MRI Heralds Secondary Nigral Lesion after Brain Ischemia in Mice: A Secondary Time Window for Neuroprotection

Abstract

Cerebral ischemia in the territory of the middle cerebral artery (MCA) can induce delayed neuronal cell death in the ipsilateral substantia nigra (SN) remote from the primary ischemic lesion. This exofocal postischemic neuronal degeneration (EPND) may worsen stroke outcomes. However, the mechanisms leading to EPND are poorly understood. Here, we studied the time course of EPND via sequential magnetic resonance imaging (MRI) and immunohistochemistry for up to 28 days after 30 minutes occlusion of the MCA (MCAo) and reperfusion in the mouse. Furthermore, the effects of delayed treatment with FK506 and MK-801 on the development of EPND were investigated. Secondary neuronal degeneration in the SN occurred within the first week after MCAo and was characterized by a marked neuronal cell loss on histology. Sequential neuroimaging examinations revealed transient MRI changes, which were detectable as early as day 4 after MCAo and thus heralding histologic evidence of EPND. Treatment with MK-801, an established anti-excitotoxic agent, conferred protection against EPND even when initiated days after the initial ischemic event, which was not evident with FK506. Our findings define a secondary time window for delayed neuroprotection after stroke, which may provide a promising target for the development of novel therapies.

INTRODUCTION

Stroke is among the principal causes of long-term disability and death worldwide. ¹ To date, treatment options are still very limited, in particular due to the narrow time window for thrombolysis.^2–4 Cerebral ischemia in the territory of the middle cerebral artery (MCA) may induce delayed neuronal cell death in non-ischemic brain regions remote from the primary lesion with histopathologic changes occurring as late as days to weeks after the initial ischemic event. ⁵ This phenomenon, also termed ‘exofocal postischemic neuronal death (EPND)’ has been associated with impaired neurologic function^6,7 and the development of vascular Parkinson's disease.^8,9 We hypothesize that, due to its delayed occurrence, EPND could also be a promising target for new neuroprotective strategies beyond the narrow time window for acute stroke treatment. However, so far, the mechanisms and characteristics of EPND have remained incompletely understood. ⁷ In particular, an examination of the time course of secondary changes in a mouse model of stroke is indispensable for studying new treatment approaches. Therefore, we here investigated the pathogenesis of EPND via sequential magnetic resonance imaging (MRI) and immunohistochemistry in a well-established murine model of mild transient brain ischemia. To further elucidate potential pathophysiologic mechanisms involved, we tested the impact of MK-801, an established anti-excitotoxic agent, and of FK506, a potent anti-inflammatory compound. ¹⁰

MATERIALS AND METHODS

Animals and Treatments

All experimental procedures were approved by the respective official committee (G0383/09, LaGeSo, Berlin, Germany) and performed in accordance with the Animal Welfare Act, the European Communities Council Directive of 24 November 1986 (86/609/EEC) and the ARRIVE (Animals in Research: Reporting In Vivo Experiments) guidelines. 129/SV male wild-type mice (BfR, Berlin, Germany) aged 8 to 12 weeks were used. Animals were held under standard laboratory conditions (room temperature: 22± 2° humidity 50%; 12-hour/12-hour light-dark cycle; free access to food and tap water). Animals were kept under specific pathogen free conditions while regularly screened for infections according to FELASA protocols. After arrival at the animal facility the mice did not undergo any experimental procedures for at least one week to allow them to adapt to the novel environment. All efforts were made to minimize the number and suffering of animals used.

The study included a total of 68 animals that were subjected to MCAo. Six animals were excluded due to insufficient drop of relative cerebral blood flow and lack of infarct on MRI (please see below). Separate groups of animals were used for histologic characterization, the MRI time course, and the evaluation of drug treatments. Starting on day 5 after MCAo, subsets of animals (n = 8 per group) received FK506 (Astellas-Pharma, Munich, Germany) 1 mg/kg bodyweight or MK-801-hydrogen maleate (Sigma Aldrich, Seelze, Germany) 1 mg/kg bodyweight, respectively. Both compounds were dissolved in normal saline and administered intraperitoneally in a total volume of 150 μL. The vehicle group received saline as a control. Animals were randomized to their respective treatments. The results section and the figure legends provide detailed information on the number of animals assessed in each experiment as well as on mortality rates.

