Abstract
Background
Given the current lack of widely adopted strategies for facilitating drug penetration into the brain, developing new techniques to increase blood-brain barrier (BBB) permeability is essential to address the increasing burden of central nervous system disorders. Osmotic blood-brain barrier opening (OBBBO), achieved through intra-arterial delivery of 25% mannitol to the cerebral vasculature, is a pioneering strategy demonstrating both safety and partial efficacy.
Purpose
To investigate the potential of 25% mannitol with 4% NaCl, a combination that doubles the osmotic power, to safely increase OBBBO efficacy.
Materials and Methods
To visualize penetration into the brain, OBBBO with intra-arterial, intravenous, or intraperitoneal infusion of molecules was performed in mice (n = 44). Brain penetration of small molecules (gadolinium-based contrast agent) and the safety of the infusion procedure were assessed via MRI and histologic examination. Brain penetration of large molecules (zirconium 89–radiolabeled antibodies) was assessed via PET imaging. A head-to-head comparison was performed of brain penetration of both these molecules using the standard approach (25% mannitol) versus the combination of 25% mannitol with 4% NaCl. The Wilcoxon signed rank test, Mann-Whitney U test, and mixed-model analysis were used for the statistical analyses.
Results
The findings showed robust OBBBO across the entire targeted hemisphere of mouse brains following administration of 4% NaCl in 25% mannitol (n = 9), surpassing that achieved with 25% mannitol alone (n = 8) (ratio of OBBBO area to intact BBB area, 1.99 ± 0.17 vs 1.41 ± 0.15; P < .001). Follow-up MRI and postmortem histologic examination confirmed the safety of this procedure. PET imaging revealed that brain uptake of radiolabeled antibodies was significantly improved using this new method.
Conclusion
Combining 4% NaCl with 25% mannitol showed promise for improving the safe delivery of therapeutic agents to the brain through the endovascular route.
© RSNA, 2025
Supplemental material is available for this article.
See also the editorial by Clement and Gaultier in this issue.
Summary
MRI and PET showed that the combination of two osmotically active agents, 25% mannitol and 4% NaCl, safely increased blood-brain barrier permeability to therapeutic antibodies.
Key Results
■ In a study of 44 mice, contrast-enhanced MRI showed that a combination of 25% mannitol and 4% NaCl opened the blood-brain barrier (BBB) more effectively than 25% mannitol alone.
■ Sequential T2- and T1-weighted MRI, along with histologic examination (hematoxylin-eosin, glial fibrillary acidic protein, ionized calcium-binding adaptor molecule 1, and NeuN), confirmed the excellent safety of the procedure for up to 7 days.
■ PET imaging of radiolabeled bevacizumab delivered after osmotic BBB opening using 25% mannitol and 4% NaCl demonstrated significantly higher uptake in the targeted brain hemisphere, confirming the high potential of the formulation to improve the delivery of therapeutic antibodies to the brain.
Introduction
The blood-brain barrier (BBB) is a formidable obstacle to effective treatment of central nervous system diseases (1–6). Many studies have aimed to find a safe and effective method to temporarily disrupt the barrier to enhance drug delivery (Table). Osmotic BBB opening (OBBBO) with an intra-arterial (IA) infusion of 25% mannitol is one strategy for transiently enhancing drug penetration into the central nervous system (7,8). This approach was quickly adopted in clinical practice and has proven to be safe (9), although variability in drug penetration has been reported (10). One clinical trial showed favorable results for IA chemotherapy in conjunction with OBBBO in patients with malignant brain tumors (11). Extensive data from more than 5000 procedures performed over 20 years support the safety of this procedure (11,12). Nonetheless, concerns have been raised regarding the fast infusion rate and nonselective delivery of osmotic agents, which could lead to maculopathy (13). In addition, preclinical studies in small animals have suggested that rapid infusion of osmotic agents may contribute to neuroinflammation, which may be missed in clinical oncological settings (14).
