A Decade of Blood-Brain Barrier Permeability Assays: Revisiting Old Traumatic Brain Injury Rat Data for New Insights and Experimental Design

Abstract

Increased microvascular permeability at the level of the blood-brain barrier (BBB) often leads to vasogenic brain edema following traumatic brain injury (TBI). These pathologic conditions compromise the integrity of the neurovascular unit resulting in severe brain dysfunction. To quantify this permeability and assess ionic equillibrium, preclinical researchers have relied on the use of various molecular weight permeable dyes such as Evans Blue that normally cannot enter the brain parenchyma under homeostatic conditions. Evans Blue, the most cited of the molecular weight dyes, has reported reproducibility issues because of harsh extraction processes, suboptimal detection via absorbance, and wide excitation fluorescence spectra associated with the dye. Our laboratory group transitioned to Alexa Fluor 680, a far-red dye with improved sensitivity compared to Evans Blue and thus improved reproducibility to alleviate this issue. To evaluate our reproducibility and increase the rigor of our experimental design, we retrospectively analyzed our controlled cortical impact (CCI) experiments over the past 10 years to evaluate effect size with larger samples and potential sources of variability. All of our BBB permeability experiments were performed with Male, Sprague Dawley rats weighing between 225–300 grams. Historically, Sprague Dawleys were randomly divided into treatment groups: SHAM, CCI, and a stem cell-based treatment from years 2007–2020. The assessment of microvascular hyperpermeability were evaluated by comparing the mean at minimum threshold, area at 1k-2k, and intensity density obtained from Alexa Fluor 680 permeability data. Studies utilizing Evans Blue were further compared by tip depth, diameter size, and the hemisphere of injury. Statistical evaluation utilizing the G Power software analysis did not yield a significant difference in sample size comparing experimental groups for Evans Blue and Alexa Fluor 680 analyzed brain tissue. Our analysis also demonstrated a trend in that recent studies (years 2018–2020) have yielded more compact sample sizes between experimental groups in Alexa Fluor 680 analyzed rats. This retrospective study further revealed that Alexa Fluor 680 image analysis provides greater sensitivity to BBB permeability following TBI in comparison to Evans Blue. Significant differences in sample size were not detected between Evans Blue and Alexa Fluor 680; there were significant differences found throughout year to year analysis at the lower range of thresholds.

Introduction

The Blood-Brain Barrier (BBB) is a dynamic interface and physiologic unit responsible for the maintenance of ionic homeostasis¹, characterized by a vascular network of brain capillaries held together by brain microvascular endothelial cells². Pathologic conditions to the central nervous system, such as traumatic brain injury (TBI), severely disrupt the molecular and structural integrity of the BBB³. Upon compromise, vasogenic edema occurs, accompanied by an influx of proinflammatory and neurotoxic macromolecules from the systemic circulation⁴. This inflammatory cascade initiates the conversion of glial cells to a reactive state that then proceeds to produce an abundance of proinflammatory cytokines, resulting in a chronic neuroinflammatory state⁵. This transition from healthy to a neuroinflammatory state has been extensively studied by our laboratory group and collaborators^6–10.

Clinically, BBB compromise is calculated as ratios between brain and serum total protein or albumin levels from patient fluid collected by either microdialysis catheters or external ventricular drains. However, to evaluate initial BBB dysfunction preclinically in animal models, dye extravasation using Evans Blue has been the preferred method for assessment¹¹. Evans Blue (EB) is a 961 Da molecular weight dye that is highly water soluble with a strong affinity to serum albumin (66 kDa)¹². Evans Blue quickly arose as the preclinical dye of choice after being reported to not significantly affect laboratory animals’ physiology, is relatively cost efficient compared to other commercial options, and its ease of quantification for statistical analysis¹³.