Model of Cerebral Ischemia

Mice were anesthetized by 1.0% (vol/vol) isoflurane in 69% nitrous oxide (N₂O) and 30% oxygen (O₂) with a facemask. Focal cerebral ischemia was induced by 30 minutes filamentous MCA occlusion (MCAo) and reperfusion as described previously.^11,12 Rectal temperature was controlled and kept constant at 36.5±0.5°. Regional cerebral blood flow was measured by means of a flexible probe and laser-Doppler monitoring (Perimed, Järfälla, Sweden). There was an equivalent decrease in regional cerebral blood flow to less than 20% of baseline at filament insertion and an equivalent increase in regional cerebral blood flow at filament withdrawal.

Magnetic Resonance Imaging Time Course

Follow-up MRI was performed in nine animals on days 1, 4, 5, 7, 14, 21, and 28 of reperfusion. Successful MCAo was confirmed by MRI on day 1. Images were acquired on 7 Tesla rodent scanner (Pharmascan 70/16 AS, Bruker BioSpin, Ettlingen, Germany) with a 16-cm horizontal bore magnet and a 300-mT/m maximum gradient capability. For imaging, a 1H-radio frequency quadrature-volume resonator was used (Bruker, 20 mm, radio frequency coil), as well as a surface coil (Bruker, 72-mm resonator for transmission and a 1H-phased-array surface coil). Data acquisition and image processing were performed with the Bruker software Paravision 4.0 (Bruker BioSpin MRI GmbH, Ettlingen, Germany). The animals were placed on a heating blanket to keep body temperature constant at 37° during the procedure. Anesthesia was performed using 1% isoflurane (Forene, Abbot, Wiesbaden, Germany) in 30% oxygen and 69% nitrogen. During anesthesia, cardiopulmonary function was monitored using a Small Animal Monitoring & Gating System (SA Instruments, Stony Brook, NY, USA). The total time for the measurement did not exceed 15 minutes. The following sequences were used: T2 sequence: repetition time: 4,059 ms; TE: 36 ms; RARE factor 8; 4 averages; 35 axial slices; slice thickness: 0.5 mm; field of view: 2.85×2.85 cm; matrix: 256×256. T2 map: repetition time: 2,000 ms; TE: 20 ms; 2 averages; 5 slices; slice thickness: 0.5 mm; field of view: 2.50×2.50 cm; mm; matrix 256×256. diffusion weighted imaging repetition time: 3,000 ms; TE: 35 ms; 1 average; 5 slices; slice thickness: 1 mm; field of view 2.80×2.80 cm; matrix 128×128; 1 diffusion direction; diffusion duration: 7 ms; diffusion separation: 14 ms; max b value: 3263.5 s/mm²; b value per direction: 0 to 1,300 s/mm². Regions of interest were drawn on the images of the bilateral substantia nigra (SN). Therefore, two regions of interest per animal were used for quantitative analysis. To get T2 values, T2 maps were generated from the Multi Slice Multi Echo images by monoexponential fitting on a pixel-by-pixel basis, using the following equation: Y = A+C × exp(-TE/T2) with A = absolute bias and C = signal intensity. The apparent diffusion coefficient (ADC) was calculated by using the formula ADC= 1/b ln(S(b)/S0) with b being the b value, S(b) the signal intensity with diffusion gradient, S0 the signal intensity without diffusion gradient (b = 0). To obtain an isotropic diffusion estimate, the mean ADC of the three orthogonal directions was calculated using MATLAB (The MathWorks, Natick, MA, USA). As indicated in the figure legends the number of animals used for quantitative MRI analysis varies. The reason for this is movement artifacts that preclude exact measurements. We do not believe that these random artifacts would have resulted in a systematic bias in our data.

