This protocol describes the surgical and technical procedures that enable real-time in vivo multiphoton fluorescence imaging of the rodent brain during focused ultrasound and microbubble treatments to increase blood-brain barrier permeability.
The blood-brain barrier (BBB) is a key challenge for the successful delivery of drugs to the brain. Ultrasound exposure in the presence of microbubbles has emerged as an effective method to transiently and locally increase the permeability of the BBB, facilitating para- and transcellular transport of drugs across the BBB. Imaging the vasculature during ultrasound-microbubble treatment will provide valuable and novel insights on the mechanisms and dynamics of ultrasound-microbubble treatments in the brain.
Here, we present an experimental procedure for intravital multiphoton microscopy using a cranial window aligned with a ring transducer and a 20x objective lens. This set-up enables high spatial and temporal resolution imaging of the brain during ultrasound-microbubble treatments. Optical access to the brain is obtained via an open-skull cranial window. Briefly, a 3-4 mm diameter piece of the skull is removed, and the exposed area of the brain is sealed with a glass coverslip. A 0.82 MHz ring transducer, which is attached to a second glass coverslip, is mounted on top. Agarose (1% w/v) is used between the coverslip of the transducer and the coverslip covering the cranial window to prevent air bubbles, which impede ultrasound propagation. When sterile surgery procedures and anti-inflammatory measures are taken, ultrasound-microbubble treatments and imaging sessions can be performed repeatedly over several weeks. Fluorescent dextran conjugates are injected intravenously to visualize the vasculature and quantify ultrasound-microbubble induced effects (e.g., leakage kinetics, vascular changes). This paper describes the cranial window placement, ring transducer placement, imaging procedure, common troubleshooting steps, as well as advantages and limitations of the method.
A key challenge to treating neurological disorders is the presence of the blood-brain barrier (BBB). The BBB limits hydrophilic, charged, polar, and large (> 400 Da) molecules from entering the brain parenchyma1. One method currently used to deliver therapeutics across the BBB into the brain parenchyma is to use stereotactic intracranial injections2. Other less invasive methods under investigation are hindered by the complexity of the techniques used, such as designing drugs for receptor-mediated delivery across the BBB3, or are limited in the spatial precision of targeted areas, such as intranasal injections4 or administration of hyperosmotic solutions5.
The use of ultrasound in conjunction with systemically injected microbubbles, an ultrasound contrast agent, has been developed as a noninvasive means to transiently increase the permeability of the BBB6. By using a focused transducer7 or a steerable phased array of transducers8,9, ultrasound can be targeted to selected areas in the brain with millimeter level precision, minimizing off-target effects. Ultrasound-microbubble treatments can be customized to each subject's brain anatomy by using magnetic resonance imaging guidance7,10,11,12,13,14 or stereotactic frames15. Furthermore, the extent of increase in BBB permeability can be controlled in real-time by monitoring acoustic emissions from microbubbles16,17,18. Clinical trials investigating the safety and feasibility of ultrasound-microbubble treatments are currently in progress worldwide (e.g., ClinicalTrials.gov identifier NCT04118764).
Ultrasound-microbubble BBB treatments are typically evaluated by confirming treatment induced increases in BBB permeability, visualized in contrast-enhanced magnetic resonance imaging, or by dye extravasation in in vivo imaging or ex vivo histology. However, most microscopic analyses have been performed ex vivo, following the completion of ultrasound-microbubble treatments11,19, thereby missing the dynamic biological responses during, and immediately following, ultrasound exposure. Real-time imaging conducted during ultrasound exposure may aid in understanding the mechanisms driving ultrasound-microbubble BBB treatments as well as downstream responses, which may increase our understanding of its therapeutic applications. Furthermore, the use of chronic cranial windows with in vivo imaging techniques would enable longitudinal studies to evaluate temporal aspects of ultrasound-microbubble treatments.