When Evans Blue is injected intravenously into laboratory animals, albumin-bound EB remains in the bloodstream under homeostatic conditions unless a defect in the structural or functional integrity of the BBB has occurred. If there is a defect, Evans Blue should be quantitively detected in the interstitial tissue. However, the use of Evans Blue has disadvantages based on the researcher’s experimental parameters. Disadvantages from using Evans Blue include low sensitivity and harsh tissue extraction. These disadvantages increase the likelihood of a false positives in Evans Blue quantification analysis to arise as unbound EB has been demonstrated to diffuse out of intact capillaries into the brain parenchyma when there is no BBB compromise¹⁴. Additionally, extraction of Evans Blue-stained tissue involves homogenizing the sample tissue, a destructive process that prevents further histologic analysis. Also, extraparenchymal EB dye cannot be easily excluded from intraparenchymal dye when placed under an absorbance microplate reader (Fig. 1), likely contributing to the high false positive rate with this technique¹⁵. Collectively, these disadvantages represent a reproducibility issue. This issue is not alleviated even with the accounting of biologic variability, as readings between treatment groups may vary significantly between experiments, within the same lab group or by the same technical expert.

Figure 1. Schematic of tissue analysis differences in Evans Blue and Alexa Fluor 680 for BBB permeability experiments.

For Evans Blue analysis: Evans Blue in phosphate-buffered saline (PBS) was injected via tail vein. 60 minutes after injection, Sprague-Dawley rats were euthanized via right atrial puncture and perfused with 4% paraformaldehyde (PFA). The brain was isolated and divided through the midline and the mass of each hemisphere measured (ipsilateral and contralateral to injury). Tissue was then homogenized and incubated overnight to allow for dye extravasation. 100 μL of solution from each sample was transferred to a 96 well plate (in triplicate) and absorbance was measured at 620 nm. All values were normalized to hemisphere weight. Alexa Fluor 680 analysis: Alexa Fluor 680 dye conjugated to 10kDa dextran in PBS was introduced via tail vein injection. Rats were allowed to recover for 60 minutes, then sacrificed via right cardiac puncture and perfused with ice cold PBS followed by 4% PFA. Brains were explanted and placed in 4% PFA and coronally sectioned into 2 mm slices. Slices were imaged using a LI-COR Odyssey CLx infrared laser scanner (LI-COR, Lincoln, Nebraska) at 700 and 800 nm.

To circumvent these pitfalls, many preclinical laboratory groups have made the transition from Evans Blue to fluorescent markers, such as the Alexa Fluor series (Thermo Fisher Scientific, Waltham, MA, USA) for their permeability studies. The Alexa Fluor dyes are able to conjugate to a variety of molecules and have been used to analyze drug transit across the BBB.^16–19 We have previously demonstrated that the implementation of Alexa Fluor 680 for BBB permeability analysis results in increased sensitivity to BBB perturbation by reducing non-specific signals, thereby allowing identification of specific sites of microvascular injury²⁰.

Even with the advantages of the Alexa Fluor dyes compared to the traditional Evans Blue, there are still limitations to the use of Alexa Fluor²¹. While Alexa Fluor dyes allow for whole brain slice imaging and image analysis, the slicing matrix requires the use of slices that are thick enough to prevent some fluorescence light from fully penetrating each section. Our laboratory has previously standardized our imaging to produce 6–2 mm depth slices into the brain²². This ensures technique reproducibility even with routine changes in research personnel. This reproducibility is critical for the conclusions derived from our group and collaborators, as well as comparison to other experiments for treatment efficacy evalaution²³.

The NIH has placed a growing focus on the need for rigorously designed, published preclinical studies, to ensure that such studies can be replicated and produce translational results²⁴. Reproducibility in research findings are the cornerstone of the scientific method and fundamental for the ethical justification of in vivo research²⁴. Too often, failure to replicate preclinical data produces translational failure. Proper experimental design is imperative to obtaining sound results and preventing unwanted bias²⁵. Good experimental design pays attention to details such as validity, power, and sample size²⁶. These details are necessary to justify the number of animals needed in experiments and ensure reproducible results²⁶.

To evaluate rigor and examine the reproducibility in our research group, we systematically evaluated our in vivo BBB permeability experiments over the past 10 years. In this retrospective study, we highlight a comparison of BBB permeable dyes and changes in sample size over years of preclinical experimental study. It is our goal that this study will serve as a frame of reference for preclinical researchers evaluating BBB permeability with Alexa Fluor 680 and/or Evans utilizing the controlled cortical impact (CCI) model of TBI in rats.

Materials and Methods

Animal Protocols

All protocols involving the use of animals followed the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by The University of Texas Health Science Center at Houston Animal Welfare Committee (Institutional Animal Care and Use Committee).

