1. Introduction

Stroke is the leading cause of adult disability in the developed world and the second leading cause of death worldwide [1] yet aside from thrombolytic agents, which must be administered within hours of the onset of ischemia, there are currently few treatment options available for those affected by stroke. Cerebral ischemia triggers a complex and dynamic series of cellular and molecular events which ultimately lead to tissue damage and loss of function [2, 3]. A better understanding of these events is critical to developing more effective treatments for stroke patients.

To this end, researchers have developed a number of different animal models of stroke [4]. One such model is the photochemical stroke method first described by Watson et al. in 1985 [5]. This method involves injecting animals systemically with a photo-sensitive dye (Rose Bengal) and then irradiating an area of brain tissue with light to activate the dye. This process releases reactive oxygen species, which then cause peroxidative damage to the endothelium and thrombus formation [6], resulting in a focal ischemic lesion localized in the illuminated region. The most common version of this method uses external illumination through the intact skull to create a focal cortical lesion, making it unsuitable for selectively targeting subcortical brain structures as the overlying blood vessels and tissue are also affected. This method has several advantages over other stroke models; it is simpler and much less invasive than surgical models, it is highly reproducible and, importantly, the experimenter has control over the size and location of the infarct, provided surface brain areas are targeted [4]. A more invasive technique is to deliver the light through an optical fiber, which can be surgically implanted into a selected region of the brain. Such a technique has been demonstrated in both rat [7–10] and mouse models [11]. While these photochemical stroke techniques have been gaining popularity among labs performing rodent models of stroke in recent times, there are still some fundamental aspects that are not completely understood. In particular, there is limited understanding of the nature and timing of the biochemical events surrounding infarction. To address this issue it is desirable to be able to measure and monitor physiological parameters at the lesion site in real-time, such that a profile of the evolution of the infarct can be generated.

Optical fibers have several key properties which make them amenable to in vivo sensing applications: their capacity to penetrate tissue within a small diameter needle or catheter, small additional pathological effects, flexibility and biocompatibility [12, 13]. In this work, we demonstrate a novel Fiber-Optic Photochemical Stroke (FOPS) technique that provides the means to deliver light to, and hence create lesions within, localized regions of the brain that are difficult to access. This technique provides the opportunity to selectively and specifically target deep brain structures that have been technically difficult to access either using standard extracranial photochemical stroke methodologies or intravascular occlusive models in rodents. A significant number of ischaemic strokes in the clinical setting are due to small vessel occlusion (either from thrombus or embolus) and the current study allows the simulation of this type of stroke in humans to small mammals. In addition, the same optical fiber can then be used to recapture the light emitted by the fluorescent dye, and this signal can be monitored in real-time. This allows spectral and temporal monitoring of the dye fluorescence to be performed during the dye-excitation phase of the trial.

We begin by demonstrating a correlation between fluorescence spectral shape and dye concentration in an in vitro model. We then show a successful lesion using the fiber optic approach, and compare the size of the lesion with illumination power. The fluorescence collected from the in vivo study is analyzed in the spectral and time domain, which confirms the collection of Rose Bengal dye fluorescence from inside the brain and provides insight into the temporal properties of the photochemical reactions leading to stroke events in this model. This development opens up a pathway towards better understanding of the underlying changes in the brain tissue after ischemic injury with the potential to measure the extent of damage induced by the photochemical stroke model without the need for tissue histology that is generally required to determine if a stroke has successfully been induced.

2. Experimental

2.1 Optical setup

The experimental setup used in the in vivo sensing part of this work is shown in Fig. 1.

 

Fig. 1 - Experimental setup used for the FOPS experiments. Signal collected by the delivery fiber is analyzed by the CCD spectrometer in a backscattering geometry.

Download Full Size | PDF

Excitation light from a 532 nm continuous wave laser (CrystaLaser CL532-100) was reflected off a long-pass 532 nm filter (Semrock LP03-532RU) and coupled into a multimode optical fiber (core diameter 200 μm) using a 10X microscope objective. The green laser source was chosen as it is a cold light source with minimum thermal content, to reduce thermal damage to the tissue. The multimode optical fiber was chosen to collect the maximum amount of fluorescence while maintaining a low power intensity at its tip. The free end of the fiber was threaded inside a 26G surgical needle to protect the fragile tip of the fiber and to allow penetration through the mouse skull. The power output at the free end of the fiber prior to skull insertion was measured using a calibrated visible-light silicon photodiode (Thorlabs S121VC). Light generated at the tip of the fiber was collected in a backscattering geometry through an additional 550 nm long-pass filter (Thorlabs FEL550) and focused by a 10x microscope objective onto a multimode patch cable connected to a CCD spectrometer (Ocean Optics QE65 PRO).

