Theranostics 2022, Vol. 12, Issue 10 4477 Ivyspring International Publisher Theranostics 2022; 12(10): 4477-4497. doi: 10.7150/thno.72167 Research Paper Selective brain entry of lipid nanoparticles in haemorrhagic stroke is linked to biphasic blood-brain barrier disruption Zahraa S. Al-Ahmady1,2#, Ben R. Dickie3,7#, Isabelle Aldred2, Dhifaf A. Jasim2, Jack Barrington3, Michael Haley3, Eloise Lemarchand3, Graham Coutts3, Satinderdeep Kaur1, Jessica Bates1, Sarah Curran1, Ruth Goddard1, Megan Walker1, Adrian Parry-jones4,5,7, Kostas Kostarelos2,6, Stuart M. Allan3,7 1. Pharmacology Department, School of Science and Technology, Nottingham Trent University, Nottingham, NG11 8NS, United Kingdom. 2. Nanomedicine Lab, Division of Pharmacy and Optometry, Faculty of Biology, Medicine and Health, AV Hill Building, The University of Manchester, Manchester M13 9PT, United Kingdom. 3. Division of Neuroscience and Experimental Psychology, School of Biological Sciences, Faculty of Biology, Medicine and Health, University of Manchester, Manchester, M13 9PT, United Kingdom. 4. Division of Cardiovascular Sciences, Lydia Becker Institute of Immunology and Inflammation, School of Medical Sciences, Faculty of Biology, Medicine and Health, Manchester Academic Health Science Centre, The University of Manchester, Manchester, United Kingdom. 5. Manchester Centre for Clinical Neurosciences, Salford Royal NHS Foundation Trust, Manchester Academic Health Science Centre, Salford, United Kingdom. 6. Nanomedicine Lab, Catalan Institute of Nanoscience and Nanotechnology (ICN2), Bellaterra UAB Campus, Barcelona, Spain. 7. Geoffrey Jefferson Brain Research Centre, The Manchester Academic Health Science Centre, Northern Care Alliance NHS Group, University of Manchester, United Kingdom. # Shared co-first authorship  Corresponding authors: Dr Zahraa Al-Ahmady, email: zahraa.al-ahmady@ntu.ac.uk; Prof Kostas Kostarelos, email: kostas.kostarelos@manchester.ac.uk; Prof Stuart Allan, email: stuart.allan@manchester.ac.uk © The author(s). This is an open access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/). See http://ivyspring.com/terms for full terms and conditions. Received: 2022.02.18; Accepted: 2022.04.29; Published: 2022.05.26 Abstract Haemorrhagic stroke represents a significant public health burden, yet our knowledge and ability to treat this type of stroke are lacking. Previously we showed that we can target ischaemic-stroke lesions by selective translocation of lipid nanoparticles through the site of blood-brain barrier (BBB) disruption. The data we presented in this study provide compelling evidence that haemorrhagic stroke in mice induces BBB injury that mimics key features of the human pathology and, more importantly, provides a gate for entry of lipid nanoparticles-based therapeutics selectively to the bleeding site. Methods: Haemorrhagic stroke was induced in mice by intra-striatal collagenase injection. lipid nanoparticles were injected intravenously at 3 h, 24 h & 48 h post-haemorrhagic stroke and accumulation in the brain studied using in-vivo optical imaging and histology. BBB integrity, brain water content and iron accumulation were characterised using dynamic contrast-enhanced MRI, quantitative T1 mapping, and gradient echo MRI. Results: Using in-vivo SPECT/CT imaging and optical imaging revealed biphasic lipid nanoparticles entry into the bleeding site, with an early phase of increased uptake at 3-24 h post-haemorrhagic stroke, followed by a second phase at 48-72 h. Lipid nanoparticles entry into the brain post-haemorrhage showed an identical entry pattern to the trans-BBB leakage rate (Ktrans [min-1]) of Gd-DOTA, a biomarker for BBB disruption, measured using dynamic contrast-enhanced MRI. Discussion: Our findings suggest that selective accumulation of liposomes into the lesion site is linked to a biphasic pattern of BBB hyper-permeability. This approach provides a unique opportunity to selectively and efficiently deliver therapeutic molecules across the BBB, an approach that has not been utilised for haemorrhagic stroke therapy and is not achievable using free small drug molecules. Key words: Liposomes, blood-brain barrier, drug delivery, haemorrhagic stroke, lipid nanoparticles https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4478 Introduction in getting enough drug into the injury site due to the protective nature of the blood-brain barrier (BBB) [15]. Stroke poses a major clinical threat, attributing to This highlights a real need to develop new strategies 7 % of deaths worldwide [1]. It is the second cause of to enhance drug delivery to the brain in ICH to death globally affecting 17 million people every year, increase survival rates and improve the post-stroke with only one-third of patients surviving with no quality of life for ICH patients. Under normal complications [2]. Stroke can be categorised into major conditions, the BBB creates a sterile environment for subtypes of different pathophysiology, ischaemic and neurons via tight control of the movement of haemorrhagic stroke. Ischaemic stroke is caused by substances between peripheral blood and neuronal cerebrovascular vessel occlusion resulting from tissues. This interface comprises endothelial cells and thrombus or emboli. Haemorrhagic stroke is caused the tight junction proteins that regulate paracellular by a rupture of a damaged or abnormal blood vessel transport, pericytes, astrocytes, and transcellular leading to local bleeding within the cranium and vesicles which regulate the trafficking of small accounts for 15-20 % of all stroke cases [3]. Although molecules through the endothelial cells [16,17]. haemorrhagic stroke has a lower incidence rate, Maintaining BBB function is vital to maintain normal outcomes are worse than ischaemic stroke with brain function with damage contributing to a host of almost 40 % mortality risk [4]. Haemorrhagic stroke neurological pathologies [16]. Paradoxically, the can be further sub-categorised into intracerebral protective nature of BBB also limits the delivery of haemorrhage (ICH), subarachnoid and interventri- therapeutic drugs into the brain. However, preclinical cular bleeds depending on the location of the data [18] and also human studies [19] have shown that ruptured vessel. The most common of those is ICH, BBB integrity is altered following ICH, which may which results from a direct leak of blood into the brain provide a window of opportunity to deliver drugs parenchyma due to vascular rupture secondary to effectively into the haematoma and surrounding brain hypertension, cerebral amyloid angiopathy, vascular tissue. Therefore, understanding the mechanisms and aneurysm, tumour, or predisposing genetic condi- the timeframe of BBB alteration in ICH is crucial for tions [5]. Age is another un-adjustable risk factor for the development of potential new treatments. ICH which doubles the chance of the disease every 10 Previously, we showed that we can selectively years of life [6]. The increased use of oral target the lesioned area in ischaemic stroke using lipid anticoagulant to prevent ischaemic stroke causes nanoparticles (liposomes) by taking advantage of another threat and carries a 50 % mortality rate in the enhanced transcellular and paracellular pathways occurrence of ICH due to the increased chances of that contribute to BBB disruption after cerebral re-bleeding [7,8] ischaemia [20]. Liposomes offer several advantages Secondary injury is triggered in ICH due to the compared to conventional medications including; a) breakdown of blood products and this has been small size and long circulation half-life that enable shown to stimulate several cellular mechanisms selective and efficient lesion targeting, and b) which in turn exacerbate the original injury. adaptation to encapsulate a wide range of therapeutic Extracellular contact with blood components exerts molecules and contrast agents which offer an ideal neurotoxic effects from the release of free radicals opportunity to image where the drug is being causing vasogenic oedema [9]. Brain-resident (micro- delivered [21–23]. glia) and infiltrating immune cells (macrophages) In this study, we hypothesised that similar attempt to clean up the haematoma debris, which is selective targeting is possible using liposomes in ICH, the initial driving force of the inflammatory response with BBB disruption being the primary mechanism by to stop further inflammation [10] and in parallel, a which entrance is permitted. Such an approach host of chemokines and cytokines are released [11]. provides a unique opportunity to selectively and Red blood cell (RBC) lysis also occurs, triggering efficiently deliver therapeutic molecules across the oxidative stress and free radical formation. Iron BBB, an approach that has not been utilised for ICH toxicity has also been noted to cause microglial therapy and is not achievable using free small drug activation and proliferation, enhancing the molecules due to short circulation half-life [24]. To inflammatory response. [12]. Any recovery in ICH is understand the timing of liposome penetration into thought to be dependent on haematoma resolution, the brain in ICH, we measured brain liposome oedema reduction, neuroplastic abilities of the brain accumulation using radioisotope and optical and neurogenesis [13]. detection methods. To confirm that leakage of Treatments that can salvage damaged brain liposomes was primarily through a disrupted BBB, we tissue after ICH are currently lacking [14]. The lack of used dynamic-contrast enhanced MRI to detect the treatments in ICH is partly due to the complexity of leakage of MRI contrast agent Gd-DOTA. Similar the underlying pathophysiology but also the difficulty https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4479 patterns of leakage were observed strongly indicating over time with a biphasic pattern. Higher accumu- that liposome penetrance at acute timepoints (0-72 h lation is observed both when injected at 3-24 h and post-ICH) is mediated by self-diffusion across the 48-72 h as shown in Figure 1C. Minimum liposomal disrupted BBB. accumulation into the haemorrhagic area was observed when 111In-Lp were injected 24 h after ICH Results for both detection time points (24-25 h and 24-48 h) To establish the validity of our new therapeutic and no liposomal entry was observed in the approach, we have induced ICH in mice by contralateral side from all groups as well as healthy intracerebral injection of bacterial collagenase. This animals (Figure 1B-C). Quantification of 111In-Lp in model degrades the extracellular matrix causing the brain tissues 24 h after injection confirmed the endothelial basement membrane destruction and significant biphasic increase in liposomes accumu- subsequent haemorrhage in the brain parenchyma lation at the lesion site. An increase of 27.5-fold and [25]. This model offers a non-invasive and reprodu- 11-fold was observed at 3-24 h and 48-72 h, cible platform to mimic the pathological alterations respectively, compared to healthy mice (Figure 1D). It seen clinically in ICH, including inflammatory is worth pointing out here, that consistent across all responses and haematoma expansion, which occurs in results we observed that the second episode of 73 % of clinical cases [26]. The other advantage of the hyperpermeability was less than the first, though this collagenase model compared to the autologous blood is still showing to be significant. Quantification of 111In-Lp levels in the cerebrospinal fluid (CSF) model is that the collagenase model mimics the acute cerebrovascular rupture and BBB disruptions seen in indicated no significant differences in CSF liposomal ICH patients [11,27]. Therefore, this model is more levels compared to healthy mice. This suggested relevant to the aim of our study which is focused