Characterizing infarction and selective neuronal loss following temporary focal cerebral ischemia in the rat: A multi-modality imaging study

Abstract

Background and purpose

Current models dictate that, depending on occurrence of early reperfusion, the ischemic penumbra either undergoes or escapes infarction (i.e., “pan-necrosis”). However, tissue outcome following temporary middle-cerebral artery occlusion (tMCAo) in rodents can also include selective neuronal loss (SNL), which even if subtle may impede functional recovery. In order to explore the pathophysiology of ischemic stroke, determine potential therapeutic targets and monitor effects of therapy, in vivo imaging surrogates of these varied histopathological outcomes applicable in the clinical setting would be useful. Although hyperintense signal on T₂-weighted MRI in the chronic post-stroke stage is considered a reliable surrogate of tissue infarction, SNL is not associated with T₂W abnormal signal. In the clinical setting, the neuron-specific PET ligand ¹¹C-flumazenil (FMZ) has been used to identify both pan-necrosis and peri-infarct SNL, but this inference has not been histopathological confirmed so far. Here we investigated the late tissue sequelae of tMCAo in the rodent using in vivo T₂W MRI and FMZ-PET against post mortem immunohistochemistry as gold standard.

Methods

Adult spontaneously hypertensive rats (SHRs) underwent 45 min distal-clip middle-cerebral artery occlusion and, 28 days later, FMZ-PET and T₂W-MRI, immediately followed by immunohistochemistry for neuronal loss (NeuN), activated microglia and astrocytosis. Based on standard histopathological definitions, ischemic lesions were classified into pan-necrosis, partial infarction or SNL. NeuN changes and FMZ binding across the whole hemisphere were quantified in the same set of 44 regions-of-interest according to previously validated protocols; linear regressions between these two measures were carried out both within and across subjects.

Results

Both cortical pan-necrosis/partial infarction and SNL were present in all rats except one, where SNL was isolated and extensive. Infarction/partial infarction, but not SNL, was associated with T₂W hyperintense signals and cortical atrophy. In contrast, FMZ binding was decreased in all types of lesions including SNL, in proportion with NeuN staining intensity both within (p < 0.05 to < 0.001) and across (p < 0.001) subjects, including the subject that showed pure SNL (p = 0.01).

Conclusion

This novel study revealed three main facts: i) long-term histopathological cortical changes following 45 min tMCAo in SHRs included all three of SNL, partial infarction and frank infarction; ii) T2W MRI showed conspicuous high signal lesions for complete or partial infarction, but no changes for SNL; and iii) FMZ-PET was sensitive to all three types of tMCAo-induced histopathological changes, including isolated SNL, suggesting it is a valid surrogate for the histological sequelae of focal cerebral ischemia. In addition, the finding of almost universal completed cortical infarction at 28 days differed from our previous findings at 14-day survival using the same model and rat strain, where SNL was the almost exclusive outcome, possibly representing delayed infarct maturation. Prospective studies are needed to investigate this interesting possibility.

Introduction

According to the dominant “core/penumbra” model of anterior circulation stroke, the acutely ischemic brain comprises both a core of rapidly irreversibly damaged tissue and a less severely hypoperfused, mainly cortical area termed the ischemic penumbra, which is still viable but will inevitably undergo complete tissue infarction (i.e., “pan-necrosis” Auer, 2008, Garcia et al., 1996a, Garcia et al., 1996b), unless rapidly reperfused (Astrup et al., 1981, Baron, 1999, Moskowitz et al., 2010). Given that final infarct volume strongly impacts neurological outcome (Brott et al., 1989), rescuing the penumbra has major clinical relevance (Baron, 1999, Furlan et al., 1996) and accordingly acutely administered intravenous thrombolysis strongly improves functional outcome (Lees et al., 2010).

Despite this undisputed operational validity of the core/penumbra model, studies in rodents have demonstrated that following temporary middle cerebral artery occlusion (tMCAo) tissue outcome is not dichotomic but can range from pan-necrosis right through to subtle changes in dendritic arborization, with selective neuronal loss (SNL) and partial infarction as intermediate stages (Auer, 2008, DeGirolami et al., 1984, Garcia et al., 1996a, Garcia et al., 1996b, Heiss and Rosner, 1983, Ito et al., 2007, Jones et al., 1981, Li et al., 1999). Of particular interest is SNL involving the penumbra rescued from infarction (Garcia et al., 1996b, Garcia et al., 1997, Hughes et al., 2010, Li et al., 1999, Li et al., 2000, Sicard et al., 2006a). Although SNL does impact early behavioral performance in rodent models (Sicard et al., 2006a, Sicard et al., 2006b), its long-term effects, particularly if cortical areas are affected, are less clearly established. Specifically, given the major role of peri-infarct plastic processes (Furlan et al., 1996, Jaillard et al., 2005), it is likely that SNL impedes optimal plasticity post-stroke and hence delays or dampens behavioral recovery. In turn, SNL may represent a novel therapeutic target in the post-acute phase of ischemic stroke (Baron, 2005).

In order to explore the pathophysiology of ischemic stroke, determine potential therapeutic targets and monitor effects of therapy, in vivo imaging surrogates of post-tMCAo histopathological tissue outcomes, including SNL, applicable in the clinical setting would be extremely useful. For instance, using T₂-weighted MRI in the chronic post-stroke stage, hyperintense signal is considered a reliable surrogate for tissue infarction, based on studies in rodents (Jacobs et al., 2000, Ringer et al., 2001, van Lookeren Campagne et al., 1999, Wagner et al., 2012), and accordingly is the standard in the clinical setting (or its FLAIR variant, which however may be over-sensitive). Noteworthy, SNL is not known to be associated with abnormal signal on follow-up T₂-weighted imaging in the rodent (Sicard et al., 2006a, Sicard et al., 2006b). In the clinical setting, peri-infarct SNL in the chronic post-stroke stage has been inferred time and again from observations of mildly reduced specific binding of the neuron-specific PET ligand ¹¹C-flumazenil (FMZ) — or of its SPECT derivatives — as compared to markedly reduced within infarcted areas (Dong et al., 1997, Guadagno et al., 2008, Hatazawa et al., 1995, Nakagawara et al., 1997, Saur et al., 2006, Yamauchi et al., 2007, Yamauchi et al., 2009). However, the inference that reduced FMZ binding is a valid surrogate of chronic post-stroke ischemic damage including SNL has not received histopathological confirmation so far.

In the present study, we assessed the long-term (28 days post-stroke) sequelae of temporary focal cerebral ischemia in the rat using T2-weighted MRI and FMZ-PET as in vivo imaging markers, against post mortem immunohistochemistry (IHC) as a gold standard to characterize neuronal and tissue damage, from SNL to frank infarction. Even though technically challenging relative to standard proximal MCAo models, we implemented here the distal clip MCAo model (Buchan et al., 1992, Kaplan et al., 1991), given our specific interest in cortical pathology; another advantage of the clip model relative to e.g., the filament model is that it allows direct verification of restoration of blood flow through the MCA following clip removal. Likewise, despite their higher cost, we elected to use spontaneously hypertensive rats (SHRs) in order to enhance our study's clinical relevance (Duverger and MacKenzie, 1988). Our choice of 28 days for the PET and IHC assessments was dictated by our interest here in established, as opposed to potentially still evolving, ischemic lesions, particularly given occasional reports of delayed “infarct maturation” up to 3 weeks after the ischemic insult in SHRs (Buchan, 2001, Lehrmann et al., 1997).