2.2 In vitro trials

Homogenized mouse brain tissue was combined from three mice and then split into six equal samples. Rose Bengal dye (Sigma, R3877) was then dissolved in saline, filtering through a 0.45 µm filter (Minisart, 16555). This dye solution was added to the homogenized mouse brain tissue to form sequentially higher concentrations (10⁻⁶, 10⁻⁵, 10⁻⁴, 10⁻³, and 10⁻² mg/ml). At each concentration the fiber was immersed in this solution and the spectra recorded at a power of 15 mW (intensity 48 W/cm²) at the tip of the fiber.

2.3 Photochemically induced subcortical ischemia

All animals were housed and treated in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes. The University of Adelaide Animal Ethics Committee approved all experiments prior to commencement (Animal Ethics # M-2013-085). C57BL6 Mice were intraperitoneally injected with Rose Bengal dye (100 mg/kg) prior to be being anesthetized with inhaled isoflurane (Attane, Bomac). The mice were then secured in a stereotaxic frame. An incision was made in the scalp and the skull exposed to reveal bregma. The stereotaxic frame was then used to guide the fiber to the target area (i.e. the hippocampus) using the coordinates 1.4 mm lateral (left hemisphere) and 2.4 mm posterior to bregma, 2 mm ventral, information provided by a mouse brain atlas [14]. To induce focal cerebral ischemia, the target area was illuminated for 10 min at constant power, in accordance with previous literature [7–10], after which the fiber was removed and the wound sutured. For the photochemical stroke trials, the powers used were 12 and 43 mW (intensities of 38 and 137 W/cm² respectively) at the tip of the fiber to investigate the effect of laser power on the infarct size. For the fluorescence collection trials 15 mW (48 mW/cm²) was used to prevent unnecessarily large area tissue damage. Following the procedure, animals received a subcutaneous injection of local anesthetic (0.5% bupivacaine hydrochloride, 50 µL) under the scalp and an antibiotic ointment (Bactroban®, 2% w/w mupirocin) was applied to the surgical site. Control animals were treated as above, but received an injection of saline solution instead of Rose Bengal. Four days after FOPS, mice were humanely killed using cervical dislocation. Brains were removed and immersion fixed in 4% paraformaldehyde for 48 h. 100 µm sections were cut using a vibratome and Nissl stained [15]. Sections were imaged using an Olympus DP72 digital camera mounted on an Olympus BX53 microscope.

2.4 Data analysis

The spectra collected from the optical measurements, both in vitro and in vivo, were analyzed using the expected value of each spectrum. This was defined as the sum of all intensity values multiplied by their corresponding wavelength, divided by the sum of intensities, to extract quantitative information on their relevant spectral changes. The kinetic behavior of fluorescence signals was monitored by collecting fluorescence spectra, and integrating them across all wavelengths to produce an overall signal intensity, normalized to the initial intensity value as a function of collection time.

3. Results and discussion

3.1 In vitro measurements of dye fluorescence

The fluorescence spectra for the solutions described in Section 2.2, shown in Fig. 2(a) normalized to the peak intensity of the collected light.

 

Fig. 2 - (a) Normalized spectra collected in vitro in brain solutions for varied Rose Bengal concentrations (b) Expected wavelength values calculated for the data shown in Fig. 2(a) at different dye concentrations.

Download Full Size | PDF

The spectra red-shift and become progressively broader with decreasing dye concentration as the brain autofluorescence dominates. To extract quantitative information for the observed shift in spectral features the expected wavelength value (see Section 2.4) was calculated for the different dye concentrations (Fig. 2(b)). The expected values results show the potential of this technique to provide quantitative measurements of dye concentration in brain tissue.

3.2 Photochemical stroke

Results showed a positive correlation between the light intensity and size of the resultant infarct, as can be seen in Fig. 3 for 3 different test subjects.

 

Fig. 3 - Subcortical FOPS-induced lesions in the mouse brain (a) No infarction was observed in control mice injected with saline. Note the needle tract injury in the dorsal cortex caused by implantation of the fiber (arrowhead). (b) Injection of Rose Bengal coupled with low power illumination (12mW) resulted in a small lesion in the CA1 region of the hippocampus. (c) Injection of Rose Bengal coupled with high power illumination (43mW) resulted in extensive damage to the entire left striatum. Infarct locations are indicated by the dotted boundary. Scale bars = 1mm.