The goal of this protocol is to describe the surgical and technical procedures required to conduct real-time multiphoton imaging of ultrasound-microbubble treatments for acute and chronic studies in rodents (Figure 1). This is achieved in two parts: first, to create a cranial window to enable in vivo imaging, and second, to mount a ring transducer on the top to enable concurrent sonication and imaging. Cranial windows have been extensively used by neuroscientists for in vivo imaging of neurovascular coupling20, β-amyloid pathogenesis21, and neuroimmunology22, among others. In this protocol, surgical procedures for creating acute (non-recovery) and chronic (recovery) cranial windows in the mouse and rat skull are described. Cranial window methodologies, particularly for chronic experiments, have been well-documented23,24,25. To be consistent with existing literature, the terms 'acute' and 'chronic' will be used throughout this protocol. The design of ring transducers for in vivo imaging has also been previously described26. Despite the availability of these techniques and the insights that can be gained from real-time imaging of ultrasound-microbubble treatments, there are very few research laboratories that have successfully published literature using this technique26,27,28,29,30,31,32. As such, in this protocol, the surgical and technical details of conducting these real-time ultrasound-microbubble experiments are described. While the specified sonication and imaging parameters have been optimized for BBB experiments, other effects of ultrasound exposure to the brain, such as neuromodulation33,34, β-amyloid plaque monitoring31, and immune cell responses32, can also be investigated using this technique.
All the following experimental procedures were approved by and conducted in accordance with the Norwegian Food and Safety Authority, Sunnybrook Research Institute's Animal Care Committee, and the Canadian Council on Animal Care.
1. Material preparation
2. Animal preparation
3. Placement of the ring transducer
4. Multiphoton microscopy imaging
5. Ultrasound exposure
6. Image analysis
Successful ultrasound-microbubble treatments can be detected by the extravasation of fluorescent dextran from the intravascular to the extravascular space (Figure 8), indicating an increase in BBB permeability. Depending on the pressure field of the ring transducer, pial vessels and/or capillaries will be affected.
To evaluate the vascular changes induced by ultrasound-microbubble treatments, the diameter of the vessel of interest can be measured before, during, and after ultrasound-microbubble treatment (Figure 9). This can be done manually in a commercially available software (e.g., Olympus Fluoview software). During image acquisition, bolus dextran injections and line scans can also be used to assess blood flow30,41. To evaluate kinetics of dextran leakage as a representative model for drug delivery, the signal intensity between the intra- and extravascular spaces can be evaluated using tools such as MATLAB26,27,29,41 (Figure 10).
Further image processing can be achieved using ImageJ/FIJI. ImageJ/FIJI is an open-source software that is compatible with MATLAB and is well-suited to conduct common analyses in biological image analysis, such as measuring vascular changes, or the lengths of or distance between fluorescent objects (e.g., β-amyloid plaques to blood vessels). Image processing pipelines created in ImageJ/FIJI can be automated by writing custom macros.
More complex analyses, such as 3D segmentation of blood vessels and cell tracking, can be achieved using more advanced, semi-automated software (Figure 11). Following segmentation, more specific analyses can be conducted, such as classifying blood vessels as arterioles, venules, or capillaries, based on diameter, branching, tortuosity patterns, and flow direction42,43. Machine learning algorithms have also been developed to automate blood vessel segmentation22,44.
Figure 1: General workflow of intravital multiphoton ultrasound-microbubble brain experiments. A general workflow of the intravital multiphoton ultrasound-microbubble brain experiments described in this protocol is shown. There are 6 steps: (A) Animal preparation for (A1) mice and (A2) rats, (B) Dextran injection, (C) Microbubble injection, (D) Pre-treatment imaging, (E) Treatment and imaging, (F) Post-treatment imaging and data analysis. Please click here to view a larger version of this figure.
Figure 2: Cross-section and top view of 3D-printed mold. (A) Cross-section of the mold. A thin layer of cyanoacrylate glue is applied on the top surface of the ring transducer, and a coverslip is placed on top. A stamp may be used to apply firm, even pressure on the coverslip and ring transducer. (B) Top view of the mold. A notch can be added in the mold to facilitate removal of the prepared transducer. Please click here to view a larger version of this figure.
Figure 3: Ultrasound set-up. Typical hardware for ultrasound experiments are shown. Ultrasound parameters are set and triggered by the signal generator and amplified by the amplifier. A power meter can be used to record forward and reflected powers prior to sending the signal to the matching box, which is matched to the transducer. All connections are achieved using BNC cables unless stated otherwise. Please click here to view a larger version of this figure.
Figure 4: Area of fur removal and scalp removal. (A) Fur removal should start from between the eyes and extend until the anterior half of the neck. (B) Scalp removal should be sufficient to expose the parietal bones. Bleeding must be stopped before proceeding. Please click here to view a larger version of this figure.
Figure 5: Outline of the cranial window. The cranial window is situated on a parietal bone. (A) An outline of the cranial window can be drawn onto the skull to aid in the drilling process. (B) The outline of the cranial window can be seen following drilling through the compact bone. Please click here to view a larger version of this figure.