Study Design

All animal data was derived from male, Sprague-Dawley rats (Charles River, Houston, TX) weighing 225–300 grams. Rats received either 1 mL of 3% EB (Thermo Fisher Scientific, Waltham, MA, USA) or 1 mg/kg Alexa Fluor dye (Thermo Fisher Scientific, Waltham, MA, USA) conjugated to 10 kDa dextran in phosphate-buffered saline (PBS) intravenously at time of euthanasia.

Injuries: Controlled Cortical Impact and SHAM

Controlled cortical impact (CCI) devices were used to administer unilateral brain injuries as previously described by Lighthall²⁷. Briefly, all rats were shaved and mounted on a secured stereotactic frame in the horizontal plane. After anesthetization with 4% isoflurane in oxygen, a midline incision was made and extended in depth to the skull with a 7–8 mm craniotomy. The center of the craniotomy was placed at the midpoint between bregma and lambda, ~3 mm lateral to the midline, overlying the temporoparietal cortex. Animals received a single impact at either 2.7- or 3.1-mm depth of deformation with an impact velocity of 5.8 m/s and a dwell time of 150 ms (medium severity) or 200 ms (high severity). SHAM injuries were performed by anesthetizing the rats, making the midline incision, and separating the skin, connective tissue, and aponeurosis from the cranium. The incision was then closed as previously described. Experimental injuries before 2013 and 2013 and after were performed with different CCI devices (before 2013: eCCI Model 6.3; VCU, Richmond, VA, USA; 2013 and later: Leica Impact One Stereotaxic Impactor, Buffalo Grove, IL, USA).

Evans Blue BBB Permeability Analysis

Evans Blue permeability analysis was performed as previously described by Bedi et al⁸. Briefly, 1 mL of 3% Evans Blue dye in PBS was introduced by tail vein injection. Animals were given 60 minutes for recovery to allow for perfusion of the dye. After this time, animals were sacrificed via right atrial puncture, perfused with 4% paraformaldehyde (PFA), and decapitated for brain extraction. After cerebellum dissection, the brain was divided through the midline, and the mass of each hemisphere (ipsilateral to injury and contralateral to injury) was measured for normalization. Subsequently, each hemisphere was incubated overnight in 5 mL of formamide for dye extraction. After homogenization and centrifugation, 100 μL of the supernatant from each sample was transferred to a 96 well plate (in triplicate), and absorbance was measured at 620 nm using the VersaMax plate reader (Molecular Devices Inc., Sunnyvale, CA) (Fig 1).

Alexa Fluor 680 BBB Permeability Analysis

Thirty minutes before sacrifice, 1mg/kg of Alexa Fluor 680 dye conjugated to 10 kDa dextran in 1 ml of PBS was introduced via tail vein injections. Animals were allowed to recover for 60 minutes, then sacrificed via right cardiac puncture, and perfused with ice-cold PBS followed by 4% PFA (Fig 1). Explanted brains were then placed in 4% PFA for 1 hour and transferred to PBS. The brains were then sectioned coronally into 2 mm slices using a rat brain matrix slicer (Zivic Instruments, Pittsburgh, PA). Eight anatomically standardized slices encompassing the area of injury were then placed on a plastic petri dish using a grid and imaged using the LI-COR Odyssey CLx infrared laser scanner (LI-COR, Lincoln, NB) at 700 (Alexa Fluor 680 signal) and 800 nm (background signal). The background image was subtracted from the Alexa Fluor signal image to produce the assay results.

Data Analysis

G*Power (version 3.1.9.4) was used to determine sample size for SHAM and CCI groups in all studies^28,29. The threshold for statistical significance was P≤ 0.05. All other statistical analyses were performed using Graph Pad PRISM (version 7.08, Graph Pad Software Inc., La Jolla, CA, USA). All data are represented as mean ± SD. Comparisons between means of each group were made using one-way ordinary analysis of variance (ANOVA) followed by Tukey’s post hoc test, except for comparison of SHAM and CCI groups under Alexa Fluor 680 (AF680) which was analyzed using a two-tailed t test. P < 0.05 was considered significant.