Comparison of BBB Opening Methods
| Method | Description | Advantages | Disadvantages |
|---|---|---|---|
| Osmosis | Intra-arterial administration of a hyperosmolar solution (usually mannitol) to shrink endothelial cells and open tight junctions | Well-established technique; relatively simple procedure; wide openings for large molecules; cost-effective | Requires catheter placement inside arteries; image guidance may improve precision and spatial specificity |
| Focused ultrasound | Use of acoustic energy combined with intravenous microbubbles to create temporary openings in the BBB | Noninvasive; precise spatial targeting; reversible effects; real-time monitoring possible with MRI; can be repeated multiple times | Requires specialized equipment; high initial cost; may cause minor tissue damage if parameters are not optimized; limited to accessible brain regions |
| Chemical agents | Use of compounds such as bradykinin, histamine, or alkylglycerols to modify tight junctions | Can be administered systemically; multiple agents available; moderate cost; well-studied mechanisms | Limited spatial specificity; variable effectiveness; potential inflammatory responses; duration of opening hard to control |
| Electromagnetic radiation | Application of electromagnetic fields or radiation to increase BBB permeability | Noninvasive; can cover larger areas; no chemical agents required; potentially repeatable | Limited depth penetration; possible thermal effects; limited equipment availability; radiation-induced carcinogenesis |
| Biologic agents | Use of biologic molecules that bind to the endothelial surface to shuttle molecules across BBB | Natural transport mechanisms; high specificity possible; no BBB opening per se | Immune response risks; complex manufacturing; high development costs; toxicity related to targeting endothelial molecules |
Note.—BBB = blood-brain barrier.
Recent work has further enhanced the precision, safety, and efficacy of OBBBO in the context of a superselective approach. This work has shown the critical role of real-time MRI guidance in predicting the region in which OBBBO occurs (15), and MRI-guided OBBBO has been successfully implemented twice in the same patient in a clinical setting (16). Studies have shown that using a relatively slow IA infusion rate prevents detrimental changes to murine brain tissue (17). However, this slower infusion rate significantly reduces the spread and efficacy of OBBBO (18). Therefore, strategies have been sought to safely enhance the efficacy of OBBBO.
To date, OBBBO procedures in both preclinical and clinical settings have primarily used 25% mannitol as the osmotic agent. However, the solubility limit of mannitol prohibits further increases in osmotic power. Thus, the potential benefit of combining 25% mannitol with other osmotically active substances has been investigated in pilot experiments. Given the widespread clinical use of hypertonic saline, it emerged as the natural first choice to comply with regulatory requirements for the clinical translation of new techniques. Specifically, 4% NaCl was selected because it is the most frequently used clinical formulation of hypertonic saline for intravenous (IV) infusions.
The aim of this study was to evaluate BBB permeability using gadolinium-based contrast agent (GBCA; a representative small molecule) and antibodies (a representative large molecule and also the most commonly registered class of biologic drugs).
Materials and Methods
Study Design
All procedures were approved by the Institutional Animal Care and Use Committee at the University of Maryland, Baltimore. The study design involved IA, IV, and intraperitoneal infusions of molecules in mice (n = 44 for the entire study) and visualization of the penetration of these molecules into the brain (Fig 1). There were three parts of the study design. First, there was an investigation of the degree of brain penetration of GBCA (a small molecule) after OBBBO and assessment of the safety of the infusion procedure. For this purpose, the following procedures were compared: OBBBO with 25% mannitol plus intraperitoneal GBCA (eight mice) and OBBBO with 4% NaCl in 25% mannitol plus intraperitoneal GBCA (nine mice). Second, there was an assessment of the brain penetration of antibodies (large molecules) after OBBBO. For this purpose, the following procedures were compared: OBBBO with 25% mannitol plus IA radiolabeled antibody (eight mice) and OBBBO with 4% NaCl in 25% mannitol plus IA radiolabeled antibody (eight mice). Third, IA (six mice) and IV (five mice) routes of delivery of radiolabeled antibodies following OBBBO with 4% NaCl in 25% mannitol were compared. OBBBO procedures were performed as previously described (18). OBBBO was performed on the right hemisphere.
Figure 1:
Diagram of the study design. BBB = blood-brain barrier, BBBO = BBB opening, 89Zr-BV = zirconium 89 bevacizumab-deferoxamine, IA = intra-arterial, IP = intraperitoneal, IV = intravenous.
GBCA distribution within the brain parenchyma was assessed using MRI. Antibody biodistribution was assessed using a previously established method that involved PET/CT after injection of conjugated bevacizumab-deferoxamine radiolabeled with zirconium 89 (89Zr). The percentage of injected dose per gram of tissue and standardized uptake value (SUV) for selected organs and regions were determined. Follow-up MRI on days 3 and 7 and postmortem histologic examination on day 7 were used to detect potential adverse effects and evaluate overall safety. Further methodological details can be found in Appendix S1.
MRI Acquisition and Analysis
MRI was performed using a Bruker 9.4-T small animal MRI scanner. T1-weighted and T2-weighted images were acquired in an axial plane in an anteroposterior readout orientation with a section thickness of 0.7 mm, field of view of 14 × 14 mm, and flip angle of 90°. For T1-weighted images, echo time was 6.7 msec and repetition time was 350 msec. For T2-weighted images, echo time was 10 msec and repetition time was 2500 msec. T1-weighted scans obtained before and immediately after injection of 0.05 mmol of GBCA (1 mmol/mL) during the OBBBO procedure, as well as on days 3 and 7 after OBBBO, were used to assess BBB integrity. T2-weighted scans obtained after injection of GBCA during the OBBBO procedure, as well as on days 3 and 7, were used to evaluate potential brain injury. Each animal underwent MRI a total of three times.
