Animal Models: Cerebral Hemorrhage

Introduction

Worldwide, stroke is the second leading cause of death, affecting approximately 6 million people in 2012. Intracerebral hemorrhage (ICH) accounts for 13% of all stroke cases and is associated with high mortality and morbidity. As many as 50% of patients die within 1 month after ICH onset, and only 20% of patients have functional independence at 6 months. However, preclinical research is less focused on ICH than on ischemic stroke, and only a few phase 3 clinical trials have been completed. To better mimic the clinical pathological progress of ICH, researchers have developed several preclinical animal models in pigs, dogs, rabbits, cats, and rodents, and in different brain regions, including striatum, cerebroventricles, cortex, hippocampus, and hypothalamus. Two well-established and commonly used preclinical ICH models are the whole blood model and the collagenase model ( Fig. 64.1 ). Besides these two, researchers also use the microballoon insertion model to mimic the mass effect of ICH, and hypertensive mice for induction of spontaneous ICH.

Blood Injection Model

The blood injection model was first used in rats and was later adapted for use in mice . In this model, whole blood is drawn from the animal itself or from a donor animal and injected directly into the brain. Fresh or heparinized blood can be used, as can blood components or metabolites such as platelets, thrombin, hemin (oxidized heme), or iron. The volume of the blood or blood metabolite used varies among studies and species. In most studies with mice, researchers have injected 10–40 μL of whole blood directly into striatum (x: 0.8; y: 2.0; z: 2.8) at 0.5–2 μL/min. In rats, up to 280 μL is injected at 10 μL/min (3.5 mm to the right of and anteroposterior to the bregma; z: 6 mm). However, because only a limited volume of blood can be infused into the mouse striatum before it begins to flow back along the needle track, a double-injection method was devised. The investigator injects a small amount of blood (one-third of the total volume), waits for it to clot (5–10 min), and then pushes the needle deeper and injects the remaining blood . This method largely prevents the blood backflow and makes the size of the hematoma more reproducible. Even a triple-injection model has been used .

Blood metabolite injection can be used to study the effects of a single element on brain damage, whereas the whole blood model mimics almost every aspect of ICH pathology. The use of autologous whole blood mimics ICH better than donor blood because the donor blood causes more severe brain edema and induces more inflammatory responses . However, donor blood or blood components can be drawn from fluorescently labeled animals, enabling researchers to observe blood cell diffusion, absorption, and even engulfment by phagocytes.

The blood injection model is easy to perform and reproducible and permits control of hematoma size, but the time required to complete an injection in one animal is three times that of the collagenase model. Additionally, the blood model does not fully represent human ICH pathology because it lacks the underlying vascular pathology and rupture; the hematoma size does not continue to increase in the early hours, whereas most patients with ICH have continuous bleeding; and the injected blood can accumulate under the corpus callosum . However, it simplifies the disease in a way that allows investigators to better study the effects of blood itself and may be a good model for studying white matter injury because extravasated blood travels along white matter fiber tracks.

Collagenase Injection Model

In the second commonly used ICH model, bacterial enzyme collagenase is injected into the striatum. Injection of 0.0345–2 U/0.5 μL collagenase causes a breakdown of the blood–brain barrier (BBB) and results in active breeding for 6 h or longer during the acute phase of ICH . The procedure is very easy and quick and does not require as much technical expertise as the blood injection model. An experienced investigator can generate 16 animal models in one day. The hematoma location is very reproducible, but its size may vary because different animals react differently to collagenase. Some researchers believe that bacterial collagenase may cause more severe inflammation than blood injection, although our group and others have not observed activation of microglia in cultures treated with collagenase in vitro . Interestingly, Kleinig et al. reported that hemoglobin crystallization is more prominent in the autologous blood injection model than in the collagenase injection model, and may exaggerate associated inflammatory responses.

The collagenase model is relevant to the clinical condition because bleeding continues in 14–20% of all patients with ICH and lasts for over 6 h in 17% of cases . As a result of BBB breakdown, many blood cells infiltrate the hematoma (B cells, T cells, monocytes, macrophages, neutrophils, etc.). Thus the collagenase model provides a good platform to study blood cell infiltration and the role of various blood cells after ICH (inflammation and brain repair). Of course, this model does not mimic clinical ICH in every respect. Collagenase induces rupture and bleeding of many small vessels, whereas a deep small artery rupture is the main cause of ICH in humans. Additionally, the dosage of collagenase injected into the brain is much greater than the level of endogenous collagenase in human ICH .

Comparison of Blood and Collagenase Models