Results

Experimental Design

Using an a priori calculation, sample size was calculated for studies involving Evans Blue and Alexa Fluor 680 at the minimum threshold. After analysis, an overall sample size of 13 was determined to be the minimum number of rats required for studies using Alexa Fluor 680 and Evans Blue (Fig 3). Experimental parameters (tip and depth size) for CCI also changed during the course of these experiments. Sample size for Evans Blue studies with a depth size of 2.7mm depth and a 5mm tip was determined to be 12 per treatment group. Increasing the depth to 3.1mm, yielded an increase in variability requiring 15 per experimental treatment groups. Studies utilizing Evans Blue also evaluated optimal tip size. A depth of 2mm with a 6mm tip size yielded a sample size of 4 between experimental groups treatment group.

Figure 3. Timeline of experiments utilizing Evans Blue and Alexa Fluor 680 for BBB permeability analysis.

(A). Equipment and personnel changes during the past 12 years of experimentation with the CCI model in our lab group. (B). Calculated minimal sample size per year after retrospective analysis by BBB permeability assay.

BBB Permeability Assay Analysis – Evans Blue

To determine the extent of injury, dye extravasation was used as a measure of leakage into the brain parenchyma. BBB permeability measurements were initially completed using Evan’s Blue for SHAM and CCI analysis. Figure 4 demonstrates Evans Blue characterized by tip size (Fig 4A) and contralateral or ipsilateral to injury site (Fig 4B). In studies conducted with a 5 mm tip size, there was a significant difference between SHAM (0.09 ± 0.02) and CCI of 2.7 mm depth (0.24 ± 0.15). There was also a significant difference between SHAM (0.09 ± 0.02) and CCI injury of 3.1 mm depth (0.24 ± 0.17). There was no significant difference between the 2.7 mm (0.24 ± 0.15) and 3.1 mm depth size using a 5 mm tip (0.24 ± 0.17). In studies conducted using a 6 mm tip, there was a significant difference between the SHAM (0.09 ± 0.03) and 2.7 mm depth size CCI groups (0.16 ± 0.02). There was also a significant difference between SHAM (0.09 ± 0.02) and CCI injury of 3.1 mm depth (0.59 ± 0.25). In addition, there was significant differences between the 2.7 mm and 3.1 mm depth sizes using a 6mm tip size. Studies when data was compared from being ipsilateral to the injury revealed a significant difference between SHAM (0.07 ± 0.04) and CCI (0.11 ± 0.06). Studies comparing contralateral to the CCI injury also revealed a significant difference between SHAM (0.10 ± 0.06) and CCI (0.41 ± 0.20).

Figure 4. Evans Blue data breakdown by depth, tip size, and hemisphere.

(A). Overall Evans Blue data characterized by tip and depth size. 5mm tip SHAM (n=50), 2.7mm depth (n=46), and 3.1mm depth (n=43). 6mm tip SHAM (n=34), 2.7 mm depth (n=11), and 3.1mm depth (n=5). (B). Overall Evans Blue data characterized by contralateral or ipsilateral to injury. The Y axis OD/Wt refers to optical density (absorbance) per rat weight. Contralateral Evans Blue SHAM (n=17) and Evans Blue CCI (n=40). Ipsilateral SHAM (n=17) and CCI (n=40).

BBB Permeability Assay Analysis – Alexa Fluor 680

BBB Permeability analysis after Evans Blue extravasation transitioned to Alexa Fluor 680. The mean minimum signal for CCI rats was 1112.1 ± 402.75 was significantly greater than SHAM 807.27 ± 208.67, which was significant (Fig 2A). For the CCI animals, the CCI rats the mean area at the 1K threshold was 54450 ± 5469.8 while the mean area for SHAM rats was 94723.41 ± 58805.22 (Fig 2B). At the 2k threshold, the mean area to the signal for SHAM rats was 3882.9 ± 5455.5 while the mean area for CCI rats 33105 ± 16602, which was significant between the two groups (Fig 2C). The integrated density of CCI rats at the 1K threshold was determined to 7.6654 × 10⁷ ± 8.6317 × 10⁷ while SHAM rats were determined to be 2.1771 × 10⁸ ± 1.4973 × 10⁸, this comparison revealed significant between the two groups (Fig 2D). At the 2K threshold, the mean intensity of the integrated density for SHAM rats were determined to be 8.2261 × 10⁶ ± 1.3924 × 10⁶, while CCI rats were determined to be 1.2651 × 10⁸ ± 7.5416 × 10⁷ (Fig 2E).