on minimal clearance of liposomes after entry into the BBB disruption, unlike the autologous blood model brain (Figure 1E). which is often used to study blood clotting effects. The observations reported above with In this study, we injected a liposome platform SPECT/CT were confirmed using fluorescently (HSPC:CHOL: DSPE-PEG2000) intravenously at labelled liposomes coupled with optical imaging and different time points following ICH (3 h, 24 h & 48 h) histological analysis. For this, we injected DiI-labelled and performed a comprehensive analysis of their liposomes (DiI-Lp) at the same time points used for accumulation in the brain and correlated this to SPECT/CT imaging. Each group of mice received a changes in BBB integrity and ICH pathology. These single I.V injection of DiI-Lp at either 3, 24, or 48 h liposomes are around 120 nm in size and slightly post-ICH. Mice were then sacrificed either 2 h after negatively charged (for full characterisation data of injections (3-5 h, 24-26 h & 48-50 h) or 24 h later (3-24 liposomes, refer to Figure S1). h, 24-48 h & 48-72 h) to evaluate the liposomal distribution in the brain and whether this changes Liposomes accumulate at the lesion site in a over time. biphasic pattern and maintain long term IVIS Lumina II imaging (Figure 2B) confirmed localisation after selective entry into the brain the selective accumulation of liposomes into the To study liposomal translocation into the brain haemorrhage site (right side) of the brain as early as 2 after ICH we performed dynamic SPECT/CT imaging h after DiI-Lp injection in the acute phase (3-5 h). after I.V administration of 111In-DTPA-liposomes Quantification of the total fluorescent signal of DiI-Lp (111In-Lp). For full physicochemical characterisation, in the brain 2 h after injection indicated a significant radiolabelling efficiency and stability see Figure S1. increase by 3.3-fold from the 3-5 h group compared to Following induction of ICH with collagenase, 111In-Lp an increase by 1.8-fold and 2-fold detected from 24-26 were injected at three different time points post-ICH h and 48-50 h groups, respectively (Figure 2B-C). In (3 h, 24 h and 48 h) designed to cover the different agreement with SPECT/CT images, evaluation of stages of ICH pathology (Figure 1A). Each injection of liposomal brain accumulation 24 h after injection 111In-Lp was detected using SPECT-CT either 1 h (3-4 revealed a biphasic entry into the brain when injected h, 24-25 h & 48-49 h, Figure 1B) or 24 h later (3-24 h, 3 h and 48 h post-ICH (Figure 3). Quantification of the 24-48 h & 48-72 h, Figure 1C). Each time point optical images indicated a significant increase by represents a separate group that received a single 6.6-fold in the 3-24 h group and 3-fold in the 48-72 h injection of 111In-Lp I.V. Selective liposomal group compared to healthy mice injected with DiI-Lp. accumulation was seen as early as 1 h after 111In-Lp Liposome accumulation was most prominent in the injection when administered acutely (3-4 h) after ICH. ipsilateral side, where the haemorrhage was induced. Repeated SPECT/CT imaging on the same animals Less accumulation was observed when DiI-Lp was indicated that liposomal entry into the brain increased administered 24 h post-ICH and almost no https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4480 fluorescence signal was detected in the contralateral following DiI-Lp administration in sham mice (Figure side at all time points tested. We also found no 3E-F). evidence of liposomal accumulation in the brain Figure 1: Imaging and quantification of liposomes accumulation into the brain after inducing ICH into the right hemisphere (A) Schematic presentation of experimental design and the time frame of 111In-Lp I.V injection after ICH. Each time point represents a separate group that received a single injection of 111In-Lp (8-9 MBq) I.V. Representative SPECT/CT imaging for the same mice over B) the first hour after I.V administration and C) 24 h later confirmed the selective accumulation of the liposomes into the ipsilateral side of the brain (right) compared to the contralateral side. Images shown here are for the same mouse from each group. The scale bar is the same in B and C. Healthy mice (where no ICH was induced) showed no detectable 111In-Lp level in the brain. The signal observed in healthy mice is due to the presence of 111In-Lp in the blood that is mainly https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4481 found in the circulation outside the brain. This reduced substantially 24 h after injection of 111In-Lp as they start to clear from the blood. D) Quantification of the 111In-Lp level in the brain 24 h after I.V injection revealed a biphasic entry pattern with maximum accumulation observed between 3-24 h and 48-72 h post-ICH. Values are expressed as % of I.D ± SEM per gram brain tissue. (E) Quantification of 111In-Lp levels in the CSF indicated no significant differences in CSF liposomal level compared to healthy mice. The data in D & E were analysed by one-way ANOVA followed by Tukey multiple comparison tests (n = 4-5). The data were considered significant if p values < 0.05. Figure 2: Evaluation of selective liposomal entry into the haemorrhagic brain shortly after I.V administration. A) Represents a schematic presentation of experimental design and the time frame of liposomes injection. Like the previous studies, all mice received a single I.V injection and each time point represents a separate group. B) Imaging DiI-Lp entry into the brain with IVIS Lumina II optical imaging system confirmed the selective accumulation of liposomes into the ipsilateral side (right) as early as 2 h after DiI-Lp I.V administration whereas no signal was observed on the contralateral side of the brain (left). Likewise, control groups which include naïve healthy mice injected with PBS and healthy mice injected with DiI-Lp showed no detectable signal. C) Quantification of the fluorescent signal of DiI-Lp in the brain by IVIS Lumina imaging software was performed by drawing a region of interest (ROI) that covers the whole brain and is expressed as total efficiency. Colour scale of epi-fluorescent signal range from min =2x10-4 (red) to max= https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4482 1 x10-3 (blue). Quantification data showed the significant entry of DiI-Lp into the brain only over the time frame from 3-5 h after ICH (D) to exclude any interference of tissue autofluorescence, control groups of ICH mice at a similar time point were imaged without injection of DiI-Lp which indicated no fluorescent signal in the brain. E) Histological analysis of brain tissues confirmed the presence of liposomes (yellow signal) that colocalised with the region of brain damage (represented by pale region with cresyl violet stain) and hematoma region as shown with H&E. IgG stain was performed to evaluate any endogenous IgG leakage into the brain (outline by dashed black lines) that is used as an indication of BBB disruption. Statistical analysis was performed using one-way analysis of variance followed by the Tukey multiple comparison test and p values < 0.05 were considered significant. n = 3-4 in each group. Figure 3: Analysis of delayed liposomal entry into the haemorrhagic brain 24 h after I.V administration. A) Schematic presentation of experimental design indicating the time frame of liposomes injection. Each time point represents a separate group of mice that received a single I.V injection. Analysis of liposomal entry into the brain was performed 24 h after I.V injection. B) IVIS Lumina II optical Imaging of DiI-Lp accumulation in the brain indicated that liposomes maintained their selective accumulation in the brain even 24 h after injection. Liposomal accumulation also revealed a biphasic entry pattern which was further confirmed by quantification of the fluorescent signal of DiI-Lp in the brain using https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4483 IVIS Lumina imaging software as indicated in C). No signal was observed on the contralateral side of the brain (left) and control groups which include naïve healthy mice injected with PBS and healthy mice injected with DiI-Lp. D) IVIS images of ICH brain tissues in the absence of DiI-Lp injection did not show any fluorescent signal in the brain. E) Representative optical images of the brain after injection with DiI-Lp at the indicated time points in sham mice. F) Quantification of DiI-Lp entry from IVIS Lumina II optical images confirmed that no significant increase in liposomes entry into the brain was observed in sham mice which indicated that selective accumulation of liposomes into the ipsilateral side (right) after ICH is due to brain injury induced by ICH rather than any brain damage induced by the procedure. G) Histological analysis of brain tissues confirmed biphasic entry of liposomes (yellow signal) into the brain that colocalised with the region of brain damage (represented by pale region with cresyl violet stain) and hematoma region as shown with H&E. IgG stain was performed to evaluate any endogenous IgG leakage into the brain (outline by dashed black lines) that is used as an indication of BBB disruption. Statistical analysis was performed using one-way analysis of variance followed by the Tukey multiple comparison test and p values < 0.05 were considered significant. n = 3-4 in each group. Liposomal brain entry co-localises with brain of liposomes increases the chances of extravasation damage, haematoma, and BBB disruption into the lesion through areas of BBB disruption which The overlap of liposomes accumulation detected is one of the advantages liposomes offer compared to from SPECT/CT and IVIS with the ipsilateral small drug molecules with short circulating t1/2. Our hemisphere was confirmed histologically on serial previous work using a similar type of liposomes in brain sections, which showed that the localisation of C57 mice indicated that more than 40 % of the ID is DiI-Lp signal is confined to the lesion site. Figure 2E still in circulation at 6 h after IV injection [21]. Also, shows the highest DiI-Lp accumulation in the brain previous studies with other types of liposomes when injected in the acute phase (3-5 h), staining is including drug-loaded liposomes (e.g. Doxil) restricted to the infarct region with little deviation indicated a circulation half-life of 1-2 days and an from the borderline of the haemorrhage. This increase in total drug level of up to 60 folds compared indicates that liposomes are transported at the site of to the free drug [28,29]. active bleeding through damaged vessels and DiI-Lp cerebral accumulation covered both the preferentially reside within the area of damage. ipsilateral cortex and striatum and maintained similar Cerebral accumulation detected from the 24-26 h distribution to the areas of BBB damage and infarct as group is minimal and restricted to the perihematomal seen with IgG infiltration and Cresyl violet stain region (rim area). At this time point, it can be seen that respectively (Figure 2E). Mice from all time points the area of infarct extends beyond the haemorrhage displayed altered neurological symptoms such as volume with oedema and mass effect major circling behaviours, gait and body posture that are contributors to the observed cellular damage. At 48-50 indicative of unilateral brain damage (Figure S2A&B). h liposomes are seen, in a similar patterned All mice showed an increase in neurological score and distribution to that observed in the 24-26 h group, at contralateral foot-slips measures. Histological the lesion rim, but also beyond the infarct region close measurements of haematoma volume (Figure S2C) to regions of BBB damage. The lack of liposomes correlated with changes in infarct volume (Figure within the core of the haemorrhage in the 