Manual segmentation of brain tissue on T1-weighted scans after OBBBO was performed by one author (G.Q., with 18 years of experience in rodent image analysis), who was not blinded to any data. A custom Python (version 3.12.1; Python Software Foundation) script was used to analyze the pixel intensity histogram to determine a cutoff that could separate two Gaussian distributions. The entire brain volume was divided into areas above and below that cutoff. For safety assessment at follow-up MRI, the brain hemispheres were segmented on T1- and T2-weighted sections using ImageJ, and the mean intensities were calculated for comparison.
PET/CT Imaging
Approximately 5.5 MBq (150 μCi) of 89Zr-bevacizumab-deferoxamine (1 mg or 0.1 mg) reconstituted in 0.2 mL of saline was infused IA at a rate of 0.15 mL/min, and each animal was transferred to a Siemens Inveon preclinical microPET/CT scanner immediately thereafter. PET imaging was performed for 30 minutes. Then, CT was performed for anatomic coregistration. PET data were reconstructed in the axial plane using the OSEM3D/SP-MAP reconstruction algorithm, with a matrix size of 128 × 128, two ordered subsets expectation maximization (OSEM) iterations, and 18 maximum a posteriori (MAP) iterations, with a target size of 1.5 mm. Regions of interest, including the right hemisphere, left hemisphere, heart, lungs, and liver, were delineated in the axial plane. SUVs were calculated using the Fiji ImageJ PET/CT plugin. The radioactivity biodistribution in different brain regions was derived using PMOD 3.711 (PMOD Technologies) and a built-in MRI-based mouse brain map. The final PET/CT images and whole-body volume-rendered images were generated using Amira3D version 2022.2 (Thermo Fisher Scientific).
Statistical Analysis
Comparisons of the spread areas of OBBBO and mean intensities at follow-up MRI between groups, hemispheres, or time points were evaluated by one author (G.Q.) using the Wilcoxon signed rank test, Mann-Whitney U test, or mixed-effects analysis in GraphPad Prism 10.2.1 (GraphPad Software). Values are presented as means ± SDs unless stated otherwise, and differences between groups were considered significant at P < .050. Ex vivo biodistribution data, PET-based SUV data, and the distribution of radioactivity across brain regions were analyzed by two authors (D. Gulisashvili and M.J.) using the PROC MIXED procedure in SAS Studio (version 9.4; SAS Institute). Comparisons that did not show a statistically significant difference were further investigated with the Mann-Whitney U test by one author (G.Q.).
To explore the peak performance characteristics of antibody delivery, the mice in each treatment group were sorted by (a) ipsilateral-to-contralateral hemisphere ratio or (b) the highest value in the ipsilateral hemisphere for the percentage of injected dose per gram of tissue, mean SUV (SUVmean), or maximum SUV (SUVmax) (see Table S1). For the top four mice in each group, correlation analysis was performed by two authors (D. Gulisashvili and M.J.) using the PROC CORR procedure in SAS Studio to assess the correlation between PET metrics (percentage of injected dose per gram of tissue, SUVmean, and SUVmax) and ex vivo biodistribution data. One author (D. Gulisashvili) fitted a line with simple linear regression in GraphPad Prism to demonstrate the correlation between percentage of injected dose per gram of tissue and SUV measurements (Fig S1).
Results
BBB Permeability to Small Molecules Assessed Using MRI
A solution of 25% mannitol effectively opened the BBB, but a combination of 25% mannitol and 4% NaCl opened the BBB with greater spread and efficacy, as shown at contrast-enhanced MRI (Fig 2A). Using a custom Python script, intact BBB and OBBBO areas were segmented (Fig 2B) (workflow shown in Fig S2). Statistical analysis confirmed the success of OBBBO with 25% mannitol (P = .003) (Fig 2C, left). However, the combination of 25% mannitol and 4% NaCl resulted in a more pronounced effect (P < .001) (Fig 2C, right). The ratio of OBBBO area to intact BBB area in the whole brain was higher in the 25% mannitol and 4% NaCl group than the 25% mannitol group (1.99 ± 0.17 vs 1.41 ± 0.15; P < .001) (Fig 2D, left). The difference in the OBBBO territory was also significant (Fig 2D, right): In the 25% mannitol group, the OBBBO area covered less than 20% of the brain area, whereas in the 25% mannitol and 4% NaCl group, it extended to almost 50% of the brain area (P < .001).