Figure 6: Cranial window and transducer alignment. (A) The cranial window is created on a parietal bone. The bone island has been removed, exposing the brain underneath. (B) The cranial window is complete when a glass coverslip is sealed onto the skull using cyanoacrylate glue. (C) The transducer is centered to the cranial window and adhered using cyanoacrylate glue. Please click here to view a larger version of this figure.
Figure 7: Positioning of objective lens and transducer. (A,B) The objective lens is centered to the ring transducer. (C) Blood vessels filled with fluorescent dextran are visible through the eyepieces, under epifluorescence. Please click here to view a larger version of this figure.
Figure 8: Maximum projection multiphoton images of ultrasound-microbubble induced increases in BBB permeability. Maximum projection images of vasculature (A) before and (B) after ultrasound-microbubble treatments. Successful ultrasound-microbubble treatments can be confirmed by observing increases in BBB permeability following treatment, visualized as fluorescent dextran extravasation (arrows). Scale bar: 50 µm. Please click here to view a larger version of this figure.
Figure 9: Analysis of vasomodulation induced by ultrasound-microbubble treatments. Maximum projection images of cerebral blood vessels before, during, and after ultrasound-microbubble treatments. Microbubbles are present in all images. Compared to (A) pre-treatment conditions, clear vasomodulation can be observed (B) during ultrasound-microbubble treatments (red arrows). Ultrasound-microbubble mediated increases in BBB permeability are also evident following treatment from the leakage of fluorescent dextran from the intravascular to the extravascular space (yellow arrows). (C) When ultrasound is turned off, vascular diameters return to pre-treatment, baseline sizes. (D) Vascular changes can be analyzed by plotting the diameter of the vessel of interest before, during, and after ultrasound-microbubble treatment. Scale bar: 100 µm. (Unpublished work). Please click here to view a larger version of this figure.
Figure 10: Analysis of leakage kinetics following ultrasound-microbubble treatments. Increase in BBB permeability is visualized as leakage of fluorescent dextran from the intravascular to the extravascular space. Changes in BBB permeability are evident when comparing image stacks acquired (A) before and (B) after ultrasound-microbubble treatments. (C) Leakage kinetics can be analyzed by tracking the intensity, volume, and speed of dextran in extravascular compartments (yellow rectangle). Scale bar: 50 µm. (Unpublished work.) Please click here to view a larger version of this figure.
Figure 11: Blood vessel segmentation of multiphoton microscopy XYZ stack. (A) Depth (XYZ) stack of blood vessels in a transgenic EGFP rat. Blood vessels are visualized via intravenous injection of fluorescent Texas Red 70 kDa dextran (red). The green channel shows fluorescent cells and tissue autofluorescence. (B) 3D reconstructions of blood vessels are created, and then color-coded according to blood vessel type to facilitate type-specific analyses. Vein/venules are blue, arteries/arterioles are red, and capillaries are cyan. Scale bar: 50 µm. Reconstructions created using Bitplane Imaris. Please click here to view a larger version of this figure.
Intravital multiphoton microscopy monitoring of the brain is a valuable tool to study brain responses during ultrasound exposure. To our knowledge, the protocol described here is the only method of conducting multiphoton microscopy imaging of the brain parenchyma during ultrasound-microbubble treatments. The creation of cranial windows and the use of ring transducers allow real-time monitoring of vascular, cellular, and other downstream responses to ultrasound-microbubble treatments at high spatial and temporal resolution. Other groups have performed multiphoton microscopy imaging following the completion of ultrasound-microbubble treatments, thereby missing the real-time response of the brain parenchyma to treatments19. The procedure described offers improved temporal control, allowing the collection of data that may help illuminate the acute mechanisms behind ultrasound-microbubble treatments. Quantitative and qualitative data can be extracted and analyzed from the acquired image stacks, such as extravasation kinetics27,29,30, changes in β-amyloid plaque volume31, and cell dynamics32.
Several troubleshooting steps were highlighted throughout the protocol. First, surgical steps that are particularly susceptible to operator error were emphasized, such as use of agarose during cranial window surgery and placement of the transducer. Steps to prevent animal discomfort and death were also provided, including monitoring animal physiology during surgery, and thoroughly vortexing the dextran prior to injection. Second, physical specifications of the transducer, and alignment of the objective lens, transducer, and cranial window, were also highlighted. The specifications of the ring transducer and its acoustic properties must be determined in consideration of the objective lens used as well as the animal model. Specifically, the inner diameter of the ring transducer must be large enough to surround the objective lens, but small enough to be mounted securely onto the animal's skull. In addition, the focal area of the transducer must align with the range of the objective lens used.
