Spontaneous ischaemic stroke lesions in a dog brain: neuropathological characterisation and comparison to human ischaemic stroke
Neuropathological changes in the affected area of the dog brain corresponded well to what has previously been described for 3-day-old infarcts in humans [33, 34] and experimental murine models [35], and included the presence of injured neurons, reactive microgliosis and astrocytosis as well as neutrophil granulocyte and macrophage infiltrations in the peri-infarct area.
In the canine ischaemic stroke brain, reactive microglia were found in the peri-infarct. The pathological changes observed in the present study are similar to those in the murine permanent MCA occlusion experimental model [36, 37]. Microglia are known to monitor the microenvironment of the brain and to react instantly to injury by undergoing morphological and functional changes [37, 38], thus neuronal death is suspected to induce transformation to phagocytic microglia with ameboid morphology as observed closest to and in the necrotic tissue in the present study [39]. Following the dynamic role of microglia in relation to the formation of the ischaemic lesion, microglia are the subject of a growing research interest [40, 41].
Astrocytosis was demonstrated in the cortical peri-infarct zone. Astrocytes function among other by maintaining the vascular tone changes following neuronal activity, and are capable of both secreting and absorbing neural transmitters. Immediately following injury to the brain, reactive astrocytosis develops. While a negative effect of astrocytosis by increasing infarct size has been shown [42], astrocytes at the same time have the potential to decrease the detrimental excitotoxicity [43, 44]. It is further known, that astrocytes in damaged tissue can induce a microglial response [37]. Whether astrocytes are primarily beneficial in terms of recovery or only exacerbate lesion progression is thus controversial [45]. Accordingly, this cell type should be further studied in animal models of ischaemic stroke, including the dog.
Neutrophil granulocytes were recognized based on nuclear morphology, which is a method that has previously proved reliable when evaluating TB stained sections [46]. In the present study, infiltration of neutrophil granulocytes into the necrotic centre of the canine brain parenchyma was observed (Fig. 3). This is in accordance with previous reports from experimental studies in rats and mice, which have shown that neutrophil migration into the parenchyma of a brain affected by ischaemic stroke peaks within the first 48 h [47, 48]. However, neutrophilic reactions following ischaemic stroke are not fully understood [49–52]. In humans, neutrophilic granulocytes are known to play a potentially harmful role with regard to infarct progression [53, 54]. Consequently, neutrophils in ischaemic stroke have been studied with the aim of developing novel treatments. Investigated potential targets include inhibiting activation, recruitment, and transmigration of neutrophilic granulocytes [49]. In humans, the proportion of leukocytes made up of neutrophils in the peripheral blood is approximately 50–70% [55]. In contrast, neutrophils in mice only constitute around 8–24% of the peripheral blood leukocytes [56], while the dog, interestingly, has a peripheral blood composition highly similar to humans with neutrophils forming approximately 60–80% of the peripheral blood leukocytes [57]. It would therefore be of interest to investigate the relationship between neutrophils and blood–brain barrier breakdown, haemorrhagic transformation, and the impact on final neurological outcome [49] in dogs with spontaneous ischaemic stroke.
When evaluating the dog as a potential spontaneous animal stroke model, it seems relevant whether the ischaemic stroke was caused by a local thrombus or by an embolus. In the present study, a thrombus or embolus was neither identified at necropsy nor at histological examination even though this was the suspected underlying cause. This might, however, be explained by the fact that embolus reduction in vivo as well as post-mortem in dogs usually takes place within a few hours [58]. In the present case, however, an embolus as the underlying cause of the infarct was strongly suspected due to the presence of petechial haemorrhages indicating haemorrhagic transformation, which is typically seen with embolic infarcts in humans [59]. In humans, the majority of ischaemic stroke events are caused by thromboembolism [55]. Atherosclerosis, which is the most frequent type of vascular pathology associated with arterial thrombosis in humans, seems rare in dogs and is most often associated with diabetes mellitus or hypothyroidism [16, 58]. Even though the T4 and free T4 levels were low and TSH was increased in the dog reported here, there were no clinical signs of concurrent hypothyroidism and no atherosclerosis was identified on histopathology. This further support the hypothesis of an embolus having caused the ischaemic stroke in the dog investigated.
The most common subtype of ischaemic stroke in humans is MCA territory infarcts [60], and the majority of animal models therefore aim at mimicking this subtype [61]. MCA occlusion is also a common subtype of spontaneous stroke in dogs [2], and thus offers an interesting spontaneous animal stroke model. So far, experimental studies have provided a substantial insight into the pathophysiology of ischaemic stroke, but effective neuroprotective drugs in experimental studies have failed when tested in human patients. The translational gap may, in part, be a result of the animal models not being able to mimic the complexity of the human disease appropriately [62]. A benefit of studying the pathophysiology of spontaneous stroke in dogs is that confounding factors such as anesthesia and surgical trauma of experimental models are avoided. Further, the similarities between the basic neuroanatomy of the canine and the human brain might explain the resemblance between the clinical disease observed in dogs and in humans with regard to associated neurological deficits and final outcome [2].
Ischaemic stroke seems to be less common in dogs than in humans [63]. The reasons for this remain unclear, but possible explanations could be the presence of vascular anastomoses in the canine brain, the rare occurrence of atherosclerosis in dogs [64] and the rapid dissolution of clots in dogs [58]. The low incidence of ischaemic stroke in dogs poses a hindrance to a widespread use of the dog as a spontaneous animal model for human ischaemic stroke. However, as studies regarding drug development for ethical reasons cannot be carried out in dogs, dogs could never fully replace existing animal stroke models. Instead, important investigations of the pathophysiology of spontaneous ischaemic stroke in dogs may contribute to bridge the translational gap between human patients and experimental animal models.
Our results are based on investigations of a single dog brain and thus cannot stand alone. In future, they should be followed by larger comparative studies, preferably using a multicenter design, which can ensure a high number of brains and support evidence-based conclusions. It would be of interest to perform further neuropathological characterisation of the reactions of neurons and neuroglia at different post stroke time points and investigations of vascular pathology seem highly relevant. Furthermore, white matter neuropathology has previously been linked to clinical deficits in humans with ischaemic stroke [65]. It would therefore also be of interest to investigate such white matter lesions in dogs.