Download Full Size | PDF

For low excitation power (12 mW, 38 W/cm²) at the tip of the fiber, a mouse injected with Rose Bengal showed small to moderate stroke infarcts (0.32 mm²) in the targeted area of the hippocampus. When the excitation power was increased to 43 mW (137 W/cm²), the resulting infarct was much larger (1.37 mm²) as the light intensity is sufficiently high throughout a more extensive volume of brain tissue to cause dye activation. In this case, the infarcted area was greatly expanded such that ischemic damage was observed in surrounding brain regions including the ventral neocortex and striatum. Control mice injected with saline solution and exposed to the same conditions showed no infarct, confirming the photo-activation of the dye as the driving mechanism of photochemical stroke.

3.3 In vivo fluorescence measurements

Trials were performed on four test mice using a relatively low excitation power (15 mW, 48 W/cm²) as this power was shown to reproducibly produce a small infarct at the tip of the fiber. The autofluorescence background of the brain tissue, as seen from the in vitro trials at low dye concentrations, spectrally overlaps with the dye fluorescence. To differentiate between the two, a pre-bleaching step was incorporated for each mouse, whereby the mouse brain was illuminated with laser light for 10 min prior to the dye injection, enough time to completely quench the observed autofluorescence and consistent with the main illumination step for stroke infarct creation. The laser illumination was then blocked, and dye injected into the anesthetized mouse and was allowed to circulate for 10 min with the fiber still in place. Laser illumination was then resumed post dye injection for another 10 min and spectra were collected at 1 s intervals. Figure 4(a) shows the initial spectra recorded during the trials, showing brain autofluorescence spectra from the pre-bleaching step and a combination of autofluorescence and dye emission after dye injection. The contribution from dye fluorescence is clearly visible in the majority of the measurements taken post dye injection, providing confirmation of the ability to identify the Rose Bengal dye in vivo.

 

Fig. 4 - (a) Spectra collected for different in vivo trials in the case of pre-bleaching, showing brain tissue autofluorescence (dashed lines), and after dye injection (solid lines). (b) Expected values for the spectra recorded from the pre-bleaching (circles) and post-dye injection (diamonds) trials. Right axis: estimated dye concentration determined by in vitro experiments. Dashed horizontal lines represent the mean value across each group.

Download Full Size | PDF

Interestingly, by comparing the expected values for the two groups, it is possible to differentiate between them based on the shape of the recorded fluorescence. This is shown in Fig. 4(b), where the distinction between the two groups is shown. Most of the post-injection spectra expected values are below 630 nm and are separated as a group from the autofluorescence spectra that show expected values over 630 nm. Comparing the expected values for the in vivo dye spectra to those collected from pre-determined dye concentration in brain tissue pulp suggests a localized dye concentration, in the area of illumination at the beginning of each trial, of around 10⁻⁴ mg/ml. This is promising as it could allow for localized dye concentration measurements and their correlation to the properties of the generated stroke infarct. The kinetic behavior of the two sets of signals is shown in Fig. 5 (normalized to the initial intensity values) for the first 5 minutes of the measurements.

 

Fig. 5 - Kinetics of the total fluorescence intensity in the case of pre-bleaching, corresponding to brain tissue autofluorescence (dashed lines), and after dye injection, (solid lines) over the time of in-vivo trials.

Download Full Size | PDF

These results indicate that that the Rose Bengal dye is being bleached much more rapidly than the background autofluorescence (which results from the brain tissue as opposed to the injected dye), with the majority of the dye bleached within the first 60 s of the experiment, suggesting that associated photothrombotic events have been initiated within this timeframe. This is yet another means of discrimination between dye and autofluorescence based on decay dynamics.

4. Conclusion

In this work we have explored the development of photochemical stroke events in real-time using optical fibers to monitor changes in fluorescence intensity of the stroke-inducing dye Rose Bengal. These results offer, for the first time, an insight into the localized concentration (10⁻⁴ mg/ml) and the time scale (<60 s) over which the Rose Bengal-induced reaction occurs within the brain, in the order of 1 minute, significantly less than the duration used in the literature (>5 min). Future work will expand on the number of trials to establish statistical significance for the correlation between the dye concentration, the kinetics of this reaction and the size and rate at which a stroke infarct forms in the mouse hippocampus. We envisage that, with some modifications, this technique could be applied to the real-time measurement of dynamic physiological changes that occur in the brain as a result of ischemia. This technique could become a predictive tool to gauge whether a stroke had been successfully induced without needing to perform labor-intensive histology or expensive magnetic resonance imaging. We suggest this technology has the potential to advance our understanding of ischemic stroke in real time in vivo, which is fundamental to investigating new therapeutics in mammalian model systems.