Figure 2:
Blood-brain barrier (BBB) permeability to small molecules assessed using MRI. (A) Axial gadolinium (Gd)–enhanced T1-weighted brain MRI scans in a mouse that underwent osmotic BBB opening (OBBBO) with 25% mannitol (left) and a mouse that underwent OBBBO with 4% NaCl in 25% mannitol (right). The approximate principal axes, left (L)–right (R) and anterior (A)–posterior (P), are shown, and the sections are ordered from superior to inferior from left to right. Arrows indicate areas of OBBBO. (B) Segmentation of pixels below (red) and above (green) a threshold of pixel intensity identified using a custom Python script to represent intact BBB and OBBBO areas, respectively. (C) Box-and-whisker plots show significant differences in pixel intensity between brain tissue with intact BBB (red) and OBBBO areas (blue) in OBBBO with 25% mannitol (left) and OBBBO with 4% NaCl in 25% mannitol (right). (D) Box-and-whisker plots compare of the ratio of OBBBO area to intact BBB area (left) and percentage of the brain area that is in the OBBBO area (right) between the 25% mannitol group (red) and the 4% NaCl in 25% mannitol group (blue). In the box-and-whisker plots, the box indicates the IQR, center line indicates the median, whiskers indicate the range, and circles and triangles indicate individual data points. P values calculated using the Mann-Whitney U test: ** = P < .01, *** = P < .001, **** = P < .0001. au = arbitrary units.
Safety and Long-term Consequences of OBBBO
No animal mortality was observed during or after surgical procedures, OBBBO, or GBCA injections and for 7 days after OBBBO. T2-weighted MRI scans obtained on days 3 and 7 after OBBBO revealed no abnormal findings, such as edema, stroke, inflammation, or microhemorrhages, in either group (Fig S3A). T1-weighted images obtained at these same time points revealed no contrast enhancement in the brain, indicating that the BBB had reverted to its original permeability without long-term consequences (Fig S3B). These MRI findings suggest the safety of the OBBBO procedure for up to 7 days in mice (Fig S3).
Consistent with the MRI results, hematoxylin-eosin staining revealed no signs of injury or neuroinflammation in the OBBBO regions (Fig S4A). Immunohistochemical analysis of glial fibrillary acidic protein and ionized calcium-binding adaptor molecule 1 revealed no activation of astroglia or microglia, respectively (Fig S4B, S4C). Additionally, neuronal staining with NeuN revealed no loss of neurons (Fig S4D). These histologic findings were further validated with statistical analyses (Fig S4E–S4G), which indicated the safety of OBBBO for up to 7 days in mice.
PET Imaging of Brain Antibody Uptake
Successful radiolabeling was confirmed using mass spectrometry and instant thin-layer chromatography (Fig S5). PET imaging using IA 89Zr-bevacizumab-deferoxamine revealed that the 25% mannitol and 4% NaCl group had significantly greater radioactivity accumulation in the ipsilateral hemisphere (SUVmean and SUVmax of 1.59 ± 1.24 and 2.77 ± 1.80, respectively) than the 25% mannitol group (SUVmean and SUVmax of 0.51 ± 0.14 and 1.01 ± 0.23, respectively; P = .002 and <.001, respectively) (Fig 3). In the comparison of IA versus IV antibody delivery after OBBBO with 4% NaCl in 25% mannitol, IA delivery resulted in significantly greater radioactivity accumulation in the ipsilateral hemisphere (mean SUVmean and SUVmax of 1.54 ± 0.70 and 2.49 ± 1.16, respectively) than IV delivery (mean SUVmean and SUVmax of 0.40 ± 0.37 and 0.64 ± 0.52, respectively; P < .001 for both) (Fig 4). With IV antibody delivery after OBBBO, there was no evidence of a difference in SUVmean between the ipsilateral and contralateral hemispheres (P = .97). Also, there was no evidence of a difference in SUVmax in the contralateral hemisphere between IA and IV antibody delivery (P = .98).
Figure 3:
PET/CT imaging after osmotic blood-brain barrier opening (OBBBO). (A, B) Images in mice that underwent OBBBO with (A) 4% NaCl in 25% mannitol or (B) 25% mannitol, followed by intra-arterial (IA) injection of zirconium 89 bevacizumab-deferoxamine and whole-body PET/CT with three-dimensional (3D) reconstruction. Coronal sections from left to right are ordered from anterior (A) to posterior (P); horizontal sections are ordered from superior (S) to inferior (I); and sagittal sections are ordered from left (L) to right (R). (C, D) Box-and-whiskers plots show (C) mean standardized uptake value (SUVmean) and (D) maximum standardized uptake value (SUVmax) in the ipsilateral and contralateral brain hemispheres after OBBBO with 4% NaCl in 25% mannitol (blue) or 25% mannitol (red). In the box-and-whisker plots, the box indicates the IQR, center line indicates the median, whiskers indicate the range, and circles and triangles indicate individual data points. P values calculated using the Mann-Whitney U test: ** = P < .01. P values calculated using mixed-model analysis: ## = P < .01, ### = P < .001, #### = P < .0001.