24-26 h and S2D) and body weight (Figure S2E). Changes in 48-50 h groups likely relates to the absence of active haematoma volume and body weight peaked at 24 h bleeding and clot stabilisation at these timepoints. and started to reduce over time. Widespread accumulation of liposomes Liposomes do not penetrate brain tissue in the following injection in the acute phase (3-24 h group) is absence of pathology seen (Figure 3G) suggesting that they were once In healthy control animals, there was a lack of within the core, as shown from the 3-5 h group in fluorescent signal, indicating no DiI-Lp accumulation, Figure 2E, but that over time they disperse more both with IVIS optical imaging (Figure 2B & Figure widely. The distribution of liposomes injected 24 h 3B) and histologically (Figure 2E and Figure 3G). post-ICH did not change over time and showed Normal physiology of these animals is demonstrated restricted co-localisation to the lesion rim. with no cellular damage, absence of haemorrhage, Interestingly, a dramatic change in liposomes intact BBB and hemisphere of equal sizes indicating distribution was observed in the 48-72 h group with the absence of oedema. Imaging brain tissues from DiI-Lp staining detected within the core of the animals that had ICH induction but no DiI-Lp haemorrhage again, as seen in the 3-24 h group, as injection also showed no increase in fluorescent signal well as the lesion rim. This is unlikely due to (Figure 2D and Figure 3D), confirming that the secondary bleeding and may be linked to the increase in the fluorescent signal detected with IVIS is haematoma resolution process and infiltration of due to liposomal accumulation, and not auto-fluores- immune cells. cence from RBCs. Histological analysis of brain tissues Collectively, our data identified a biphasic entry of ICH mice at different time points after ICH in the profile of liposomes into the brain after ICH with the absence of DiI-Lp administration confirmed the lack highest accumulation observed 24 h after injection. of autofluorescence (yellow signal) in the injury site This indicated that the long blood circulation half-life due to the presence of RBCs in the brain (Figure S3). https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4484 MRI revealed two windows for BBB opening ICH lesions exhibited increased but stable T1 that correlate with the timing of selective compared to naïve mice liposomal entry into the brain Elevated spin-lattice relaxation time (T1) was Using dynamic contrast-enhanced MRI to observed in the ICH lesions (white columns) measure the BBB leakage rate of Gd-DOTA (Ktrans compared to normal tissue in the same location in [min-1]), a biomarker for BBB integrity, revealed an sham mice (grey shaded columns) indicating the identical biphasic entry pattern post-ICH to that seen presence of oedema (likely vasogenic) (Figure 4F). No with Dil-Lp (Figure 4B-C). The mean Gd-DOTA significant changes in T1 values were observed over leakage rate (Ktrans) tended to increase between 48-72 the time frame tested in our experiment (up to 72 h) h, but was not statistically significant. To investigate which suggested the persistence of brain oedema past further, we assessed only the areas of greatest BBB 72 h. Analysis of T2-TurboRARE images for the disruption, calculating the 95 % percentile of lesion changes in ipsilateral/contralateral ventricle size ratio Ktrans values (Figure 4D). In this analysis, leakage was in ICH mice showed no significant differences at this significantly higher at 72 h indicating that while the stage of the disease (Figure 4 G-H). average BBB integrity may be similar between 48-72 h, at 72 h certain focal areas of the lesion exhibit ICH lesions exhibit increases in free iron hyper-permeability. These observations were further content, T2*-weighed MRI signal intensity and confirmed in a larger cohort of mice (n=9-10) by decreases in RBCs over time measuring Ktrans values at 24 h and 72 h post-ICH Accumulation of iron products in the brain was (Figure 4E) which showed a significant increase in assessed by gradient-echo MRI (T2* signal). At 3 h mean Ktrans values confirming the second phase of BBB post-ICH, a significant reduction in T2* signal disruption. Given the similarity in the temporal and intensity was observed in ICH mice (white columns) spatial pattern of liposome and Gd-DOTA compared to sham mice (grey columns), likely accumulation in the brain, it is likely that their reflecting penetrance of iron-rich RBCs into brain penetrance is governed by a common entry process tissue from the initial bleed. By 48 h, T2* signal mediated by BBB disruption. These findings strongly intensity returned to the level observed in healthy suggest that leakage of both molecules into the brain tissue of naïve mice, indicating dispersion of RBC iron are mediated by self-diffusion across the disrupted into the lesion due to RBC lysis. BBB. The spatial distribution of T2* changes covered the entire lesion at 3 h and progressed to affect only Both transcellular transport and tight junction the perihematomal rim at 48-72 h (Figure 5A-B). This disassembly contribute to the entry of may reflect either i) preferential clearing of RBCs from liposomes into the brain after ICH the lesion centre leaving RBCs in the rim unaffected, To investigate the mechanism of liposomal entry ii) migration of RBCs to the peri-haematoma rim, or into the brain, IHC staining for Caveolin-1 (Cav1) and iii) accumulation of free-iron in the peri-haematoma Claudin-5 was performed to evaluate the contribution rim to levels that cause detectable T2* effects. of transcytosis transport and tight junction (TJ) To understand contributions to T2* signal disassembly respectively. A significant increase in intensity, histological analysis of RBCs and free iron Cav1 expression 3-5 h after ICH which suggests that deposition based on H&E and Prussian Blue stain was enhancement in transcytosis transport is contributing performed at three different coordinates from the to liposomal brain accumulation at this time. This Bregma; +0.14 (Figure 5C-D), -0.58 (Figure S6 A-C) agrees with our previous work in ischemic stroke that and -1.22 (Figure S6 B-D) both at the lesion core and showed enhanced Cav1 expression between 0.5-2.5 h perihematomal area (rim). Analysis of RBCs positive after the middle cerebral artery occlusion (MCAo) areas with H&E confirmed a significant decrease in model of stroke. In addition, we stained for Claudin-5 RBCs in the core over time, correlating with the to evaluate TJ remodelling after ICH, a significant progressive increase in T2* signal intensity. reduction in Claudin-5 being seen at 3-5 h and 48-50 h, Quantification of RBCs accumulation in the brain which suggests a contribution of paracellular assessed via H&E staining did not show significant transport to liposomal accumulation at both time change over time in the lesion rim (Figure 5D). points. Overall, our data (Figure S4-5) suggest Conversely, free iron as measured by histology using liposomal entry into the brain in ICH is mediated by; Prussian Blue stain increased with time in both the 1) an early phase of enhanced transcellular transport core and rim (Figure 5 C-D and Figure S6), supporting and TJ disassembly, and (2), a delayed phase (48 h the progressive release of iron from RBCs as a result after ICH) of TJ disassembly. of haemolysis, and possible migration of free iron from the centre to the edge of the lesion. Lack of https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4485 correlation of free-iron changes in the core with T2* area. These changes in iron content and haematoma MRI likely reflects the dominant effect of RBC (bound resolution could have affected liposomal transport iron) towards T2* signal loss. At 72 h, T2* signal loss after leakage into the brain and may explain some was prevalent only in the rim where free iron was also variation in liposomal brain accumulation observed found to be high. These findings may reflect highly between timepoints i.e the changes in redistribution of concentrated free iron or potential rebleeding in this liposomes into the core in the second phase 48-72 h. Figure 4: Characterisation of the blood-brain barrier and tissue changes in the haemorrhaged area by MRI at 3 h, 24 h, 48 h, and 72 h post-ICH. A) Schematic presentation of experimental design showing the time points of repeated MRI scans following ICH. 0.5 M Gd-DOTA was delivered by intravenous injection through a tail vein cannula at a volume https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4486 of 3.3 µL / g at a rate of 0.25 mL / min. B) Assessment of BBB leakage was performed by measuring Gd-DOTA leakage into the brain showing similar temporal and spatial patterns to liposome accumulation, suggesting a common entry process mediated by BBB disruption. C, D & E) measurements of Gd-DOTA leakage rate (Ktrans) acquired in 2 separate cohorts (cohort 1: C, D; cohort 2: E) indicated biphasic enhancement in BBB leakage. F) Repeated measurements of spin-lattice relaxation time (T1) indicated elevated values in the ICH model (white columns) compared to sham mice (grey shaded columns). No significant changes in T1 values were observed over time post-ICH. G) analysis of T2-TurboRARE MRI for the changes in ipsilateral/contralateral ventricle ratios showed no significant differences which indicates that no brain atrophy occurs at this stage of the disease which can also be seen on H) T2 Turbo rare MR images taken at different time points after ICH and sham surgeries. Statistical analysis was performed using a one-way analysis of variance followed by the Tukey multiple comparison test and p values < 0.05 were considered significant. n = 4-7 in each group. Figure 5: A) Accumulation of iron in the brain was assessed by T2* signal intensity, H&E and Prussian Blue staining. B) Our data showed a sharp reduction in T2* signal intensity at 3 h, indicative of deposits of blood-products (heme iron) within the lesion (white columns) compared to sham mice (grey columns), likely reflecting penetrance of iron-rich RBCs into brain tissue from the initial bleed. By 48 h, T2* signal intensity increased back to the level observed in the healthy tissue of naïve mice, indicating clearance of blood products from the brain. These results were confirmed by histology using H&E and Prussian Blue stain in the lesion core (C) and lesion rim (D). At each brain section measurements of positive RBCs, areas and positive free iron stain were done on four rim regions and two core regions. H&E images indicated RBCs positive area decreased over time in the lesion core while the values are not significantly changed in the lesion rim. Prussian Blue stain on the other hand indicated a continuous rise in free iron as a result of RBCs haemolysis reaching significant levels around 72 h after ICH. Values are expressed as mean ± SEM. Statistical analysis was performed using one-way analysis of variance followed by the Tukey multiple comparison test and p values < 0.05 were considered significant. Microglial activation is a dynamic process pathology, thus we next investigated if liposomes following ICH and co-localised with co-localised with those cells after access into the brain. Confocal images showed that microglia are ramified DiI-liposomal accumulation illustrating their transition into reactive, phagocytic Microglia and astrocytes play a key role in ICH and amoeboid cells which were observed both at the https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4487 lesion core and periphery (Figure S7). Astrocytes, timing of BBB disruption between the two conditions, identified as glial fibrillary acidic protein (GFAP) +ve the underlying pathological mechanism is different. area, were observed to be most evidently activated at In ischaemic stroke, the