Figure 2. Global statistics of Alexa Fluor 680 usage breakdown.

Significant differences in integrated signal between CCI and SHAM was identified in the low + narrow threshold range of 1K–2K. Column informational statistics are demonstrated with (A.) mean at minimum threshold (SHAM n=94 and CCI n=118), (B.) area at 1K (SHAM n=71 and CCI n=93), (C.) Area at 2K (SHAM n=44 and CCI n=54), (D.) Integrated density at 1K (SHAM n=90 and CCI n=122) and (E.) Integrated density at 2K (SHAM n=34 and CCI n=34). Data is presented as SEM. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001 (Two tailed t-test).

Non-Specific Signal Reduction with Alexa Fluor 680

Alexa Fluor 680 reduces non-specific signals between CCI and SHAM groups (Fig 5) by 37%, a fourfold reduction. High signal in SHAM group of Evans Blue rats is attributed to intraparenchymal dye that could not be excluded from the microplate reader.

Figure 5. Blood Brain Barrier Permeability Assays.

Non-specific signals are reduced with Alexa Fluor compared to Evans Blue dye as evidenced by the high signal from SHAM animals receiving Evans Blue. The Y axis OD/Wt refers to optical density (absorbance) per rat weight. The X axis Integrated Density refers to the cumulated signal. Alexa Fluor SHAM (n=94), Alexa Fluor CCI (n=118), Evans Blue SHAM (n=50), Evans Blue CCI (n=105). Data is presented as SEM. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001 (Ordinary One-way Anova).

Alexa Fluor 680 BBB Permeability Visualization

Coronal 2mm brain slices are arranged rostral to caudal (left to right) and imaged using the Licor Odyssey. False colored images were overlaid to display the Alexa Fluor 680 (red) against non-specific fluorescence detected in the 800 channel (gray) (Fig. 6A). The distribution of Alexa Fluor 680 can be processed to false colorize different intensity bins to assist with visualizing the localization of BBB damage. In Fig 6B, we have applied a common simple pseudo color scheme to separate non-specific background fluorescence (blue) from low intensity staining that may include diffuse vascular leakage (green). Meanwhile, the injury penumbra features more moderate Alexa Fluor 680 signals (yellow) and the center of the lesion has an intense fluorescence (red).

Figure 6. Representative montages of CCI and SHAM rats.

A representative scan of Alexa Fluor 680 10,000 MW dextran dye extravasation from an experiment with an n=5. Brains were sliced into six 2mm sections and arranged rostral to caudal (left to right) in rows. A. The Alexa Fluor 680 dye is imaged and colorized according to intensity, where blue represents background auto fluorescence (RFU 257–1031), green represents diffuse low levels of fluorescence (RFU 1032–2063), yellow represents moderate intensity (RFU 2064–5031) and red represents high fluorescence intensity staining (RFU 5032–65535). B. A montage containing the two experimental groups after merging the two detection channels (700nm - Alexa 680 and 800nm - background). Red areas represent presence of Alexa Fluor 680 dye.

Histogram Analysis of Alexa Fluor 680 Distribution

In some recent studies, we have made an effort to use quantitative analysis to preferentially analyze different injury components. To this end, we have begun analyzing different quantitative aspects of the image across the histogram. By plotting the differences in the area of intensity, mean fluorescence, and the integrated density as a function of an increasing threshold, we can start to easily see where quantitative differences between CCI and SHAM injured rats are occurring at both low and high fluorescence intensities in each measurement (Fig 7). The differences between groups can be used to quantitatively define intensity ranges and measurements that are associated with specific injury processes and can be visualized directly (as in Fig 6B).

Figure 7. Plot Profile Analysis.