Figure 4:
PET/CT imaging after intra-arterial (IA) versus intravenous (IV) antibody delivery after osmotic blood-brain barrier opening (OBBBO) with 4% NaCl in 25% mannitol. (A, B) Images in a mouse that underwent OBBBO followed by (A) IA or (B) IV injection of zirconium 89 (89Zr) bevacizumab-deferoxamine (89Zr-BVDFO) and whole-body PET/CT with three-dimensional (3D) reconstruction. Coronal sections from left to right are ordered from anterior (A) to posterior (P); horizontal sections are ordered from superior (S) to inferior (I); and sagittal sections are ordered from left (L) to right (R). (C, D) Box-and-whiskers plots show (C) mean standardized uptake value (SUVmean) and (D) maximum standardized uptake value (SUVmax) in both brain hemispheres after OBBBO with 4% NaCl in 25% mannitol, followed by IA (blue) or IV (green) injection of 89Zr-bevacizumab-deferoxamine. In the box-and-whisker plots, the box indicates the IQR, center line indicates the median, whiskers indicate the range, and triangles and squares indicate individual data points. P values calculated using mixed-model analysis: ### = P < .001, #### = P < .0001.
PET imaging also provided spatial data on antibody biodistribution, when overlaid with a mouse brain template using PMOD software (Fig 5A, 5B). This method allowed estimation of the volumes of mouse brain regions and assessment of the uptake of radioactivity per region. Among the 20 brain regions considered, 15 regions presented significantly greater mean radioactivity per cubic centimeter in the 25% mannitol and 4% NaCl group than in the 25% mannitol group (Fig S6). Moreover, 14 regions presented higher values of percentage of injected dose per gram of tissue in the 25% mannitol and 4% NaCl group (Fig S7). The percentage of the total injected dose delivered to each brain region indicated significantly greater doses to the cerebral cortex (P < .001), thalamus (P < .001), cerebellum (P = .02), and right hippocampus (P = .015) in the 25% mannitol and 4% NaCl group than in the 25% mannitol group (Fig 5C). Conversely, 25% mannitol alone resulted in an increased share of total radioactivity in the brainstem (P < .001) and cerebellum (P = .001) (Fig 5D). These results align with those of prior studies showing that 25% mannitol primarily targets the posterior brain, a limitation that 4% NaCl in 25% mannitol overcomes by enhancing delivery to the anterior brain regions.
Figure 5:
Distribution of zirconium 89 bevacizumab-deferoxamine in the brain, delivered intra-arterially after osmotic blood-brain barrier opening (OBBBO). (A, B) Representative coronal, horizontal, and sagittal PET images overlaid with a mouse brain template (lines) are shown for OBBBO using (A) 4% NaCl in 25% mannitol versus (B) 25% mannitol. (C) Bar graph shows the percentage of the total injected dose (ID) delivered to different brain regions. (D) Bar graph shows the percentage of total brain radioactivity across different brain regions. P values calculated using mixed-model analysis: # = P < .05, ### = P < .001, #### = P < .0001. SEM = standard error of the mean.
Ex Vivo Biodistribution
The biodistribution of 89Zr-bevacizumab-deferoxamine was further evaluated through ex vivo analysis. Consistent with earlier studies (19) and PET imaging results, high radioactivity uptake was detected in the lungs, spleen, heart, kidneys, and liver (Fig S8A, S8B). In the 25% mannitol and 4% NaCl group, there was significantly greater uptake in the ipsilateral hemisphere than the contralateral hemisphere (mean percentage of injected dose per gram of tissue, 5.88% ± 2.17 vs 1.53% ± 0.17; P = .036). In the 25% mannitol group, the uptake was 3.39% ± 0.63 in the ipsilateral hemisphere and 1.40% ± 0.20 in the contralateral hemisphere, which was not shown to be significantly different (P = .300) in the mixed-model analysis but was significantly different according to the Mann-Whitney U test (P = .004). The analysis of biodistribution did not reveal a statistically significant difference in uptake in the ipsilateral hemisphere between the 25% mannitol and 4% NaCl group and the 25% mannitol group (P = .060) (Fig S8C). Compared with the IV route, the IA route resulted in significantly greater radioactivity accumulation in the ipsilateral hemisphere, with a mean percentage of injected dose per gram of tissue of 10.30% ± 4.12 after IA injection compared with 2.59% ± 2.45 after IV injection (P < .001). The percentage of injected dose per gram of tissue differed between the ipsilateral and contralateral hemispheres after IA injection (P < .001), whereas there was no evidence of a difference between hemispheres after IV injection (P = .520) (Fig S8D).