lack of oxygen and glucose the peripheries of the lesion, indicating reactive supply to the brain initiate the acute stage of BBB gliosis. Similar to our previous experimental design, disruption whereas delayed BBB damage is associated DiI-Lp were injected at three different time points with reperfusion injury and infiltration of after ICH (3, 24 and 48 h) and representative confocal inflammatory cells to the brain [19]. In ICH, the images were taken 24 h after DiI-Lp injection. leakage of blood components into the brain activates Quantification of DiI-Lp co-localisations with Iba1 pathways related to cell injury and inflammation +ve cells (activated microglial marker) indicated that which has a major role in BBB disruption. The role of maximum uptake of DiI-Lp by microglia was cerebral ischaemia in ICH injury is not yet conclusive. observed when injected at 3 h (44 %) and 48 h (32 %) Perihematomal blood flow measurements in animal after ICH, which was significantly higher than that and human studies are not significantly low to trigger observed when injected 24 h post-ICH (9 %). These ischaemic brain injury [30–32]. The differences in the data correlate with biphasic liposomal uptake into mechanism of BBB disruption between the two ICH lesions and suggested two windows for optimum models may also explain the higher extent of microglial targeting to shift or direct their activation. liposomal entry into the brain we observed here in Unlike uptake by microglia, only limited uptake of ICH, which is 10 times higher than what we reported DiI-Lp was detected in astrocytes at all time points in ischaemic stroke [20], offering the potential for (Figure S7 and S8). enhanced drug delivery in this, particularly devastating stroke subtype. Discussion The onset of BBB disruption in ICH reported in Recently we have demonstrated in an previous studies is rather variable depending on the experimental ischaemic stroke model that liposomes animal model. Induction of ICH by injecting blood can infiltrate BBB at the ischaemic site taking into the brain showed no early BBB disruption at 4 h, advantage of enhanced transcytosis and tight junction but BBB damage was observed at 1-2 days [33]. disassembly [20]. Liposomes also have great potential Inducing ICH by collagenase injection resulted in BBB for targeted drug delivery to damaged tissue disruption within 30 minutes followed by a following ICH. Here we aimed to confirm the prolonged, but lower level of BBB disruption that was penetrance of liposomes in ICH, possible underlying shown to last for 7 days [18] which agrees with the mechanisms by which liposomes enter the brain, and biphasic BBB disruption we observed. Another study to evaluate the optimal timepoint for drug delivery. with the collagenase ICH model showed a continuous We have demonstrated that liposomes accumu- increase in BBB disruption over time that peaks late in the lesion site after ICH and maintain long term around 3-4 days before returning to normal level at co-localisation in the injured area. A biphasic entry 7-10 days [34]. However, that study did not show BBB pattern was identified, indicating 3 h and 48 h permeability measurements in the first few hours post-ICH as the optimum windows for therapeutic after ICH which could explain the absence of biphasic intervention. We also observed that liposomal BBB disruption. In humans, ICH is reported to cause distribution in the brain varies over time and is likely both early [35] and delayed vascular permeability [36] linked to haematoma resolution. Activated microglia and, in some patients, continuous bleeding and uptake the liposomes at both early and late time extravasation of contrast agent were observed in the points which could be related to enhanced phagocytic first 24 h post-injury [35,37]. Therefore, the findings in capacity after ICH. By this, we have demonstrated the this study confirm that the collagenase model ability of a clinically approved nanomedicine recapitulates the clinical picture of BBB disruption. formulation to cross the BBB in a rodent model of The detection of endogenous IgG infiltration into the ICH, providing promising data for future therapeutic lesion site is frequently used to evaluate BBB studies. disruption as indicated in our findings. While this Our data showed that the biphasic entry pattern method is adequate in concluding the presence of BBB of the liposomes into the brain correlates with leakage damage, it is not possible to determine the exact of Gd-DOTA, as detected using DCE-MR. These timings at which the damage occurred. Therefore, in observations are very similar to our previous findings this study, it was essential to correlate our in ischaemic stroke which also showed two windows observations with serial in-vivo MRI to determine the of liposomal entry into the brain that peaks around 3 h time-course of BBB disruption post-ICH. In our and 48 h after brain ischaemia [20]. Although these previous study, investigating the mechanism of observations suggested that there is an overlap in the liposomal entry into the brain after ischaemic stroke, https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4488 we demonstrated that liposomes access to the brain is start within minutes after bleeding and lasts for days mediated by enhanced caveolae-mediated afterwards [45]. Our data showed a sharp reduction in transcytosis in the acute phase (0.5-2.5 h after stroke) T2* signal intensity at 3 h, indicative of deposits of when no damage to the TJ is detected [20]. In this blood-products (heme-iron) within the lesion. This study, we observed that Cav1 is also significantly appears to resolve by 48 and 72 h post-ICH, with T2* enhanced 3-5 h after ICH which suggests that signal returning to normal first in the core, and last in liposomes transport across a disrupted BBB is linked the lesion rim. This suggests that the second phase of to caveolae-mediated transport. Our data also suggest BBB hyper-permeability at 72 h does not result in the contribution of TJ remodelling to liposomal entry significant re-bleeding of RBCs into the lesion. Free into the brain both at 3-5 h and 48-50 h after ICH as iron was found to accumulate over time, likely the indicated by the reduction in claudin-5 positive areas. result of RBC lysis. Taken together, these findings Previous studies in rodent models of ICH reported a suggest that in the context of ICH, T2* signals are reduction in various TJ proteins 1-3 days after ICH dominated by iron within RBCs. We hypothesize that [38–41] using immunohistochemistry, western blot or as these cells rupture and release iron into the mRNA expression. Similarly, a clinical study in ICH surrounding area, the iron disperses resulting in a patients showed elevated TJ protein levels in CSF reduction in susceptibility gradients and an increase samples measured on admission which indicate a role in T2*. Much higher levels of free-iron were observed in TJ remodelling in BBB dysfunction very early after in the lesion rim (2-3 fold) compared to the lesion ICH [42]. core. This indicates that the rim may also contain a In the present study, we assessed BBB disruption higher density of RBCs (the source of free iron), or is a and liposomal entry into the brain up to 72 h site of significant free iron accumulation, which is post-ICH, however, other studies in animals [34,43] supported by Figure 5D and Figure S6 C and D. Iron and human [36] reported that BBB disruption lasts and heme-iron lead to the induction of reactive over a longer period before reaching baseline level. oxygen species which can lead to an iron-dependent This suggests that liposomal entry into the brain could cell death known as ferroptosis [46,47]. Although the be extended beyond the 72 h tested in this study. concept is still evolving, ferroptosis is now considered Further research will be needed to confirm the exact a unique form of regulated cell death characterized by mechanism of entry and whether this would change the cytosolic accumulation of iron and executed by the over time given the variation in the underlying massive induction of lipid ROS that leads to pathological events associated with ICH. intracellular oxidative stress and damage to nucleic In this study, we have used serial measurements acid, proteins, lipid membrane and consequently cell of spin-lattice relaxation time (T1) to evaluate brain death [48]. water content (oedema) after ICH [44]. Our data Previous studies in animal models of ICH indicated elevated levels of oedema detected in the showed that efficient haematoma clearance is usually ICH model compared to naive mice. Using a delayed process that starts 3 days post ICH which T2-weighted MRI and apparent diffusion coefficient correlates with the peak of microglial activation and is (ADC) measurement Yang et al showed an increase in completely resolved around 3-4 weeks [34,49,50]. This perihematomal oedema after ICH reaching a peak agrees with our data that showed no significant level at day 3 followed by a gradual decrease over 7 reduction in RBCs deposition area, haematoma days [34]. However, no earlier measurements were volume and infarct volume up to 72 h post-ICH. performed in that study. In agreement with our Microglia are activated within minutes after ICH findings that study also showed diffuse microglial onset and play a key neuroprotective role by clearing activation that covers both the lesion core and blood products, apoptotic cells and cellular debris periphery, while astrocytes activation was observed through phagocytosis [50,51]. This process has an in the perihematomal area. Free iron deposition was important role in providing a nurturing environment also observed mainly in the lesion rim at day 3 to promote tissue recovery, however excessive indicating the start of RBCs lysis. These data microglial activation and infiltration of peripheral collectively suggested that activated glial cells and inflammatory cells can also exacerbate ICH injury RBCs lysis at the lesion periphery may play a role in through the release of proinflammatory mediators increasing BBB permeability and vasogenic oedema in [50]. CD36, a member of the class B scavenger receptor this region. family, is an important mediator of phagocytosis and The rupture of the microvascular wall in ICH it is highly expressed in microglia and mediates the exposes neurons, astrocytes, microglia and other cells phagocytosis of apoptotic cells, bacteria and around the haemorrhagic area to blood-derived iron, low-density lipoproteins [52,53]. Previous studies either bound or free. Erythrocyte lysis is reported to suggested that CD36 has a critical role in controlling https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4489 microglial phagocytic activity. The expression of improved functional recovery compared to the free CD36 in microglia is transcriptional regulated via the drug [62]. Another approach using polylactic- activation of peroxisome proliferator-activated co-glycolic acid nanoparticles of human recombinant receptor (PPAR) which accelerates their RBCs erythropoietin decreased the rate of lethal events and phagocytic capacity and haematoma resolution [54]. improved neurological outcomes when tested in a Treatment with PPAR agonist (e.g. rosiglitazone) traumatic intracerebral haemorrhage model [63]. showed enhanced RBCs clearance by increasing CD36 Other strategies that have been widely expression and phagocytosis both in vitro and in vivo investigated to enhance brain prenetration include and in turn promoting hematoma resolution [10]. active targeting. In this case, a high-affinity ligand PPAR activator also modulates the proinflammatory against an influx system or highly expressed receptors mediator expression to reduce secondary brain on the brain endothelial cells is utilised to shuttle damage