Histology

On days 4, 7, 14, and 28 of reperfusion, groups of animals (4 days, n = 6; 7 days, n = 6; 14 days, n = 3; 28 days, n = 4 animals per group) were killed with an overdose of pentobarbital and transcardially perfused with 0.9% saline followed by 4% paraformaldehyde in 0.1 mol/L phosphate buffer. Brains were stored in the fixative for 48 hours and then transferred into 30% sucrose in 0.1 mol/L phosphate buffer for 24 hours. Coronal sections of 40 μm thickness were cut from a dry ice-cooled block on a sliding microtome (Leica, Bensheim, Germany). For the SN, three sections located from approximately bregma −2.5 mm to bregma − 3.5 mm were used. For each staining, the interval between tissue sections was 240 μm. For the striatum, three sections located from approximately bregma 1.0 mm to bregma − 1.0 mm were used. For each staining, the interval between tissue sections was 480 μm. Brain slices were stored at − 20° in a cryoprotective solution (25% ethylene glycol, 25% glycerine, 0.05 mol/L phosphate buffer). Immunohistochemistry followed the peroxidase method with biotinylated secondary antibodies (Jackson ImmunoResearch Laboratories, West Grove, PA, USA), ABC Elite Reagent (Vector Laboratories, Burlingame, CA, USA) and diaminobenzidine as chromogen. Primary antibodies were applied in the following concentrations: anti-NeuN (mouse, 1:100; Millipore, Darmstadt, Germany), anti-tyrosine hydroxylase (TH; rabbit, 1:250; Abcam, Cambridge, UK) and anti-Iba1 (rabbit, 1:500; Wako, Neuss, Germany).

Cell Counts

The number of NeuN-positive cells, the number of TH-immunoreactive cells, and the number of Iba-1 positive cells were quantified essentially as described previously. ⁶ Ipsilateral cell densities (i.e., on the same side as the primary ischemic lesion) are expressed as the percentage of cell densities on the contralateral side of the same animal. ⁶ Briefly, the StereoInvestigator platform (Microbrightfield, Williston, VT, USA) attached to a Leica DM RA microscope (Bensheim, Germany) was used. Each reference space was delineated at a low magnification (×5) and cells in the ipsilateral and contralateral midbrain (TH, Iba-1, and NeuN) or in the striatum (NeuN) were then counted using a higher magnification (×20) objective. Three reference sections per animal were analyzed, respectively.

Statistics

All experiments were performed in a blinded manner. Values are given as means ±s.e.m. All statistics was performed using GraphPad Prism for Mac (Version 6.0, GraphPad Software, San Diego, CA, USA). Ipsilateral cell densities were calculated as the percentage of cell densities on the contralateral side of the same animal. Mann-Whitney U test was used for comparison of ipsilateral versus contralateral cell densities for the respective time point. For a comparison of the different time points, a Kruskal-Wallis test followed by post hoc analysis using Dunn's method was performed. The MRI time course was analyzed using repeated ANOVA measurements and Tukey post hoc testing. Comparisons between vehicle and treated groups were performed with Student's t-test or Mann-Whitney U test depending on normality of distribution with level of significance set at 0.05 and two-tailed P values. P < 0.05 was considered as statistically significant.

RESULTS

Magnetic Resonance Imaging Time Course After Middle Cerebral Artery Occlusion

At day 1 after 30 minutes MCAo/reperfusion, all mice showed an increase in left striatal T2 signal intensity, delineating the primary ischemic lesion in the left striatum. However, T2 alterations in the midbrain (Figure 1A, upper row) were not apparent on day 1. As the primary lesion in the left striatum started to fade at approximately 4 days after MCAo, an ipsilateral delayed T2 hyperintensity (Figure 1A, lower row) associated with changes in diffusion weighted imaging and ADC (Figure 1B) emerged in the midbrain. No differences in ADC values or T2 values in the SN were observed at 24 hours after MCAo. By contrast, a significant increase in T2 values (P<0.02) and a significant decrease in ADC values (P<0.05) were detected in the ipsilateral as compared with the contralateral SN starting on day 4. These MR alterations persisted till day 7 (Figures 1C and 1D). No significant differences in T2 and ADC values were detectable at any of the later time points (i.e., days 14, 21, and 28). Maximum alterations in ADC and T2 values were found at day 5 after MCAo (Figures 1C and 1D; T2 value: P < 0.005; ADC value: P < 0.02).