Peak Performance Evaluation
Substantial variability in BBB permeability was observed among the mice, particularly with the 25% mannitol and 4% NaCl mixture. To assess peak performance, additional analyses were performed restricted to the four mice in each treatment group with the highest ipsilateral-to-contralateral hemisphere ratio or the highest value in the ipsilateral hemisphere for SUVmean, SUVmax, or percentage of injected dose per gram of tissue. Significant differences in SUVmean and SUVmax were observed between groups when the analysis was restricted to the four mice in each treatment group with the highest value (Fig 6A, 6B). A similar pattern was noted for percentage of injected dose per gram of tissue when analysis was restricted to the four mice in each treatment group with the highest value (Fig 6C).
Figure 6:
Box-and-whisker plots of peak performance evaluation of osmotic blood-brain barrier opening using 4% NaCl in 25% mannitol versus 25% mannitol. (A, B) Peak performance measured using PET metrics (A) mean standardized uptake value (SUVmean) and (B) maximum standardized uptake value (SUVmax) when the analysis was restricted to the four mice in each treatment group with the highest value. (C) Peak performance measured using the ex vivo biodistribution metric percent of injected dose per gram of tissue (%ID/g) when the analysis was restricted to the four mice in each treatment group with the highest value. In the box-and-whisker plots, the box indicates the IQR, the center line indicates the median, whiskers indicate the range, and dots and triangles indicate individual data points. P values calculated using the Mann-Whitney U test: * = P < .05. P values calculated using mixed-model analysis: # = P < .05, ### = P < .001, #### = P < .0001.
Discussion
We enhanced osmotic blood-brain barrier (BBB) opening by combining 4% NaCl and 25% mannitol, surpassing the previous standard of 25% mannitol alone. Our new method effectively and safely increased the permeability of the BBB across the entire targeted brain hemisphere in mice. The duration of the BBB opening needs to be reported in the future, but it did not seem to exceed half an hour.
Our approach represents an improvement over previous efforts that sought to achieve more extensive OBBBO by drastically increasing the injection rate to nearly 5 mL/min in rats (14). In our study, we maintained an upper rate limit of 0.15 mL/min in mice. Even when accounting for differences in animal mass, the disparity in infusion rates between these studies is notable. Previous research demonstrated that increasing the infusion rate even just to 0.25 mL/min can damage the mouse brain (17). Another study focusing on stem cell delivery also demonstrated the damaging effects of greater infusion rates on the rat brain (20). We achieved highly effective and safe OBBBO by increasing osmotic power rather than increasing the infusion rate.
While 25% mannitol has been the predominant agent for OBBBO, there are reports of the use of perillyl alcohol for this purpose (21). However, perillyl alcohol has not been used in studies involving radiolabeled antibodies. This agent has inherent antitumor properties (22) and could be too toxic to use outside neuro-oncology; thus, its broader usefulness as an OBBBO agent remains uncertain. Importantly, its potential adverse effects may only become apparent in a larger patient population.
The rapid expansion of endovascular procedures for stroke treatment has increased the availability of endovascular infrastructure and experienced personnel, even in middle- and low-income countries, making these countries more prepared to implement OBBBO procedures (23). Furthermore, there is substantial movement toward less invasive radial approaches for catheter entry (24), as well as robotization of procedures (25), which are expected to transform OBBBO into a time-efficient outpatient procedure with the potential for global deployment.
We also emphasize the importance of the extent of BBB opening with our method, which substantially improved antibody penetration into the brain. This improvement is evidenced by the interhemispheric difference in the PET signal intensity, which was generally 5–10 times greater and, in some cases, over 20-fold greater with our method than with the standard method. In the PET imaging of 89Zr-bevacizumab-deferoxamine, baseline radioactivity from the circulating antibodies was present in the contralateral hemisphere. Therefore, even a single-digit increase in SUV can represent an increase in the penetration of antibodies in the brain hundreds or thousands of times over that achieved with standard methods (26).
This study has several limitations. It was performed in rodents, so the effects of doubling osmotic power in the brains of large animals and humans remain unknown. We have studied the efficacy of hypertonic saline up to 4% NaCl, but this limit might be extended, especially in large animals and patients, through superselective IA cerebral infusion (27), the spatial selectivity of which may help minimize the systemic adverse effects of osmotic agents. Furthermore, our technique has not been applied in disease settings, and we have studied only GBCA as a representative small molecule and one type of antibody as a representative biologic drug. Expanding the testing molecule portfolio to include nanobodies, nanoparticles, viruses, cells, and other entities is encouraged. Moreover, while our method had a favorable safety profile in mice, there is no guarantee that it will be similarly safe for use in larger animals. Additional limitations include the small number of animals per group, manual segmentation of the brain, a single non-blinded investigator for image analysis, and, lastly, the use of ranking tests for statistical analysis.