Following imaging, the extravasation of Alexa Fluor 680 was analyzed for both intensity and distribution using quantitative image analysis. The area, integrated density, and mean fluorescence (top to bottom) of CCI and SHAM-injured rats was measured as a function of an increasing minimum threshold by segmenting the histogram from each brain into 256 bins (left column). The contribution of background auto fluorescence is readily visible in the initial decrease of the Area as the intensity decreases, as visible in the difference between the blue and green colors in Figure 6A. As the minimum threshold pixel intensity increases, the difference between the CCI and SHAM injury becomes clear from diffuse low-intensity staining across the brain, most likely associated with microvascular BBB leakage. As the lower threshold reaches higher intensities, analysis becomes limited to the high intensity fluorescence areas localized to the focal injury. The difference between the CCI and SHAM injured groups (right column) further indicates the intensity ranges where each measurement (area, integrated density, and mean fluorescence) are most different between the groups.

Discussion

In this study, we performed a retrospective analysis to investigate the rigor and reproducibility of the BBB permeability assays performed in our laboratory over the past 10 years by comparing all SHAM and CCI animal experimental groups. The NIH has defined scientific rigor as ‘the strict application of the scientific method to ensure robust and unbiased experimental design, methodology, analysis, interpretation and reporting of results. This application extends into full transparency in reporting experimental details so that others may reproduce and extend the findings³⁰.” This is a critical component for the translation of bench side science into bedside applications. As a translational research group, we took a retrospective approach to evaluate and determine if refinement in sample size by evaluating our previously published data was needed to further strengthen our experimental design. Our experimental design has typically consisted of three treatment groups: SHAM, CCI injury, and a variable stem cell therapy-based treatment group^10,31–34. As stem cell treatment groups over the years have changed, we decided to highlight and focus attention to the SHAM and CCI groups as the parameters of reproducibility. The SHAM and CCI groups are incorporated into every experimental design, with SHAM serving as the control group and CCI as injury, which has been highlighted for its reproducibility^35–38.

In preclinical models of TBI, the CCI model is one of the most commonly employed techniques used to mimic in vivo TBI due to the flexibility and reproducibility of the injury parameter. In particular, the effects of CCI have been well categorized in rats^38,39. This model has demonstrated diverse histopathological and functional changes consistent with what occurs in clinical TBI cases, including but not limited to: vasogenic edema, neuroinflammation, and dysfunction of the BBB. We have previously standardized the injury parameter (e.g., depth, tip size, anesthesia type) based on the injury severity to facilitate our research goal.

Initially, we analyzed increases in sample size over time (Fig 3), we noted significant increases in the minimum required sample size needed to reach significance whenever the CCI impactor has undergone more than 3 years without calibration and servicing. For laboratory equipment to be useful for experiments that are underway it has to be reliably functioning and suitably calibrated. Even with accounting for changes in laboratory personnel, we noticed that after calibration years (2011 & 2016) decreases in necessary sample size occurred in both our Evans Blue and Alexa Fluor 680 data. Previously, we have utilized Evans Blue for analysis of BBB permeability dysfunction¹⁰. Before the widespread use of dextrans, sodium fluoresceins, and radiolabeled ligands, Evans Blue was the gold standard for evaluating BBB dysfunction in our laboratory and others. As extensively outlined in Saunders et al 2015, Evans Blue is an unsatisfactory tool for truly assessing BBB dysfunction. The reason is that Evans Blue analysis can only offer global insight into the aggregate injury to the microvasculature. Evans Blue may be ideal for looking at an initial disruption, but provides little insight into the other complex pathophysiological aspects of TBI. As Evans Blue forms a conjugate-complex with albumin, the determination of pathophysiologic effects is limited to the size of albumin²². This limits the translational relevance of this dye choice for assessment. Examining our Evans Blue data, we also discovered that there was a significant change in permeability when comparing the depth of the injury and tip size of the impactor.

Recommended CCI parameters for rats are an impactor tip diameter that ranges from 5–6 mm, with a dwell time of 50–250 msec and a depth of 1–3 mm at the parietal cortex and midline³⁷. Most of our studies utilizing Evans Blue have been performed with a 5 mm tip (Fig 4A). When protocols switched depth size from 2.7 mm to 3.1 mm, there was a trend that demonstrated variability among the injury. Increasing the tip size to 6 mm did not produce an increase in variability at the 2.7 depth size as demonstrated in the 5 mm size. These transitions, however, did not result in any significant differences in the extent of BBB breakdown or functional assessments in some of our previously reported studies^{8,9,31–33,40,41}.