Figure 1.

Magnetic resonance imaging (MRI) time course after middle cerebral artery (MCA) occlusion (MCAo). (A) Representative MRI showing evolving changes in T2WI in the territory of the MCA (left column) and at the level of the substantia nigra (SN) (right column). In the lateral striatum T2 signal hyperintensity was found day 1 after MCAo (white arrow), while no changes in T2 were detectable in the SN. At day 4, T2 hyperintensity was clearly detectable in the ipsilateral SN (circled red). (B) The ipsilateral SN showed low signal intensity in apparent diffusion coefficient (ADC) mapping (upper white arrow) on day 4 after 30 minutes of MCAo. (C) Days 4, 5, and 7 after MCAo the T2 value of the ipsilateral SN was significantly higher than that on days 1, 14, and 21, compared with the contralateral SN. (D) On days 4, 5, and 7 after MCAo, the ADC was significantly lower in the ipsilateral SN compared with the contralateral side. Values are mean ± s.e.m.; *P < 0.05, **P < 0.02, ***P < 0.005; repeated ANOVA with Tukey post hoc testing; (T2 value, n = 9, ADC, n = 5 animals per group).

Immunocytochemical and Histologic Evaluation

Immunocytochemical and histologic evaluation of the ipsilateral SN as compared with the contralateral side showed no significant difference in the number of TH-immunoreactive and NeuN-positive cells on day 4 after MCAo, while Iba-1 cells showed a significant increase at this early stage. However, on day 7, the number of TH-positive as well as NeuN-positive cells had decreased significantly compared with the contralateral side. (Figures 2A–2C, 3A, and 3B). In both cell populations, NeuN- and TH-positive cells, the changes persisted until later time points, notwithstanding slight increases for NeuN-positive cells on day 28 compared with earlier time points (Figures 3C and 3D). Comparison across the different time points showed a highly significant decrease in the densities of ipsilateral NeuN-positive cells on day 7 relative to day 4 (P < 0.02) as well as for TH-positive cells on days 14 and 28 compared with day 4 (P < 0.03 and P < 0.001, respectively). This was associated with a reduction in ipsilateral NeuN and TH counts to almost 50% of the contralateral ones (Figures 2D–2F and 3B–3D). Iba-1 immunoreactivity in the ipsilateral SN persisted on an elevated level with a peak at 7days compared with the other time points (Figure 3B). Iba-1-positive cells clearly displayed the morphology of activated microglia (Figures 2G–2I and 3A–3D; 4 days, n=6; 7 days, n=6; 14 days, n=3; 28 days, n=4 animals per group).

Figure 2.

Nigral histology after middle cerebral artery occlusion (MCAo). Boxed areas in (A, D and G) are shown in higher magnification in (B, C, E, F, H and I), respectively. (A) Representative image illustrating reduced density of TH-positive cells in the ipsilateral substantia nigra (SN) day 14 after MCAo (arrow). (D) Reduced density of NeuN immunohistochemistry in the ipsilateral SN day 14 after MCAo (arrow). (G) Upregulation of the activated microglial marker Iba-1 day 14 after MCAo showing enhanced microglial invasion and activation in the ipsilateral SN (arrow).

Figure 3.

Nigral histologic changes after middle cerebral artery occlusion (MCAo). (A) Four days after MCAo no significant decrease in the number of TH-positive or NeuN-positive cells in the ipsilateral substantia nigra (SN) was found as compared with the contralateral side, while Iba-1 cells showed a significant increase at this early stage. (B) The number of TH- and NeuN-positive cells in the ipsilateral SN was significantly decreased as compared with the contralateral side, starting day 7 after MCAo. The reduction in TH- and NeuN-positive cells persisted at (C) 14 days and (D) 28 days after MCAo. An increase in Iba-1-positive cells was found reaching its maximum day 7 after MCAo then declining but remaining on a higher level. Values are mean ±s.e.m.; *P < 0.05, **P < 0.03; Mann-Whitney test (4 days, n = 6; 7 days, n = 6; 14 days, n = 3; 28 days, n = 4 animals per group).