and protect other brain cells from oxidative molecules across the BBB irrespective of the damage imposed by microglia during phagocytosis. alterations of BBB integrity. Transferrin receptor Previous studies showed that Toll-like receptor (TLR) (TfR), insulin receptor and glutathione (GSH) may play a role in regulating CD36 receptor function transporter are some examples of receptors present on both in phagocytosis and inflammation. Upregulation the BBB that have been utilised to enhance drug brain of CD36 expression was demonstrated in TLR-4 penetration. Using this strategy in mice to deliver knockout mice suggested that TLR-4 suppression erythropoietin [64] or type II tumour necrosis factor-α could promote hematoma resolution [49,55]. The by fusion with TfR monoclonal antibody resulted in enhanced uptake of liposomes observed in microglia high brain accumulation reaching 2-3 % of ID / g compared to astrocytes could be linked to activated tissue [65]. This enhanced brain penetration is in the phagocytosis in microglia due to their critical role in same range observed in our ICH model where we RBCs clearance and haematoma resolution. demonstrated 2-3 % of ID / g brain tissue of liposomal Since microglia are activated as early as 1 h after accumulation 3-24 h after ICH. This is also ICH and remain activated for the first few days, the comparable to what was reported before from highly difference in liposomal uptake after ICH reflects the lipophilic molecules such as diazepam reaching 1.8% differences in liposomal brain entry levels showing a of ID / g brain tissue within 10 minutes of I.V dual uptake pattern. Given the biphasic therapeutic injection [66]. In addition Targeting receptors at the windows observed with liposomes, a similar dual BBB improves the brain penetration of therapeutic approach can be designed to enhance immunoliposomes and their encapsulated drug into haematoma clearance and minimise the deleterious the brain. From those TfR and GSH targeted proinflammatory responses that could trigger further immunoliposomes have been widely investigated for brain damage. brain tumours [67,68], and multiple sclerosis [69,70]. The critical role of activated microglia in Using immunoliposomes delivery method increased haematoma resolution and their localisation both in drug accumulation in the brain by 3-4 folds was the lesion core and periphery may also explain the observed compared to the free drug reaching up to difference in liposomal localisation in the brain 0.06 % of ID / g tissue in some examples [71]. parenchyma observed 48-72 h after ICH. Activation of An exciting observation in this study when MMPs that are involved in extracellular matrix considering clinical translation is the extended remodelling is also peaked 24-72 h after ICH [51,56]. therapeutic window provided by the significant Their critical role in delayed BBB disruption and liposomal accumulation when injected in the acute neurovascular remodelling suggested that they could phase (3 h post-ICH) and the subacute phase (48 h be linked with the alterations in liposomal post-ICH). The broad treatment timing provided by distribution over time [57–59]. liposomes means a larger subset of patients can The benefits of nanomedicine-based drug benefit and the choice of therapeutic molecules could delivery in stroke research have been demonstrated be diverse based on the underlying pathological recently, however, these have been largely restricted mechanisms dominating at that time. The develop- to ischaemic stroke models [20,60,61]. No liposomal ment of liposomal drug delivery systems is versatile formulation has been tested before in ICH, but very in terms of treatment choice and has already limited studies with other types of nanoparticles demonstrated promising results in preclinical models showed promising results in haemorrhagic stroke. of ischaemic stroke models [20,60,61]. The safety One study tested the therapeutic effect of profile of liposomes is already established through its nanoemulsion formulation of quercetin in a current licenced use in other pathologies [72] and thus collagenase rat model and demonstrated superior holds the best chances of being fast-tracked to the antioxidant effect, reduce haematoma volume and clinic. Modifying liposomes to scavenge iron or https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4490 modulating microglial activity towards enhanced Mice and diets haematoma resolution may enable therapies the will We have used C57BL/6 male mice (11-12 enable the reduction of these pathologies, the effect of week-old, weighing 25-30 g) provided by Charles which could be tracked using MRI. Moreover, this River, United Kingdom. The mice were housed in study will provide a potential direction for the groups of 4–5 with free access to diet and water and at development of liposome-based imaging contrasts for a constant ambient temperature of 21 ± 2 °C and the ICH model in the future in line with other humidity of 40–50 %, on a 12 h light, 12 h dark cycle. clinically relevant fluorescent imaging applications Experiments involving animals were all performed [73–75]. according to the United Kingdom Animals (Scientific Procedures) Act, 1986 (P28AA2253) and the work is Conclusion already approved by the Home Office and the local This study provides evidence that liposomes can Animal Ethical Review Group, University of selectively accumulate within the lesion site in a Manchester and reported in compliance with the collagenase model of ICH. Liposomal entry into the ARRIVE guidelines. Most of the studies conducted brain peaked when injected 3 h and 48 h after ICH and included 4-5 animals per group and analysis was correlated with the changes in leakage of Gd-DOTA performed blinded to the treatment conditions. into the lesion detected using DCE-MRI, indicating the primary mechanism of penetrance was via Induction of Intracerebral haemorrhage (ICH) self-diffusion across the disrupted BBB. Liposomal Before induction of ICH, mice were acclimatised accumulation observed is very specific to the to mash and had their bodyweight assessed. To haemorrhage area and showed 40 % co-localisation induce ICH, mice were anaesthetised with a mixture with activated microglia at the lesion site. These of 4% isoflurane in a mixture of 30 % oxygen and 70 % observations offer an extremely promising technology nitrous oxide. to shave the region between eyes and for ICH treatment enabling time-specific new ears, before being transferred to a feedback-controlled therapies and spur the rejuvenation of drugs that have heating pad where the anaesthesia was maintained at been tested before in ICH but have been missed due to 1.5-2 % of isoflurane using a nose cone. A longitudinal failed efficacy or unwanted peripheral toxicity. midline incision through the skin was made and soft tissue retracted. During the surgical procedure, core Materials and Methods body temperature was monitored using a rectal probe and maintained at 37 ± 0.5 °C, using a thermal blanket. Materials A burr-hole was created at the following coordinates; The lipids used in this study including bregma 0, lateral -2.0 and dorsoventral -2.5 using an hydrogenated soy phosphatidylcholine (HSPC) and electric hand drill. A glass micropipette was used to 2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[m inject 0.03 U collagenase VII S, dissolved in 0.5 µl ethoxy(polyethylene glycol)-2000] (DSPE-PEG2000) sterile saline, at a rate of 1 µl / min. The needle was were a kind gift from Lipoid GmbH (Ludwigshafen, left in place for 10 minutes before a gradual Germany). 18:0 PE-DTPA 1,2-distearoyl-sn-glycero-3- withdrawal over a period of 60 seconds. The wound phosphoethanolamine-N-diethylenetriaminepentaace was then closed with nonabsorbable sutures. Before tic acid (ammonium salt) was from Avanti Polar recovery, all mice were given saline (1 ml / 100 g Lipids (USA). Chloroform and methanol were body weight) and buprenorphine (0.05 mg / kg S.C). purchased from Fisher Scientific. Phosphate buffer Following the surgery, the mice were then housed in a saline, cholesterol, paraformaldehyde, sucrose, goat cage, with a choice of both mash and pelleted food, in serum, collagenase and cresyl violet were purchased a heated cabinet (~25 °C). Mice were monitored until from Sigma. Polycarbonate extrusion filters they had recovered from the anaesthetic and were (Whatman) 800nm, 200nm, and 100nm were from completely conscious and re-gained their righting VWR, UK. PD-10 desalting columns were bought reflex. After surgery, mice were weighed every day from GE-Healthcare Life Sciences. 1,1'-Dioctadecyl- and assessed for their general well-being. Bodyweight 3,3,3',3'-Tetramethylindocarbocyanine Perchlorate data were presented as a % weight change compared (DiI) was purchased from Invitrogen Detection with bodyweight on the day of surgery. Assessment Technologies. Biotinylated anti-mouse IgG, NHS of neurological function following ICH was blocking solution, DAB peroxidase (HRP) substrate performed with the 28-point neurological scoring kit and Vectastain Elite ABC HRP Kit were all system [20]. purchased from Vector Laboratories. https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4491 Assessment of brain damage with cresyl violet imaging and gamma counting. Liposomes were staining radiolabelled with radioactive indium (111In). Briefly, Brain sections were stained with cresyl violet 25 mM (total lipid concentration) of HSPC: and the infarct volume was calculated by measuring Chol:DSPE-PEG2000:PE-DTPA 56.3:38.2:5.5:1 mol/mol the areas of neuronal loss at eight defined coronal % liposomes were prepared as described above using levels as previously described [20]. On each section, the thin-film hydration method. Hydration of the the area of damage was measured using ImageJ (NIH, lipid film was done with freshly prepared ammonium Bethesda, MD, USA), adjusted for oedema and the acetate buffer (0.095 M, pH 5.5) at 60°C followed by volume of damage was calculated by integrating of extrusion to reduce the size of the liposomes. areas of damage with the distance between coronal Subsequently, liposomes were radiolabelled by 1 h levels using GraphPad Prism 7, Software. The volume incubation with radioactive 111InCl3 (11 MBq / 2.5 of damage was expressed as the total amount of µmol lipids) in 0.2 M ammonium acetate pH 5.5. ischaemic damage. Incubation was carried out at room temperature with continuous vortexing every 5 minutes. At the end of Haematoxylin and Eosin staining incubation 0.1 M, EDTA (1/20 of the total volume) For the assessment of haemorrhagic regions, was added to chelate any free 111In. To determine the haematoxylin and eosin (H&E) staining was radiolabelling efficiency, any unbound 111In and 111In-EDTA were removed with PD-10 column performed. Frozen sections 20 µm thick were first fixed in formal alcohol for 20 seconds and rinsed pre-equilibrated with HBS pH 7.4. Aliquots of each quickly in distilled water. Following that we stained final product were diluted five folds in PBS and then 1 in haematoxylin for 2 minutes and washed in Scott’s μl was spotted on silica gel impregnated glass fibre tap water. The slides were then stained in eosin for 10 sheets (PALL Life Sciences, UK). The strips were seconds, and rinsed in tap water, before being developed with a mobile phase of 50 mM EDTA in 0.1 gradually dehydrated in alcohol (70-100 %) in three M ammonium acetate and allowed to dry before steps. Once the slides are dehydrated, they were analysis. This was then developed and the autoradio- cleaned twice in Xylene each for 1 minute before being activity was quantitatively counted using a cyclone cover slipped using a resinous mounting medium. phosphor detector (Packard Biosciences, UK). The Slides were scanned using a slide scanner (3D Histec immobile spot on the TLC strips indicated the Pannoramic 250) and images were captured using percentage of radiolabelled 111In-Lp, while free 111In Pannoramic Viewer. The area of red blood cells was was detected as the mobile spots near the solvent measured in the same way as the infarct volume was front. Very minimum free 111In was detected to yield calculated and compared between the groups. radiolabelling efficiency of > 85 % as displayed in Figure S1. The radiolabelling stabilities of the final Preparation of DiI-labelled liposomes (DiI-Lp) product of 111In-Lp were studied after five times DiI-labelled liposomes composed of HSPC:Chol: dilution in both 50 % serum or PBS and then DSPE-PEG2000 56.3:38.2:5.5 mol/mol % were prepared incubated at 37 °C up to 48 h. At different time points by thin-film hydration method followed by extrusion (0, 1 and 24 h), 1 μl of the aliquots was spotted on [20]. Briefly, lipids dissolved in chloroform: methanol silica gel impregnated glass fibre sheets and then mixture (4:1) were mixed in a round bottom flask and developed and quantified as described above. 