Overall, 25% mannitol with 4% NaCl for osmotic blood-brain barrier (BBB) opening warrants further technical development and optimization before it is ready for therapeutic interventions. Nonetheless, these encouraging results may inspire other research groups to explore enhanced BBB opening through osmotic pressure modulation.
Funding: Supported by the National Institutes of Health (grants R01NS120929 and R01DA056739).
Disclosures of conflicts of interest: G.Q. Patent application with the University of Maryland. C.C. Patent application with the University of Maryland. D. Gulisashvili Patent application with the University of Maryland. S.S. No relevant relationships. L.K. No relevant relationships. L.F.P. No relevant relationships. J.O. No relevant relationships. W.G.L. No relevant relationships. D.D. Honorarium for invited presentation from the International Society for HDX-MS. M.W. Developed an open-source Python tool for dynamic MRI segmentation (April 6–May 20, 2022) and assisted researchers in using the tool. L.Z. No relevant relationships. A.J. No relevant relationships. Y.L. No relevant relationships. D. Gandhi Grants or contracts from MicroVention, National Institutes of Health, Focused Ultrasound Foundation, and University of Maryland Medical Center and patent planned, issued, or pending with University of Maryland. P.W. Support for the present manuscript from the National Institute on Drug Abuse; grants or contracts from the National Institute of Neurological Disorders and Stroke; patent planned, issued, or pending with University of Maryland, Baltimore; and stock or stock options from Ti-com and Intra-Art. M.J. Support for the present manuscript from the National Institute of Neurological Disorders and Stroke; patent planned, issued, or pending with the University of Maryland, Baltimore; and stock or stock options from Ti-com and Intra-Art.
Abbreviations:
- BBB
- blood-brain barrier
- GBCA
- gadolinium-based contrast agent
- IA
- intra-arterial
- IV
- intravenous
- OBBBO
- osmotic BBB opening
- SUV
- standardized uptake value
- SUVmax
- maximum SUV
- SUVmin
- minimum SUV
References
- 1. Bhowmik A , Khan R , Ghosh MK . Blood brain barrier: a challenge for effectual therapy of brain tumors . Biomed Res Int 2015. ; 2015 : 320941 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Arvanitis CD , Ferraro GB , Jain RK . The blood-brain barrier and blood-tumour barrier in brain tumours and metastases . Nat Rev Cancer 2020. ; 20 ( 1 ): 26 – 41 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Kadry H , Noorani B , Cucullo L . A blood-brain barrier overview on structure, function, impairment, and biomarkers of integrity . Fluids Barriers CNS 2020. ; 17 ( 1 ): 69 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Noorani I , de la Rosa J . Breaking barriers for glioblastoma with a path to enhanced drug delivery . Nat Commun 2023. ; 14 ( 1 ): 5909 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Timbie KF , Mead BP , Price RJ . Drug and gene delivery across the blood-brain barrier with focused ultrasound . J Control Release 2015. ; 219 : 61 – 75 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Partridge B , Eardley A , Morales BE , et al . Advancements in drug delivery methods for the treatment of brain disease . Front Vet Sci 2022. ; 9 : 1039745 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Neuwelt EA , Frenkel EP , Diehl JT , et al . Osmotic blood-brain barrier disruption: a new means of increasing chemotherapeutic agent delivery . Trans Am Neurol Assoc 1979. ; 104 : 256 – 260 . [PubMed] [Google Scholar]
- 8. Neuwelt EA , Maravilla KR , Frenkel EP , Rapaport SI , Hill SA , Barnett PA . Osmotic blood-brain barrier disruption. Computerized tomographic monitoring of chemotherapeutic agent delivery . J Clin Invest 1979. ; 64 ( 2 ): 684 – 688 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Neuwelt EA , Hill SA , Frenkel EP , et al . Osmotic blood-brain barrier disruption: pharmacodynamic studies in dogs and a clinical phase I trial in patients with malignant brain tumors . Cancer Treat Rep 1981. ; 65 ( Suppl 2 ): 39 – 43 . [PubMed] [Google Scholar]
- 10. Neuwelt EA , Diehl JT , Vu LH , Hill SA , Michael AJ , Frenkel EP . Monitoring of methotrexate delivery in patients with malignant brain tumors after osmotic blood-brain barrier disruption . Ann Intern Med 1981. ; 94 ( 4 pt 1 ): 449 – 454 . [DOI] [PubMed] [Google Scholar]