Treatment with FK506 and MK-801

In subsets of animals, treatment with MK-801 or FK506 was initiated on day 5 after 30 minutes MCAo/reperfusion. Treatment with MK-801, but not with FK506, conferred neuroprotection based on histologic evaluation on day 28 after MCAo (FK506, n = 6; MK-801, n = 7; vehicle, n = 7 and n = 5 respective animals per group). The neuroprotective effect of MK-801 was evident for both cell populations, NeuN- and TH-positive cells (Figures 4A and 4B). By contrast, no effect of either treatment on the number of Iba-1-positive cells was found (Figure 4C). One animal in the MK-801 group and two animals in the FK506 group died (in the respective control groups, one and three vehicle-treated mice died). Thus there was no apparent influence on mortality of the different treatments. Furthermore, no differences between treatment and vehicle groups regarding neuronal damage within the primary ischemic lesion were found (Figure 4D).

Figure 4.

Treatment with FK506 and MK-801. (A) Delayed treatment with FK506 showed no protective effect on TH-positive cells while MK801 significantly preserved TH-positive cells, ipsilateral substantia nigra (SN) compared with contralateral side 28 days after middle cerebral artery occlusion (MCAo). (B) A similar effect of MK-801 was found for NeuN-positive cells. (C) No effect of treatment with FK506 or MK-801 on the number of Iba-1-positive cells was found. (D) There was no difference between the treatment and vehicle groups regarding neuronal damage within the primary ischemic lesion in the striatum. Values are mean ± s.e.m.; *P<0.05, **P < 0.005; Mann-Whitney test (FK506, n = 6; MK-801, n = 7; vehicle, n = 7 and n = 5 respective animals per group).

DISCUSSION

This study has the following major findings: (1) Secondary neuronal degeneration in the SN occurs within a week of striatal ischemia and is characterized by marked neuronal cell loss. (2) Before significant cell loss on histology, transient MRI changes can be detected in the SN, indicating the onset of cellular edema and tissue damage. Our results suggest that MRI can be used to predict EPND in the SN after striatal infarction. (3) Treatment with MK-801, an NMDA receptor antagonist, but not with FK506, a potent anti-inflammatory drug, conferred neuroprotection even when treatment was initiated as late as on day 5 after MCAo. (4) Hence, our findings broaden the concept of neuroprotection after stroke and show a ‘delayed time window’ for potential interventions extending days after the ischemic event.

Together with the striatum, the SN forms a unique projection system that holds great importance for motor and behavioral functions. It has been suggested that EPND causes neurologic impairments that cannot be attributed to the primary ischemic lesion.^6,9 If this were so, then EPND would be of great functional significance to stroke survivors. Furthermore, due to its delayed occurrence and prolonged development, EPND could also be a promising new target for neuroprotective treatments. The mechanisms underlying EPND are complex and only poorly understood. ⁷ So far, data from studies in mice are sparse. Our study therefore addressed the question of the dynamic changes involved in EPND in SN after striatal infarction. Furthermore, we investigated the effects of an anti-excitotoxic as well as of an anti-inflammatory drug on EPND when treatment was initiated in the subacute phase after MCAo.