A common challenge is that the cranial window and ring transducer are angled relative to the objective lens. Proper centering (XY) and alignment (Z) of the objective lens with the cranial window and transducer ensures that the focal area of the transducer, and thus the region of treated brain tissue, aligns with the imaging field-of-view, and reduces the risk of collision between the objective lens and transducer during imaging. Alignment can be achieved by adjusting the head position of the animal and/or by rotating the stereotactic frame that it is fixed in.
Microscope components (e.g., detectors, beam splitters) and image acquisition parameters should be selected based on the aim of the study. Here, an objective lens with a long focal length (> 2 mm) is used due to the presence of the coverslip(s) and ring transducer located between the objective lens and the brain. An upright microscope is also recommended as it allows for more space to maneuver the animal, particularly for brain experiments. To capture the kinetics of ultrasound-microbubble induced leakage of the intravascular dye, a high temporal resolution is favorable, which can be achieved by using a resonance scanning system. Combining this with a high sensitivity detection system, such as gallium arsenide phosphide (GaAsP) detectors, will also result in more favorable images.
The experimental procedure presented has several limitations. First, the surgical procedure is quite invasive, and has been reported to cause inflammation45, although inflammation can be minimized46. Moreover, immune responses induced by cranial window surgeries were observed to resolve by 2-4 weeks following surgery23,24,25. In addition, the drilling process, particularly when conducted with excessive force or speed, can cause damage to underlying tissue due to the generation of heat, vibration, and pressure applied. Cranial window surgeries and multiphoton imaging have also been observed to affect brain temperature47. These limitations can be reduced to an extent through careful creation of pristine cranial windows, proper recovery of animals with chronic cranial windows, and maintenance of normothermic body temperature using a heating source with feedback control. Second, the imaging depth is limited by the microscope and objective lens used. For example, the effect of ultrasound-microbubble treatment in deeper brain structures, such as the hippocampus, cannot be studied without more invasive measures, such as the removal of overlying cortical tissue48, or the use of microlenses in conjunction with cortical penetration49. Using an objective lens with a long working distance could resolve this issue to an extent, but light penetration is also limited at greater depths.
While the representative images of this protocol were acquired from wild-type rodents, the presented experimental procedure can also be applied to transgenic animals and disease models, such as Alzheimer's disease31. Ultrasound experiments unrelated to BBB modulation, such as ultrasound-induced neuromodulation, can also be monitored using this protocol33,34. Other possible applications can be achieved by using different microscope or detector set-ups, such as pairing a confocal microscope with an ultra-high-speed camera50. While photobleaching and phototoxicity are comparatively worse in confocal microscopes due to the large excitation volume, ultra-high-speed imaging may enable visualization of brain capillary endothelial cell-microbubble interactions with high temporal resolution, which could further illuminate the mechanisms driving ultrasound-microbubble BBB treatments. To conclude, the protocol described provides a method to monitor vascular and cellular effects induced by ultrasound-microbubble BBB experiments in real-time, providing a tool to further determine the mechanisms driving these treatments, as well as illuminating the downstream responses of the brain parenchyma to ultrasound-microbubble treatments.
The authors have nothing to disclose.
Housing of the animals was provided by the Comparative Medicine Core Facility (CoMed, NTNU). Figure 3 was created in BioRender.com. Video recording and editing was done by Per Henning, webmaster at the Faculty for Natural Science at NTNU. The project was funded by the Norwegian University of Science and Technology (NTNU, Trondheim, Norway), Research Council of Norway (RCN 262228), Canadian Institutes of Health Research (FDN 154272), National Institute of Health (R01 EB003268), and the Temerty Chair in Focused Ultrasound Research at Sunnybrook Health Sciences Centre.