5mol% of DiI in ethanol (1 mg / ml) was added to the lipid mixture. Organic solvents were then evaporated Liposomes characterization to produce the lipid film. Lipid films were kept Liposome size and surface charge were protected from light and hydration was performed measured using Zetasizer Nano ZS (Malvern, with HBS (20 mM HEPES, 150 mM NaCl, pH 7.4) to a Instruments, UK). Samples were diluted 100 times final lipid concentration of 12.5 mM. To produce small with purified distilled water before measurements. unilamellar liposomes, the size was reduced by Triplicates measurements were recorded, and the data extrusion through 800 nm and 200 nm polycarbonate were expressed as average ± S.D. The fluorescence filters 5 times each then 20-40 times through 100 nm intensity of DiI-Lp was recorded using Carry Eclipse membranes using a mini-Extruder (Avanti Polar fluorescence spectrophotometer, (Agilent Techno- Lipids, Alabaster, AL). logy). Samples were first diluted 200 times in HBS and recorded at 518 nm / 565 nm excitation/emission Preparation and characterisation of wavelengths (slit 5 / 10). 111In-labelled liposomes (111In-Lp) To evaluate the accumulation of liposomes in the brain after ICH, we have used real-time SPECT/CT https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4492 Assessment of DiI-Lp accumulation into the dilution of the injected dose. The results were brain with optical imaging represented as the percentage of the injected dose (% To confirm the accumulation of the liposomes in ID / gm tissue ± SEM), n = 4-5 mice per group. the brain, ex vivo optical imaging was used by Assessment of BBB permeability to IgG detecting the optical fluorescence signal of the To assess BBB permeability, we have stained for liposomes. IVIS imaging was performed shortly after endogenous IgG infiltration into the brain was cardiac perfusion with ice-cold 0.9 % saline followed visualised using peroxidase-based immunohisto- by 4 % PFA to remove any DiI-Lp that are still chemistry. Serial brain sections (20 µm thick) were circulating in the blood. DiI-Lp were administered at mounted on slides and washed 3 times with PBS. the following time point after ICH (3 h, 24 h and 48 h) Following that endogenous peroxidase activity and by I.V and the brain tissues were extracted either 2 h non-specific staining were blocked by 10 min or 24 h after I.V administration. Imaging was incubation in 0.3 % H2O2 and then washed (3 times in performed with the IVIS Lumina II imaging system PBS, 10 min each). blocking was performed with 10 % (Caliper Life Sciences Corp., Alameda, CA) at 535 nm normal horse serum (NHS) in 0.3 % Triton X-100 PBS / DsRed excitation and emission filters with 0.5 sec (PBST) for 1h before overnight incubation with exposure. DiI-Lp total fluorescence intensity in the biotinylated anti-mouse IgG (1:250 in 0.3 % PBST, brain was quantified by drawing a region of interest Vector Laboratories) at 4°C. Sections were then (ROI) that covers the whole brain and is expressed as incubated with avidin-biotin-peroxidase complex, total efficiency. and colour-developed using freshly prepared Single-photon emission computed diaminobenzidine (DAB) solution. To ensure tomography (SPECT/CT) comparable DAB staining between sections is Mice were subjected to anaesthesia via the achieved, the time of the colour change was recorded, inhalation of 2.5 % isoflurane in a mixture of 30 % and DAB was applied for each subsequent sample for oxygen and 70 % nitrous oxide. Each animal was then the same amount of time. intravenously injected with 200 ul of the radioactive Immunohistochemistry staining for Cav1 and 111In-Lp (8-9 MBq). At different time points after CD31 injection (t= 0-1 h, & 24 h) SPECT/CT imaging was Free-floating serial brain sections (20 µm thick) carried out using a Nano-Scan® SPECT/CT scanner were washed 3 times in PBS for 10 min and blocked (Mediso, Hungary). SPECT images were obtained in for 1 h in 10 % normal goat serum (NGS) in 0.3 % PBS 20 projections over 40-60 min using a 4-head scanner triton (PBST). This was followed by overnight with 1.4 mm pinhole collimators. CT scans were taken incubation with primary antibodies in 2 % NGS in at the end of each SPECT acquisition using a PBST at 4 °C. The primary antibodies used are; rat semi-circular method with a full scan, 480 projections, anti-mouse CD31 (BD Pharmingen, 550274, 1:100) and maximum FOV, 35 kV energy, 300 ms exposure time mouse anti-mouse Cav1 (BD Biosciences, 610407, and 1-4 binning. Acquisitions were done using the 1:50). After incubation with the primary antibodies, Nucline v2.01 (Build 020.0000) software (Mediso, sections were washed 3 times in PBS and incubated Hungary), while the reconstruction of all images and with fluorescently labelled secondary antibodies in 2 fusion of SPECT with CT images was performed % NGS in PBST. To visualise the primary antibodies, using the Interview™ FUSION bulletin software the following secondary antibodies were used; goat (Mediso, Hungary). The images were further analysed anti-mouse Alexa Fluor® 488 conjugate (Invitrogen using VivoQuant 3.0 software (Boston, US) where the A11001, 1:500), goat anti-Rat Alexa Fluor® 647 SPECT images with scale bars in MBq were corrected conjugate (Invitrogen, A-21247, 1:200). In the end, for decay and the slight differences in radioactivity in samples were washed 3 times in PBS and transferred the injected doses between animals. For a quantitative on non-gelatine coated slides and left to dry overnight assessment of 111In-Lp in the brain, a cut and count before slides were then coverslip with ProLong Gold method was used. Mice were anaesthetized by Antifade Mountant (Thermo Fischer Scientific, Inc., isofluorane inhalation and each mouse was injected USA). Images were collected on either SP5 inverted via the tail vein with 200 μl containing 111In-Lp microscope using 40x objective. labelled with approximately 8-9 MBq. 24 h after Brain sections injected with DiI-Lp at 3 h and 48 injection, mice were perfused with iced cold saline h were used for this study and analysed 24 h after (0.9 %) followed by PFA (4 %) to remove any 111In-Lp liposomal administration. The ipsilateral ICH lesion from the blood before brain tissues were collected. site and the corresponding areas on the contralateral Each sample was weighed and counted on a gamma slide were analysed. For each brain region, 4 discrete counter (Perkin Elmer, USA), together with a serial https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4493 areas were imaged at a magnification of 40 x. Each being rehydrated in PBS for 10 min. Following that area was analysed in ImageJ by applying fluorescence the slides were quickly washed in tap water and left to intensity threshold and the number of pixels from dry vertically at 37 ˚C for 1 h. For Iba-1 staining only, each colour (black, red, green, yellow, blue, magenta) antigen retrieval was first performed. For this, the were then automatically measured as previously slides were placed inside a closed glass Coplin jar described [76]. These values enabled us to ascertain filled with the EDTA solution (pH 6.0). The jar was the area of immunolabeling from each cell type and sealed with parafilm and placed inside a 500 ml glass their co-localisation with the liposomes. Data beaker filled with boiling water and placed on a hot expressed as average ±STD. plate for 10 min. After boiling the slides were To complement confocal images, widefield immediately rinsed with ultrapure Type I water at overview images were obtained at x40 magnification room temperature for 5 min followed by two without oil utilising Leica Thunder imager successive rinses in ultrapure Type I water for an microscope. All tissue areas were processed under additional 5 min. The slides were then blocked for 1 h identical conditions to demonstrate liposomes (red), at room temperature in 5 % goat serum solution CD31 (blue) and caveolin-1 (green) immunoreactivity. prepared in 1 % BSA, 0.1 % Triton X-100 (as After initial imaging concluded, mosaic merge penetration enhancer) and 0.05 % Tween 20 (used processing via Leica application suite X v3.7.4.23463 detergent and surface tension reducer). This was combined tiles into one image and large volume followed by overnight incubation with primary computational clearing, such as lightning and thunder antibodies in 2 % normal goat serum, 1 % BSA and 0.3 programme, reduced out of focus areas to produce % Triton X-100 in PBST at 4 °C. The primary high quality images. antibodies used in the study are explained in detail below; Chicken antiGFAP (Abcam AB4674, 1:200), Immunohistochemistry staining for claudin-5 Rabbit antiIb1a (Wako 019-19741, 1:100). After Serial frozen sections of brain tissues (20 µm in incubation with the primary antibodies sections were thickness) were cut on a cryostat, mounted on slides washed 3 times in 0.1 % tween solution PBS and and allowed to warm to RT for 20 minutes before incubated with fluorescently labelled secondary being rehydrated in PBS for 10 min. Following that antibodies prepared in 0.1 % PBS tween solution. To the slides were quickly washed in tap water and left to visualise the primary antibodies, the following dry vertically at 37 ˚C for 1 h. The slides were then secondary antibodies were used; goat anti-Ck Alexa blocked for 1 h at room temperature in 5 % goat serum Fluor® 488 conjugate (Invitrogen A11039, 1:500), goat solution prepared in 1 % BSA, 0.1 % Triton X-100 (as anti-Rabbit Alexa Fluor® 647 conjugate (Invitrogen, penetration enhancer) and 0.05 % Tween 20 (used A21244 1:200). In the end, samples were washed 3 detergent and surface tension reducer). This was times in 0.1 % PBS tween and left to dry overnight followed by overnight incubation with rabbit before slides were then coverslip with ProLong Gold anti-mouse Claudin-5 in 2 % Donkey serum in 1% Antifade Mountant with DAPI (Thermo Fischer Bovine serum albumin, 0.05 % Tween 20, 0.3 % Triton Scientific, Inc., USA). Images were collected on an X-100, 1:100 (Abcam, Cambridge). After incubation inverted microscope using either 40x or 63x objective with the primary antibodies sections were washed 3 in the striatum and the outer cortex at bregma -0.58 times in 0.1 % tween solution PBS and incubated with mm. donkey anti-rabbit Alex 488, 0.1 % Tween 20 1x PBS, Analysis of liposomal uptake in microglia and 1:200 (Invitrogen, USA). In the end, samples were astrocytes was done by Imaris software based on washed 3 times in 0.1 % PBS tween and left to dry Z-stack confocal images reconstructed into 3D images overnight before slides were then coverslip with using a method previously described [77,78]. Using ProLong Gold Antifade Mountant with DAPI Imaris software a mask is first