m/s, 2 mm impact depth appears in (B) (black filled circles), while (D) shows the strain distribution of a mouse brain subject to these cortical impact conditions. References used: FE-Human (, , –64), FE-Round (65), FE-Porcine (66); PM-Human (23), PM-Primate (24), PM-Porcine (43), FE CCI-Rodent (67, 68), PM CCI-Cat (69), PM FPI-Cat (45). FE, finite element; PM, physical model.

Figure 2

Parameters of mild CCI. The mild CCI (mCCI) uses a hemispherical silicone tip, actuated by a solenoid (A). A potentiometer records the displacement of the impactor, while the angle and height of the impactor are fully adjustable. A comparison of displaced tissue volume between mild CCI (2 mm impact depth, 0.43 m/s impact velocity) and traditional CCI (1 mm impact depth, 4–6 m/s impact velocity) is shown in (B). (C) Shows the range of impact speeds between tCCI (2.0–6.0 m/s) and the velocity range that would generate clinical TBI strain rates (17–104 s⁻¹ for 0.1–0.6 m/s, respectively). The red line is the impactor speed used in this study (0.43 m/s for a strain rate of approximately 75 s⁻¹). Brains perfused 8 days after a sham injury (D), mCCI with 2 mm impact depth (E), and tCCI with an impact depth and speed of 1 mm and 6.0 m/s, respectively (F).

Figure 3

Mild controlled cortical impact induces extravasation of the blood–brain barrier (BBB) at the injured region. The pattern of extravasation is shown and pictorially depicted (A). Co-labeling of EB and Neurotrace^® for the cortex (B), CA3 (C), and dentate gyrus (D) show that the EB positive cells are mostly neurons.

Figure 4

Neuronal degeneration appears 24 h following mild controlled cortical impact. Fluoro-Jade B (FJB) positive cells appeared in the cortex (iv), dentate gyrus (v), and CA3 (vi) on the ipsilateral side; no labeling was observed in the contralateral hemisphere [(i–iii), respectively] (A). (B) The number of FJB positive cells in injured animals was significantly increased compared to sham in both the cortical and hippocampal regions for bregma sections −1.5 to −3.0 (cortex p < 0.001 for each section −1.5 to −3.0, for section −3.5 p = 0.0635, hippocampus p = 0.0113 for section −1.5, p < 0.001 for sections −2.0 to −3.5; n = 5 sham, n = 10 injured). Data are expressed as media ± SEM.

Figure 5

Neuronal degeneration persists 8 days after mild controlled cortical impact. Eight days after injury, Fluoro-Jade B staining was located in the cortex (ii), dentate gyrus (iv), and thalamus (vi) (A). The contralateral regions did not show similar FJB staining (i, iii, v). (B) Degeneration in the cortex was still significantly elevated at 8 days compared to sham (p < 0.001 for all sections). (C) In the hippocampus, all sections show significant increase in FJB staining (bregma −1.5 p = 0.0105, −2.0 p = 0.0018, −2.5 p = 0.0102, −3.0 p = 0.0058, −3.5 p = 0.0338). (D) In the thalamus, all sections showed significant levels of FJB staining (−1.5 p = 0.0024, −2.0 and −2.5 p < 0.001, −3.0 p = 0.0081, −3.5 p = 0.0323). Samples sizes were n = 5 for sham, n = 11 for injured. Data are expressed as media ± SEM.

Figure 6

Axonal injury appeared in the subcortical white matter after cortical impact injury. (A) The amyloid precusor protein (APP) staining in the subcortical white matter on the ipsilateral hemisphere (ii) show a number of varicosities (insert a). Contralateral (i) regions between injured and sham brains (B) do not show similar axonal pathologies. (C) Quantification of the axonal varicosities shows significant number of APP varicosities near and at the epicenter of impact (p < 0.001 for all three bregma sections, n = 5 sham, n = 11 injured). Data are expressed as media ± SEM.

Figure 7

Astrocyte reactivity is evident 8 days following mild cortical impact injury. (A) In the ipsilateral cortex, there was no prominent glial scar (ii), although there were clear regions of astrocyte reactivity. The lesion center (LC) also exhibited less GFAP expression than the lesion edge (LE). (i) The contralateral cortex did not significant GFAP staining. In the hippocampus, there was slight increase of GFAP expression in the ipsilateral compared to contralateral hemisphere [(iv) versus (iii)], especially in the dentate gyrus. The ipsilateral thalamus also saw an increased GFAP expression [(vi) versus (v)]. (B) Comparing the contralateral and ipsilateral GFAP expression averaged showed that astrocyte reactivity was significantly elevated in the injured hemisphere across all five sections (p < 0.001 for each brain region); between sham and injured on the ipsilateral side, there was a significant increase for all three brain regions (p < 0.0001). Sample size n = 5 sham, n = 11 injured. Data are expressed as media ± SEM.

Figure 8

Microglial activation occurs 8 days following injury. (A) The ipsilateral cortex showed increased microglia migration, especially to regions that also showed FJB staining at 8 days. The ipsilateral cortex showed greater presence of activated microglia in the lesion edge (LE) compared to the lesion center (LC) (E). (B) The contralateral cortex showed minimal microglia presence and no activated microglia. The ipsilateral hippocampus (F) and thalamus (G) show increased microglia migration compared to contralateral [hippocampus (C), thalamus (D)] but both had less presence of activated microglia compared to the cortex. Data are expressed as media ± SEM.

Figure 9

Mild controlled cortical impact leads to a transient behavioral impairment. A behavioral testing paradigm to examine both cortical and hippocampal function after impact is shown in (A). Rotarod (RR) training occurred the day before injury; RR was given on days 1–3 post injury. Spatial object recognition (SOR) was implemented on days 4 (training) and 5 (test day). Contextual fear conditioning (CFC) was given on days 7 (training) and 8 (test). The animals were perfused on day 8 after CFC. The rotarod results (B) indicate injured animals faulted at lower speeds with significantly lower fault and fall latency times as determine by a repeated measures ANOVA (p = 0.0009 and 0.0017, respectively). In neither the SOR (C) nor the CFC (D) (p = 0.2345, p = 0.2182, respectively) did injured mice show significant level of altered behavior. Sample size n = 17 sham, n = 19 injured. Data are expressed as media ± SEM.