Ring transducer placement | |||
Agarose (powder) | Sigma-Aldrich | A9539 | |
Beaker or Erlenmeyer flask (50 ml) | VWR | 213-0462 or 214-1130 | |
Cyanoacrylate glue (gel) | Loctite | 1363589 | |
Glass coverslips (13 mm) | Thermo Fisher Scientific | CB00130RA120MNT0 | Coverslip for ring transducer. |
Hot plate or microwave | Corning | PC-400D | To heat agarose solution. |
PBS (1X) | Sigma-Aldrich | P4417 | |
Ring transducer | Custom-made | Custom-made | Custom-made. E.g. https://doi.org/10.1109/ULTSYM.2014.0518 |
Rubber stopper | VWR | 217-0867 | |
Animal preparation and drugs | |||
Bupivacaine*A | Aspen | 169912 | Dose: 1 mg/kg, s.c., local anesthetic injected at incision site. |
Buprenorphine*A | Indivior | 521634 | Dose mouse: 0.05-0.1 mg/kg, s.c., opioid, administer pre-surgery. |
Buprenorphine*A | Indivior | 521634 | Dose rat: 0.01-0.05 mg/kg, s.c.. |
Carprofen*C | Pfizer | DIN 02255693 | Dose: 5 mg/kg, s.c., NSAID, adminster post-surgery. |
Depilatory cream | Veet | N/A | For complete fur removal after trimming. |
Dexamethasone*C | Sandoz | DIN 00664227, 2301 | Dose: 3 mg/kg, i.m., corticosteroid, reduces cerebral edema, administer pre-surgery. |
Enrofloxacin*C | Bayer | DIN: 02249243 | Dose: 5 mg/kg, i.p., antibiotic, administer post-surgery. |
Fur clippers | Aesculap | 90200714 | Exacta/Isis. |
Heating pad | Physitemp Instruments INC | HP-1M | |
Isoflurane | Baxter | ESDG9623C | Dose: 3% induction, 1% maintenance; anesthetic. |
Meloxicam*A | Boehringer Ingelheim Vetmedica GmbH | 25388 | Dose mouse: 2-3 mg/kg, s.c., NSAID, administer pre-surgery. |
Meloxicam*A | Boehringer Ingelheim Vetmedica GmbH | 25388 | Dose rat: 1 mg/kg, s.c. |
Pulse oximeter | STARR Life Sciences Corp | N/A | MouseOx. |
Stereotaxic frame | Kopf | Kopf 900 | |
Sterile ophthalmic ointment | Théa | 597562 | Viscotears. |
Tail vein catheter (24 G) | BD Neoflon | 391350 | |
* Discuss dosing and type of administration with veterinarian prior to use. A For acute window surgeries, C For chronic window surgeries. Dose for mice and rats are the same unless otherwise specified. | |||
Material and equipment for cranial window placement | |||
Alcohol swabs | BD | 326895 | |
Curved fine surgical scissors | Fine Science Tools | 14002-12 | |
Cotton or fibreless swabs | Chemtronics | CX50 | |
Cyanoacrylate glue (gel) | Loctite | 1594457 (gel), 230992 (liquid) | If unavailable, liquid cyanoacrylate glue can be mixed with extra-fine acrylate powder. |
Dental cement | Lang Dental | Jet Set-4 Denture Repair Package | |
Dental micromotor hand drill | FOREDOM | K.1070-2 | High speed rotary micromotor kit with 2.35 mm collet. |
Forceps | Fine Science Tools | 11152-10, 11370-40 | |
Glass coverslips | Thermo Fisher Scientific | CB00050RA120MNT0 (5 mm) | Mouse cranial windows. |
Glass coverslips | Thermo Fisher Scientific | CB00080RA120MNT0 (8 mm) | Rat cranial windows. |
Micro drill burrs (0.5 mm) | Meisinger | HM71005 (0.5 mm) | |
Micro drill burrs (0.7 mm) | Meisinger | HM71007 (0.7 mm) | |
Stereo microscope | Nikon | SMZ645 | |
Surgical gelatin sponge | Ethicon | MS0005 | |
Vetbond Tissue adhesive | 3M | 1469SB | |
Weigh boats / trays | VWR | 10803-148 | |
* Autoclave drapes, tools, materials, and gowns, and use sterile surgical gloves, for chronic cranial window surgeries. | |||
Multiphoton microscopy | |||
20x water immersion objective | Olympus | XLUMPLFLN20 XW | Numerical aperture 1.0, working distance 2.0 mm. |
Fluorescent dextran (e.g. FITC 70 kDa) | Sigma Aldrich | 46945 | Recommended 10 kDa-2 MDa. |
MaiTai DeepSee Ti:Sapphire laser oscillator | Spectra-Physics | N/A | |
SliceScope microscope | Scientifica | N/A | |
Ultrasound treatment | |||
50 dB RF Amplifier | E&I | 2100L | |
Matching circuit | Custom-made | Custom-made | Custom-made. |
Microbubbles | Bracco Imaging | N/A | SonoVue (Bracco Imaging, Europe). Dose 1 ml/kg. |
Microbubbles | Lantheus | N/A | Definity (Lantheus Medical Imaging, North America). Dose 0.02-0.04 ml/kg. |
Signal generator | Agilent Technologies | 33500B |