When the laboratory reached a consensus that Evans Blue may be unsatisfactory for the translational aspects of our research, Alexa Fluor 680 was selected as the replacement. There are numerous other choices of dyes such as fluorescein isothiocynate (FITC) that also serve as excellent models for assessing BBB damage. Alexa Fluor 680 was chosen because of the potential for further tissue analysis, is pH insensitive (over a broad pH range) and tends to be brighter compared to FITC on most instruments. In addition, since Alexa Fluor is not endogenous and cannot be confused with endogenous molecules used in IgG extravasation, it can be introduced at various times throughout the life cycle of a disease in order to determine the time of BBB disruption and in turn, provide critical information regarding metrics of the disease state.

Since brain tissue is not homogenized, brain architecture is still intact allowing for immunohistochemistry and further analysis with other cell surface markers when utilizing Alexa Fluor 680. Alexa Fluor 680 also allowed for greater sensitivity to BBB perturbation displaying at-risk regions that can serve as potential therapeutic targets. These principles of identifying areas of BBB compromise and edema are ideal as they can be translated into clinical practice⁴².

Analysis of the Alexa Fluor 680 data has demonstrated that our group has increased the precision of our injury model in recent years (Table 1). This injury data demonstrates that a decreased sample size is required in order to reach statistical significance (9 per sample group, 2020) compared to earlier years (22 per sample group, 2015). This is likely attributed to regular calibration of the impactor and retained personnel used for training sessions²⁵. This has helped our operative procedures become standardized and account for potential variability with personnel changes. This standardization helps highlight the advantages of using Alexa Fluor 680. We standardized our technique by determining which 8 brain slices are to be imaged. Slice determination is performed by displaying all slices and immediately outlining the three slices that contain the mammillary nucleus and media eminence. The remaining slices are obtained from the three preceding posterior and two anterior proceeding slices²². With this optimization, this has led to widespread reproducibility our laboratory group, which variation most likely being attributed to the initial removal of the brain from the skull. This together with the functionality of Alexa Fluor 680 has made it an ideal marker for truly assessing the extent of BBB permeability in comparison to Evans Blue.

Table 1.

Global statistics of Alexa Fluor 680 usage in experimental studies from years 2018–2020

	Mean at Min Thresh	Between Experimental Groups	Area at 1k	Between Experimental Groups	Area at 2K	Between Experimental Groups	Int Den at 1K	Between Experimental Groups	Int Den at 2K
Mean	SHAM: 854.56 CCI: 1220.7	SHAM: 854.05 CCI: 1225.4	SHAM: 69548 CCI: 1.2113 × 105	SHAM: 67851 CCI: 1.1931 × 106	SHAM: 3832.3 CCI: 35556	SHAM: 3990.8 CCI: 36521	SHAM: 1.1245 × 106 CCI: 2.8423E+08	SHAM: 1.1394 × 106 CCI: 2.8563E+08	SHAM: 6.62072 × 106 CCI: 1.41E+08
Median	SHAM: 854.017 CCI: 1186.8	SHAM: 840.79 CCI: 1182.4	SHAM: 61879 CCI: 1.2643 × 106	SHAM: 82889 CCI: 1.2054 × 105	SHAM: 2468 CCI: 33900	SHAM: 2829.25 CCI: 41786.2	SHAM: 83757980 CCI: 2.67E+08	SHAM: 1.09 × 108 CCI: 2.5E+08	SHAM: 4.9698 × 106 CCI: 1.29E+08
Lower CI	SHAM: 798.7286217 CCI: 1.038 × 105	SHAM: 775.1728 CCI: 1023.9	SHAM: 42623 CCI: 96154	SHAM: 30470 CCI: 97059	SHAM: 1393.9 CCI: 23462	SHAM: 626.90 CCI: 22088	SHAM: 5.7933 × 108 CCI: 2.0506 × 108	SHAM: 1.6799 × 107 CCI: 1.8806 × 108	SHAM: 2.7595 × 108 CCI: 7.3689 × 108
Upper CI	SHAM: 910.39 CCI: 1402.5	SHAM: 932.94 CCI: 1426.9	SHAM: 96474 CCI: 1.4609 × 105	SHAM: 105232.5 CCI: 141203.4	SHAM: 6270.6 CCI: 47651	SHAM: 7354.7 CCI: 50955	SHAM: 166980396.4 CCI: 3.63E+08	SHAM: 228052242.5 CCI: 3.83E+08	SHAM: 10481940.6 CCI: 2.09E+08
CV	SHAM: 10.767 CCI: 24.19	SHAM: 7.0119 CCI: 12.485	SHAM: 63.817 CCI: 33.465	SHAM: 41.825 CCI: 14.065	SHAM: 104.88 CCI: 55.214	SHAM: 63.993 CCI: 30.004	SHAM: 79.920 CCI: 45.239	SHAM: 76.031 CCI: 25.947	SHAM: 96.134 CCI: 77.616
Effect Size	1.6747	3.1963	1.2135	2.2004	0.5	4.0888	1.5456	2.1265	1.7337
Power Analysis Sample Size	16	8	32	12	176	6	20	12	16