- 11. Doolittle ND , Miner ME , Hall WA , et al . Safety and efficacy of a multicenter study using intraarterial chemotherapy in conjunction with osmotic opening of the blood-brain barrier for the treatment of patients with malignant brain tumors . Cancer 2000. ; 88 ( 3 ): 637 – 647 . [DOI] [PubMed] [Google Scholar]
- 12. Uluc K , Ambady P , McIntyre MK , et al . Safety of intra-arterial chemotherapy with or without osmotic blood-brain barrier disruption for the treatment of patients with brain tumors . Neurooncol Adv 2022. ; 4 ( 1 ): vdac104 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Simonett JM , Ambady P , Neuwelt EA , Skalet AH , Lin P . Maculopathy associated with osmotic blood- brain barrier disruption and chemotherapy in patients with primary CNS lymphoma . Ophthalmic Surg Lasers Imaging Retina 2020. ; 51 ( 5 ): S5 – S12 . [DOI] [PubMed] [Google Scholar]
- 14. Burks SR , Kersch CN , Witko JA , et al . Blood-brain barrier opening by intracarotid artery hyperosmolar mannitol induces sterile inflammatory and innate immune responses . Proc Natl Acad Sci U S A 2021. ; 118 ( 18 ): e2021915118 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Janowski M , Walczak P , Pearl MS . Predicting and optimizing the territory of blood-brain barrier opening by superselective intra-arterial cerebral infusion under dynamic susceptibility contrast MRI guidance . J Cereb Blood Flow Metab 2016. ; 36 ( 3 ): 569 – 575 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Zawadzki M , Walecki J , Kostkiewicz B , Kostyra K , Walczak P , Janowski M . Follow-up of intra-arterial delivery of bevacizumab for treatment of butterfly glioblastoma in patient with first-in-human, real-time MRI-guided intra-arterial neurointervention . J Neurointerv Surg 2021. ; 13 ( 11 ): 1037 – 1039 . [DOI] [PubMed] [Google Scholar]
- 17. Chu C , Liu G , Janowski M , et al . Real-Time MRI guidance for reproducible hyperosmolar opening of the blood-brain barrier in mice . Front Neurol 2018. ; 9 : 921 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Chu C , Jablonska A , Gao Y , et al . Hyperosmolar blood-brain barrier opening using intra-arterial injection of hyperosmotic mannitol in mice under real-time MRI guidance . Nat Protoc 2022. ; 17 ( 1 ): 76 – 94 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Lesniak WG , Chu C , Jablonska A , et al . A distinct advantage to intraarterial delivery of (89)Zr-bevacizumab in PET imaging of mice with and without osmotic opening of the blood-brain barrier . J Nucl Med 2019. ; 60 ( 5 ): 617 – 622 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Janowski M , Lyczek A , Engels C , et al . Cell size and velocity of injection are major determinants of the safety of intracarotid stem cell transplantation . J Cereb Blood Flow Metab 2013. ; 33 ( 6 ): 921 – 927 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Wang W , Marín-Ramos NI , He H , et al . NEO100 enables brain delivery of blood–brain barrier impermeable therapeutics . Neuro Oncol 2021. ; 23 ( 1 ): 63 – 75 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Gomes AC , Mello AL , Ribeiro MG , et al . Perillyl alcohol, a pleiotropic natural compound suitable for brain tumor therapy, targets free radicals . Arch Immunol Ther Exp (Warsz) 2017. ; 65 ( 4 ): 285 – 297 . [DOI] [PubMed] [Google Scholar]
- 23. Lim J , Aguirre AO , Rattani A , et al . Thrombectomy outcomes for acute ischemic stroke in lower-middle income countries: a systematic review and analysis . World Neurosurg X 2024. ; 23 : 100317 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Haussen DC , Nogueira RG , DeSousa KG , et al . Transradial access in acute ischemic stroke intervention . J Neurointerv Surg 2016. ; 8 ( 3 ): 247 – 250 . [DOI] [PubMed] [Google Scholar]
- 25. Bravo J , Wali AR , Hirshman BR , et al . Robotics and artificial intelligence in endovascular neurosurgery . Cureus 2022. ; 14 ( 3 ): e23662 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Al Ojaimi Y , Blin T , Lamamy J , et al . Therapeutic antibodies—natural and pathological barriers and strategies to overcome them . Pharmacol Ther 2022. ; 233 : 108022 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Riina HA , Fraser JF , Fralin S , Knopman J , Scheff RJ , Boockvar JA . Superselective intraarterial cerebral infusion of bevacizumab: a revival of interventional neuro-oncology for malignant glioma . J Exp Ther Oncol 2009. ; 8 ( 2 ): 145 – 150 . [PubMed] [Google Scholar]