Here, sequential neuroimaging examinations revealed transient MRI changes in the midbrain emerging as early as day 4 after MCAo. Significant alterations were apparent in ADC as well as in T2 values and clearly preceded histopathologic changes. Robust evidence of neuronal degeneration was only detectable from day 7 onwards. On day 7, a reduction in the density of NeuN-positive neurons to almost 50% of the contralateral side was found. This change persisted at later time points. In contrast, several studies in rats have suggested that the loss of TH immunoreactivity may only be transient.^13,14 However, our data extending up to 28 days together with the results of Kronenberg et al. ⁶ studying EPND up to 4 months after MCAo yielded a chronic and permanent reduction of nigral TH⁺ cells in mice. ⁶ Furthermore, these findings are supported by decreased dopamine levels in the ipsilateral striatum up to 4 months after 30 minutes MCAo/reperfusion.^6,15 To the best of our knowledge, this is the first study using sequential neuroimaging examinations to characterize MRI changes in the murine midbrain after experimental stroke. In rats, a comparable time course of secondary MRI alterations was found after ischemia in the territory of the MCA thus supporting our results.^5,16 A number of MRI studies of stroke survivors including a recent systematic prospective follow-up investigation at 3T have reported midbrain T2 hyperintensities as well as ADC decline developing in a delayed manner between days 6 and 10 after striatal stroke (Winter et al., in press).^17–19 We hypothesize that, in stroke survivors, secondary neuronal degeneration may manifest in the subacute phase after stroke in the form of subtle cognitive and/or neuropsychiatric deficits, which may negatively impact long-term functional outcomes. Furthermore, EPND may also reduce a patient's ‘cognitive reserve’, rendering him or her more susceptible to untoward behavioral effects of subsequent brain pathology. ⁹ In addition, as the SN represents a key structure governing motor activity, nigral EPND in particular may have crucial implications for motor recovery after stroke. In a large population-based study, Becker et al. ²⁰ found a two-fold increase in the relative risk of developing Parkinson's disease in patients with a history of ischemic stroke. The potential clinical relevance of EPND has recently garnered increased attention.^7,9,21

In line with nigral EPND, an increased number of activated microglia was found starting at 4 days after MCAo and thus before significant neuronal cell loss. Early inflammatory changes in the SN occurring before neuronal alterations have been reported recently. ⁸ Importantly, this finding might also explain the fact that, in our study, treatment with the calcineurin inhibitor FK506, a potent anti-inflammatory agent, failed to confer neuroprotection, since, in all probability, it was administered too late. However, treatment with MK-801 exerted neuroprotective effects even when treatment was begun on day 5 post event. This protective effect of MK-801 was borne out by separate analyses of NeuN⁺ and TH⁺ cells. The SN receives inhibitory GABAergic projections from the striatum and glutamatergic inputs from the subthalamic nucleus. The destruction of the GABAergic pathway after striatal ischemia may reduce inhibitory afferents to the SN, triggering increased neuronal firing with subsequently increased metabolism and finally neuronal cell death. ⁵ In line with this hypothesis, a protective effect of MK-801 against nigral degeneration has been shown previously, indicating the importance of transmission via NMDA receptors for EPND. ¹⁴ However, Yamada et al. started MK-801 24 hours after transient MCAo in the rat, whereas in our study, to preclude potential effects of the compounds on the primary lesion, treatment was only started 5 days after the ischemic insult. Delayed administration of MK-801 and confirmation of similar lesion sizes in the treatment and control groups allow us to clearly disentangle the effects of the compound on the primary lesion from those on EPND. Anterograde and retrograde degeneration of fibers may induce excitotoxicity in the SN and ultimately delayed neuronal death. Most likely, both these mechanisms are simultaneously involved in nigral EPND. The combination of a T2 hyperintensity and decreased ADC values indicates both cytotoxic and vasogenic edema. The delayed occurrence of these MRI alterations in the SN suggests that spontaneous firing with subsequently increased metabolism causes cell death over time.

Our study defines the dynamics of EPND in a murine stroke model, which is a crucial step to further elucidate the pathophysiologic mechanisms involved. These could prove to be valuable targets for the development of novel therapeutic approaches. The importance of characterizing new animal models for proper preclinical testing has just recently been emphasized. ²²

CONCLUSIONS

This is the first study to establish and characterize a reliable model of EPND in the mouse using a multimodal approach with both neuroimaging and immunohistology. Secondary exofocal neuronal degeneration in the SN occurs within the first week of striatal ischemia and is characterized by marked neuronal loss, especially of TH⁺ dopaminergic neurons. In our study, transient MRI changes in the SN preceded overt signs of EPND on histology. Together with the striatum, the SN forms a unique dual projection system, which is of critical importance for behavioral and motor functions. It has been suggested that EPND may be responsible for neurologic deficits that cannot be attributed to the primary ischemic lesion. In particular, EPND of dopaminergic neurons in the midbrain (SN and ventral tegmental area) may contribute to motor impairments and affective sequelae after brain ischemia.^6–9,23 Novel therapeutic approaches for the prevention of EPND after stroke may contribute to improved stroke outcomes.