Abbreviations in table are as listed in table 1. CI: Confidence Interval, CV: Coefficient of Variation, Mean at Min Thresh: Mean at the minimum threshold, Int Den: Integrated Density.

As previously stated, Alexa Fluor 680 functionality allows for the identification of differential degrees of BBB permeability, a great feature in preclinical TBI evaluation. This publication demonstrate that Alexa Fluor 680 is sensitive, specific and provides a platform for brain slice imaging and analysis (Fig 5). This technique identifies regional changes in BBB permeability and may be used to track the response of at-risk regions to potential therapeutics. The image analysis process has had its limitations that can occur with the threshold selection process. The threshold selection process can be a significant limitation due to subjectivity, but we standardized our process by using predetermined anatomic parameters such as penumbra selection or criteria that have been applied across experiments since Alexa Fluor 680 incorporation. We have determined that the low + narrow threshold range identified potential penumbral areas and showed significant difference in these areas compared to CCI rats. These findings previously determined that less penumbral BBB permeability occurs in treated rats versus rats with CCI injury alone²². Intensity range likely represents tissue with significant dysfunction in permeability regulation or devitalized tissue as seen with intraparenchymal contusions or microvascular hemorrhage.

In this study, we also demonstrate how we can further analyze images of Alexa Fluor 680 to understand the changes in intensity profile to evaluate changes in microvascular permeability as well as restrict analysis to the site of the lesion using unbiased histogram analysis. Differences were found in areas over threshold, integrated density over threshold, and the mean fluorescence over threshold between the CCI and SHAM can be visualized by false coloring the images to evaluate the distribution of Alexa Fluor 680 across different structures of the brain (Fig 6). Furthermore, the relative differences in the areas of microvascular permeability and the larger vascular leaks associated with the direct lesion can be determined using segmented histogram analysis (Fig 7) which can be followed by additional analysis using data-defined thresholds. This information can be correlated as a focal point for identifying areas of potential therapeutic intervention.

Our results depict a pattern: there were several years of underpowered studies observed in our previous research pattern. These underpowered years are likely attributed to changes in impactor, lack of impactor calibration, and initial variances with personnel changes. Evans Blue is a beneficial permeability assay when focusing only on the initial disruption. Alexa Fluor 680 is a superior alternative for identifying the extent of injury and also allows for better internal validation for preclinical experimentation.

Figure 8. Utilities and applications of repeated assay data.

The potential applications of this repeated assay data to the preclinical researcher. This data can be utilized for internal and external comparison to other labs performing the CCI injury model. The major utility of this data is to increase the rigor of experimental design by minimizing the number of animals and evaluating the reproducibility between experiments.

Highlights.

• Alexa Fluor 680 provides better sensitivity to injury perturbations

• Retrospective analysis of the CCI model in Sprague-Dawley rats

• In vivo BBB microvascular permeability analysis

• Evans Blue comparison of BBB permeability

Summary Statement:

This work provides a comparative analysis of BBB permeability assay techniques after CCI model of injury in rats.

Abbreviations

AF680

Alexa Fluor 680

BBB

Blood Brain-Barrier

CCI

Controlled Cortical Impact

EB

Evan’s Blue

PFA

Paraformaldehyde

TBI

Traumatic Brain Injury

TNF

Tumor Necrosis Factor

Availability of data and materials

The data for this manuscript will be made available upon reasonable request.