created over each (Thermo Fischer Scientific, Inc., USA). Images were channel corresponding to liposomes, microglia and collected on an inverted SP5 confocal microscope astrocytes. Figure S7 showing a representative images using 63x objective in the striatum. Analysis of the of liposomes (red) colocalised mainly with microglia images was performed in the same way described (cyan), while limited interaction is observed with above for Cav1 staining. astrocytes (green). All tissue samples are processed using the same acquisition settings. The Immunohistochemistry staining for astrocytes co-localisation function in Imaris is used to determine and microglia the fraction of liposomes internalised in microglia and Serial frozen sections of brain tissues (20 µm in astrocytes. thickness) were cut on a cryostat, mounted on slides and allowed to warm to RT for 20 minutes before https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4494 Perls Prussian blue staining (Iron staining) angle = 60°, matrix size = 96 x 96 x 64, FOV = 16 x 16 x To assess for the presence of free iron in the brain 24 mm. Details of MRI scan time points for each which indicates RBCs haemolysis, Perls Prussian blue mouse are shown in Supplementary Table 1. stain was performed. First, the slides were MRI analysis acclimatised to room temperature for 20 minutes and Ventricle and lesion regions of interest (ROIs) then hydrated in distilled water. The slides were then were manually delineated on the T2-Turbo RARE immersed for 20 minutes in a fresh mixture of equal images using MRIcro (version 1.0.2). Ventricle parts of 20 % hydrochloric acid and potassium volumes were calculated for each mouse by ferrocyanide. The slides were then washed in distilled multiplying the number of voxels in each ROI by the water three times and then counterstained with voxel volume. T1 maps were generated by fitting the nuclear fast red for 5 minutes and rinsed twice in RF-spoiled gradient echo signal equation to measured distilled water. The slides were then dehydrated voxelwise signal intensities as a function of flip angle. through 95 % alcohol once followed by 100 % alcohol Voxelwise signal-time series from dynamic two times over 3 minutes in each step. We cleared the T1-weighted volumes were converted to contrast slides in clear xylene twice (3 minutes each) and agent (CA) concentration time series, assuming a coverslip the slides using resinous mounting. linear relationship between 1/T1 and CA MRI concentration, and a spin-lattice relaxivity r1 = 3.5 (s Mice were imaged with MRI at 3 h, 24 h, 48 h, mm)-1. Vascular input functions were measured by and 72 h following ICH. Not all mice received scans at selecting a voxel in the superior sagittal sinus (SSS) as each timepoint due to failed cannulation and follows: the area under the concentration-time curve spontaneous deaths – details of scanning time points (AUC) for the first 5 post-contrast time points was for each mouse are given in supplementary materials. calculated. The ROI was defined as the SSS voxel with Before MRI, mice underwent tail vein cannulation a maximum AUC [79]. Ktrans maps were then under 4-5 % isoflurane in a mixture of 30 % oxygen generated by fitting the Patlak model to CA time and 70 % nitrous oxide for contrast agent series [80]. All model fitting was done using administration. Scans were acquired on an Agilent 7T ‘lsqcurvefit’ optimiser in Matlab (Mathworks, version 16 cm bore magnet interfaced with a Bruker Avance 2017a). III console running Paravision 6.0.1. During scanning, To evaluate serial changes in T1, T2* signal, and mice were anaesthetized using 2 % isoflurane in a Ktrans in the same mouse, parameter maps for a mixture of 30 % oxygen and 70 % nitrous oxide. particular mouse were co-registered to the map from High-resolution axial T2-TurboRARE images were the earliest imaging timepoint. This was done using acquired for ventricle size assessment and definition in-house image registration software written in of lesion ROIs. Acquisition parameters were: TR/TE = Matlab (Mathworks, release 2017a) based on 3592/35 ms, NEX = 2, RARE factor = 8, echo spacing = maximising the product of the source and target local 11.6 ms, matrix size = 256 x 256, 24 slices, slice image gradients. The target image was defined as the thickness = 1 mm, field of view (FOV) = 16 x 16 mm. first acquired SPGR volume (flip angle = 60°) and the Axial 3D RF-spoiled gradient-echo images were source image as the matching SPGR volume from any acquired for T1 mapping (using the variable flip angle subsequent timepoint. The estimated spatial approach) and assessment of T2* signal changes. For transformations were then applied to the associated T1 mapping the following acquisition parameters parameter maps. Visual checks of the aligned maps were used: TR/TE = 100 / 2.4 ms, flip angles = 5, 10, were completed to confirm successful registration. 40, 60, matrix size = 96 x 96 x 64, FOV = 16 x 16 x 24 Once aligned, lesion ROIs defined on T2-RARE mm. To quantify T2* signal changes, the spoiled volumes were re-gridded to the matrix size of gradient echo scan with a flip angle of 40° was parameter maps. To quantify T1, T2* signal and Ktrans repeated with a longer TE to increase the sensitivity of changes, mean values in the lesion ROIs were the scan to iron-based susceptibility gradients (TE = calculated. To investigate areas of highest contrast 4.95 ms). To quantify the leakage rate of contrast agent agent leakage, the 95th percentile of Ktrans was also (Ktrans, [min-1]) across the blood-brain barrier, calculated. dynamic T1-weighted volumes were acquired before, Statistical analyses during, and after intravenous injection of Gd-DOTA Statistical analysis of all data except MRI was (0.5 M delivered through a tail vein cannula in a performed using Graph Pad Prism 8 software. volume of 3.3 µL / g at a rate of 0.25 mL / min). The Two-tailed unpaired student t-test and one-way total duration of the dynamic scan was 40 minutes. analysis of variance followed by the Tukey multiple Acquisition parameters were: TR/TE = 50/2.4 ms, flip https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4495 comparison tests were used and p values < 0.05 were on helped with the preparation for in vivo studies. SK, considered significant. MRI data statistical analysis analyzed the MR ventricular volume and imaged and was performed in R (version 4.0.2). Mixed-effects analysed Cav1 and Claudin-5 data. JB produced the model analysis was performed instead of a one-way images taken with the Leica thunder microscope analysis of variance since not all mice were imaged at included in figure S5. SC, RG & MW performed all timepoints. Mixed-effects model analysis was Prussian blue staining and produced iron staining followed by Tukey multiple comparison tests and p data included in Figures 5 and S6. KK discussed the values < 0.05 were considered significant. For all findings and the design of the project at the initial analyses, data are represented as mean ± standard stages and provided support and infrastructure on the error of the mean (SEM), unless otherwise indicated. SPECT/CT, IVIS and radiolabeling work. SA provided continuous guidance in the conceptual Supplementary Material design of the work, data interpretation, reviewed and Supplementary figures and table. edited the manuscript. https://www.thno.org/v12p4477s1.pdf Competing Interests Acknowledgements The authors have declared that no competing The authors acknowledge Lipoid Co. (Germany) interest exists. for the lipid sample gifts, the staff at the bioimaging facility at the University of Manchester and NTU for References advice on confocal imaging and data analysis. The 1. Luengo-Fernandez R, Gray AM, Bull L, Welch S, Cuthbertson F, Rothwell PM. Quality of life after TIA and stroke: Ten-year results of the oxford vascular authors also would like to acknowledge the access to study. Neurology. 2013; 81:1588–95. 2. Sung SF, Chen SCC, Lin HJ, et al. Oxfordshire community stroke project the Leica Thunder Imager 3D Cell Culture system classification improves prediction of post-thrombolysis symptomatic funded by the Office for Students at NTU. They also intracerebral hemorrhage. BMC Neurol. 2014; 14: 1-7. 3. Qureshi AI, Mohammad YM, Yahia AM, et al. A prospective multicenter wish to thank the Biological Services Facility at the study to evaluate the feasibility and safety of aggressive antihypertensive University of Manchester for the excellent animal treatment in patients with acute intracerebral hemorrhage. J Intensive Care Med. 2005; 20: 34–42. husbandry. The cost of this study has been covered 4. Andersen KK, Olsen TS, Dehlendorff C, Kammersgaard LP. Hemorrhagic and by: a) Dr Z Al-Ahmady research fellowship (School of ischemic strokes compared: Stroke severity, mortality, and risk factors. Stroke. 2009; 40: 2068–72. Pharmacy and Optometry, University of Manchester), 5. Leppälä JM, Virtamo J, Fogelholm R, Albanes D, Heinonen OP. Different risk b) QR fund and Vice-Chancellor PhD studentship factors for different stroke subtypes: Association of blood pressure, cholesterol, and antioxidants. Stroke. 1999; 30: 2535–40. awarded for Dr Zahraa Al-Ahmady and Satinderdeep 6. Ariesen MJ, Claus SP, Rinkel GJE, Algra A. Risk factors for intracerebral Kaur, Nottingham Trent University, c) BBSRC DTP hemorrhage in the general population: A systematic review. Stroke. 2003; 34: 2060-65. Studentship awarded to Jessica Bates, Nottingham 7. Flibotte JJ, Hagan N, O’Donnell J, Greenberg SM, Rosand J. Warfarin, Trent University, d) MRC DTP studentship to Jack hematoma expansion, and outcome of intracerebral hemorrhage. Neurol. 2004; 63: 1059–64. Barrington at University of Manchester. Jack 8. Aguilar MI, Hart RG, Kase CS, et al. Treatment of warfarin-associated intracerebral hemorrhage: Literature review and expert opinion. Mayo Clin Barrington also received support from the Natalie proc. 2007; 82: 82–92. Kate Moss Trust, e) MRC grant (MR/N003586/1) for 9. Urday S, Kimberly WT, Beslow LA, et al. Targeting secondary injury in intracerebral haemorrhage-perihaematomal oedema. Nat Rev Neurol. 2015; Eloise Lemarchand at University of Manchester, and 11: 111-22. f) EPSRC grant for Dr Ben Dickie at the University of 10. Zhao X, Sun G, Zhang J, et al. Hematoma resolution as a target for intracerebral hemorrhage treatment: Role for peroxisome Manchester (EP/S021510). Graphical abstract and proliferator-activated receptor γ in microglia/macrophages. Ann Neurol. schematic graphes were performed using Biorender. 2007; 61: 352–62. 11. Wang J. Preclinical and clinical research on inflammation after intracerebral hemorrhage. Prog Neurobiol. 2010; 92: 463-477 Author Contributions 12. Healy S, McMahon J, Owens P, FitzGerald U. Significant glial alterations in response to iron loading in a novel organotypic hippocampal slice culture ZA initiated, designed, planned, and led the model. Sci Rep. 2016; 6: 1–12. study, performed the ICH collagenase model in mice 13. Xi G, Strahle J, Hua Y, Keep RF. Progress in translational research on intracerebral hemorrhage: Is there an end in sight? Prog Neurobiol. 2014; 115: and most of the experimental work including data 45-63. analysis, prepared the figures and drafted the 14. Qureshi AI, Hanley DF. Intracerebral haemorrhage. Lancet. 2009; 373: 1632–44. 15. Pulicherla KK, Verma MK. Targeting Therapeutics Across the Blood Brain manuscript. BD performed MR imaging and MR data Barrier (BBB), Prerequisite Towards Thrombolytic Therapy for analysis and participated in writing and editing the Cerebrovascular Disorders—an Overview and Advancements. AAPS PharmSciTech. 2015; 16: 223–33. manuscript. IA helped in the initial stages of the 16. Nico B, Ribatti D. Morphofunctional Aspects of the Blood-Brain Barrier. Curr project with histology data analysis and IHC Drug Metab. 2012; 13: 50–60. 17. Reese TS, Karnovsky MJ. Fine structural localization of a blood-brain barrier to experiments. DJ performed SPECT/CT imaging and exogenous peroxidase. J Cell Biol. 1967; 34: 207–17. radiolabelling stability assays. JB and EL provided 18. Rosenberg GA, Estrada E, Kelley RO, Kornfeld M. Bacterial collagenase disrupts extracellular matrix and opens blood-brain barrier in rat. Neurosci training on the ICH model and advised on the ICH Lett. 1993; 160: 117–9. 19. Keep RF, Zhou N, Xiang J, Andjelkovic A V., Hua Y, Xi G. Vascular disruption model throughout the project. MH performed CSF and blood-brain barrier dysfunction in intracerebral hemorrhage. Fluids samples collections. GC provided technical training Barriers CNS. 2014; 11: 1–13. https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4496 20. Al-Ahmady ZS, Jasim D, Ahmad SS, et al. Selective Liposomal Transport 50. Zhou Y, Wang Y, Wang J, Anne Stetler R, Yang QW. Inflammation in through Blood Brain Barrier Disruption in Ischemic Stroke Reveals Two intracerebral hemorrhage: From mechanisms to clinical translation. Prog Distinct Therapeutic Opportunities. ACS Nano. 2019; 13: 12470–86. Neurobiol. 2014; 115: 25 –44. 21. Al-Jamal WT, Al-Ahmady ZS, Kostarelos K. Pharmacokinetics & tissue 51. Wang J, Tsirka SE. Contribution of Extracellular Proteolysis and Microglia to distribution of temperature-sensitive liposomal doxorubicin in tumor-bearing Intracerebral Hemorrhage. Neurocrit Care. 2005; 3: 077–85. mice triggered with mild hyperthermia. Biomaterials. 2012; 33: 4608–17. 52. Zeng Y, Tao N, Chung K-N, Heuser JE, Lublin DM. Endocytosis of Oxidized 22. Chen S, Tam YYC, Lin PJC, Leung AKK, Tam YK, Cullis PR. Development of Low Density Lipoprotein through Scavenger Receptor CD36 Utilizes a Lipid lipid nanoparticle formulations of siRNA for hepatocyte gene silencing Raft Pathway That Does Not Require Caveolin-1. J Biol Chem. 2003; 278: following subcutaneous administration. J Control Release. 2014; 196: 106–12. 45931–6. 23. Al-Ahmady ZS, Al-Jamal WT, Bossche JV, et al. Lipid-peptide vesicle 53. IN B, R K, AV B, et al. Role of human CD36 in bacterial recognition, nanoscale hybrids for triggered drug release by mild hyperthermia in vitro phagocytosis, and pathogen-induced JNK-mediated signaling. J Immunol and in vivo. ACS Nano. 2012; 6: 9335–46. [Internet]. 2008; 181: 7147–56. 24. Kang EJ, Major S, Jorks D, et al. Blood-brain barrier opening to large molecules 54. Sundararajan S, Gamboa JL, Victor NA, Wanderi EW, Lust WD, Landreth GE. does not imply blood-brain barrier opening to small ions. Neurobiol Dis. 2013; Peroxisome proliferator-activated receptor-γ ligands reduce inflammation and 52: 204–18. infarction size in transient focal ischemia. Neuroscience. 2005; 130: 685–96. 25. Rosenberg GA, Mun-Bryce S, Wesley M, Komfeld M. Collagenase-induced 55. Fang H, Wang P-F, Zhou Y, Wang Y-C, Yang Q-W. Toll-like receptor 4 intracerebral hemorrhage in rats. Stroke. 1990; 21: 801–7. signaling in intracerebral hemorrhage-induced inflammation and injury. J 26. Chang GY. Hematoma growth is a determinant of mortality and poor outcome Neuroinflammation. 2013; 10: 1–10. after intracerebral hemorrhage. Neurology. 2007; 68: 471-72. 56. Daly MC, Atkinson SJ, Varisco BM, et al. Role of matrix metalloproteinase-8 as 27. Kirkman MA, Allan SM, Parry-Jones AR. Experimental intracerebral a mediator of injury in intestinal ischemia and reperfusion. FASEB J. 2016; 10: hemorrhage: Avoiding pitfalls in translational research. J Cereb Blood Flow 3453–60. Metab. 2011; 31: 2135–51. 57. Zhao BQ, Wang S, Kim HY, et al. Role of matrix metalloproteinases in delayed 28. A G, H S, Y B. Pharmacokinetics of pegylated liposomal Doxorubicin: review cortical responses after stroke. Nat Med. 2006; 12: 441–5. of animal and human studies. Clin Pharmacokinet. 2003; 42: 419–36. 58. Cunningham LA, Wetzel M, Rosenberg GA. Multiple roles for MMPs and 29. Bulbake U, Doppalapudi S, Kommineni N, Khan W. Liposomal formulations TIMPs in cerebral ischemia, GLIA. 2005; 50: 329–39. in clinical use: An updated review. Pharmaceutics. 2017; 9: 1-33. 59. Yong VW. Metalloproteinases: Mediators of pathology and regeneration in the 30. Zazulia AR, Diringer MN, Videen TO, et al. Hypoperfusion without ischemia CNS, Nat Rev Neurosci. 2005; 6: 931-44. surrounding acute intracerebral hemorrhage. J Cereb Blood Flow Metab. 2001; 60. Rajkovic O, Gourmel C, d’Arcy R, et al. Reactive Oxygen Species‐Responsive 21: 804–10. Nanoparticles for the Treatment of Ischemic Stroke. Adv Ther. 2019; 2: 31. Kate MP, Hansen MB, Mouridsen K, et al. Blood pressure reduction does not 1900038. reduce perihematoma oxygenation: A CT perfusion study. J Cereb Blood Flow 61. Bolan F, Louca I, Heal C, Cunningham CJ. The potential of biomaterial-based Metab. 2014; 34: 81–6. approaches as therapies for ischemic stroke: A systematic review and 32. Herweh C, Jüttler E, Schellinger PD, Klotz E, Schramm P. Perfusion CT in meta-analysis of pre-clinical studies. Front Neurol. 2019; 10: 924. hyperacute cerebral hemorrhage within 3 hours after symptom onset: Is there 62. Galho AR, Cordeiro MF, Ribeiro SA, et al. Protective role of free and an early perihemorrhagic penumbra? J Neuroimaging. 2010; 20: 350–3. quercetin-loaded nanoemulsion against damage induced by intracerebral 33. Yang GY, Betz AL, Chenevert TL, Brunberg JA, Hoff JT. Experimental haemorrhage in rats. Nanotechnology. 2016; 27: 175101. intracerebral hemorrhage: Relationship between brain edema, blood flow, and 63. O S Elizarova, S A Litvinova, V Iu Balaban’ian, I V Barskov, S V Novikova, E V blood-brain barrier permeability in rats. J Neurosurg. 1994; 81: 93–102. Stel’mashuk, et al. Neuroprotector effect of human recombinant 34. Yang J, Li Q, Wang Z, et al. Multimodality MRI assessment of grey and white erythropoietin sorbed on polymer nanoparticles studied on model of matter injury and blood-brain barrier disruption after intracerebral intracerebral post-traumatic hematoma (hemorrhagic stroke). Eksp Klin haemorrhage in mice. Sci Rep. 2017; 7: 40358. Farmakol. 2011;74:17-22. 35. Hallevi H, Abraham AT, Barreto AD, Grotta JC, Savitz SI. The spot sign in 64. Boado RJ, Hui EKW, Zhiqiang Lu J, Pardridge WM. Drug targeting of intracerebral hemorrhage: The importance of looking for contrast erythropoietin across the primate blood-brain barrier with an IgG molecular extravasation. Cerebrovasc Dis. 2010; 29: 217–20. Trojan horse. J Pharmacol Exp Ther. 2010; 333: 961–9. 36. Abid KA, Sobowale OA, Parkes LM, et al. Assessing Inflammation in Acute 65. Pardridge WM. Drug transport across the blood-brain barrier. J Cereb Blood Intracerebral Hemorrhage with PK11195 PET and Dynamic Flow Metab. 2012; 32: 1959–72. Contrast-Enhanced MRI. J Neuroimaging. 2018; 28: 158–61. 66. Greenblatt DJ, Sethy VH. Benzodiazepine concentrations in brain directly 37. Brouwers HB, Greenberg SM. Hematoma expansion following acute reflect receptor occupancy: studies of diazepam, lorazepam, and oxazepam. intracerebral hemorrhage. Cerebrovasc Dis. 2013; 35: 195-201. Psychopharmacology (Berl). 1990; 102: 373–8. 38. Keep RF, Andjelkovic A V, Xiang J, et al. Brain endothelial cell junctions after 67. Liu S, Guo Y, Huang R, et al. Gene and doxorubicin co-delivery system for cerebral hemorrhage: Changes, mechanisms and therapeutic targets. J Cereb targeting therapy of glioma. Biomaterials. 2012; 33: 4907–16. Blood Flow Metab. 2018; 38: 1255–75. 68. Johnsen KB, Burkhart A, Melander F, et al. Targeting transferrin receptors at 39. Wang T, Chen X, Wang Z, et al. Poloxamer-188 Can Attenuate Blood-Brain the blood-brain barrier improves the uptake of immunoliposomes and Barrier Damage to Exert Neuroprotective Effect in Mice Intracerebral subsequent cargo transport into the brain parenchyma. Sci Rep. 2017; 7: 10396. Hemorrhage Model. J Mol Neurosci. 2015; 55: 240–50. 69. Lee D-H, Rötger C, Appeldoorn CCM, et al. Glutathione PEGylated liposomal 40. Li Z, Chen X, Zhang X, et al. Small Interfering RNA Targeting Dickkopf-1 methylprednisolone (2B3-201) attenuates CNS inflammation and degeneration Contributes to Neuroprotection After Intracerebral Hemorrhage in Rats. J Mol in murine myelin oligodendrocyte glycoprotein induced experimental Neurosci. 2017; 61: 279–88. autoimmune encephalomyelitis. J Neuroimmunol. 2014; 274: 96–101. 41. Xie R-X, Li D-W, Liu X-C, et al. Carnosine Attenuates Brain Oxidative Stress 70. Gaillard PJ, Appeldoorn CCM, Rip J, et al. Enhanced brain delivery of and Apoptosis After Intracerebral Hemorrhage in Rats. Neurochem Res. 2017; liposomal methylprednisolone improved therapeutic efficacy in a model of 42: 541–51. neuroinflammation. J Control Release. 2012; 164: 364–9. 42. Jiao X, He P, Li Y, et al. The Role of Circulating Tight Junction Proteins in 71. Gaillard PJ, Appeldoorn CCM, Dorland R, Van Kregten J, Manca F. Evaluating Blood Brain Barrier Disruption following Intracranial Pharmacokinetics, Brain Delivery, and Efficacy in Brain Tumor-Bearing Mice Hemorrhage. Dis Markers. 2015; 860120. of Glutathione Pegylated Liposomal Doxorubicin (2B3-101). PLoS One. 2014; 9: 43. MacLellan CL, Silasi G, Poon CC, et al. Intracerebral hemorrhage models in 82331. rat: Comparing collagenase to blood infusion. J Cereb Blood Flow Metab. 2008; 72. Anselmo AC, Mitragotri S. Nanoparticles in the clinic: An update. Bioeng 28: 516–25. Transl Med. 2019; 4: e10143. 44. Clark CA, Werring DJ, Miller DH. Diffusion imaging of the spinal cord in 73. Hu W, Zhao M, Gu K, Xie L, Liu M, Lu D. Fluorescent probe for the detection vivo: Estimation of the principal diffusivities and application to multiple of hypochlorous acid in water samples and cell models. 2021; 12: 777–84. sclerosis. Magn Reson Med. 2000; 43: 133–8. 74. Li S, Yang K, Zeng J, Ding Y, Cheng D, He L. Golgi-Targeting Fluorescent 45. Keep RF, Hua Y, Xi G. Intracerebral haemorrhage: Mechanisms of injury and Probe for Monitoring CO-Releasing Molecule-3 In Vitro and In Vivo. ACS therapeutic targets. Lancet Neurol. 2012; 11: 720-31. Omega. 2022; 7: 9929–35. 46. Dixon SJ, Lemberg KM, Lamprecht MR, et al. Ferroptosis: An iron-dependent 75. Zhang Y, Zhang X, Huocheng Yang bc, et al. Advanced biotechnology-assisted form of nonapoptotic cell death. Cell. 2012; 149: 1060–72. precise sonodynamic therapy. Chem Soc Rev. 2021; 50: 11227. 47. Yagami T, Yamamoto Y, Koma H. Pathophysiological Roles of Intracellular 76. Inman CF, Rees LEN, Barker EN, Haverson K, Stokes CR, Bailey M. Validation Proteases in Neuronal Development and Neurological Diseases. Mol of computer-assisted, pixel-based analysis of multiple-colour Neurobiol. 2019; 56: 3090-112 immunofluorescence histology. J Immunol Methods. 2005; 302: 156–67. 48. Wan J, Ren H, Wang J. Iron toxicity, lipid peroxidation and ferroptosis after 77. Nemes-Baran AD, White DR, DeSilva TM. Fractalkine-Dependent Microglial intracerebral haemorrhage. Stroke Vasc Neurol. 2019; 4: 93–95. Pruning of Viable Oligodendrocyte Progenitor Cells Regulates Myelination. 49. Zhao X, Sun G, Zhang J, et al. Hematoma resolution as a target for Cell Rep. 2020; 32: 108047. intracerebral hemorrhage treatment: Role for peroxisome 78. AD N-B, TM D. Quantification of microglial contact and engulfment of proliferator-activated receptor γ in microglia/macrophages. Ann Neurol. oligodendrocyte progenitor cells in the rodent brain. STAR Protoc. 2021; 2: 2007; 61: 352–62. 100403. https://www.thno.org Theranostics 2022, Vol. 12, Issue 10 4497 79. Dickie BR, Vandesquille M, Ulloa J, Boutin H, Parkes LM, Parker GJM. Water-exchange MRI detects subtle blood-brain barrier breakdown in Alzheimer’s disease rats. Neuroimage. 2019; 184: 349–58. 80. Patlak CS, Blasberg RG. Graphical evaluation of blood-to-brain transfer constants from multiple-time uptake data. Generalizations. J Cereb Blood Flow Metab. 1985; 5: 584–90. https://www.thno.org