The pharmacodynamics, pharmacokinetics, and toxicological profiles of bradykinin and NG291 were computed using the pkCSM tool¹ as described by Pires et al. (2015) and Oladele et al. (2021). The canonical SMILE molecular structures of the compounds used in the studies were obtained from the JSME editor.² The results obtained by running the pkCSM tool on bradykinin and NG291 are summarized in Table 1.

All animals were housed and handled following protocols approved by Universidad Central del Caribe or the Veterans Affairs Puget Sound Health Care System’s Institutional Animal Care and Use Committee (IACUC). All experiments were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The animals used in this work were: Sprague Dawley (SD) rats (104 males and 57 females with 8–9 weeks of age from Universidad Central del Caribe colony weighing 250–300 g) and CD-1 mice (154 males 8–9 weeks old from Jackson Laboratories weighing 30–40 g) with food and water ad libitum with 12 h day/night cycles. Animal numbers used were based on calculations for typical studies to detect a 10% difference in the mean values (alpha = 0.05) with a power above 80%.

NG291 (Savard et al., 2013) was purchased from MediLumine.³ SD rats were anesthetized with a mixture of 1.5% isoflurane/70% nitrous oxide/30% oxygen. The anesthetized animals were injected (i.v.) with either saline, 50 or 100 μg/kg of NG291. The injected solution circulated for 1–8 h before an IV injection of 4 ml/kg 2% Evans blue (Millipore) solution (EB) dissolved in 0.9% saline. Sixteen hours after EB injection, the animals were submitted to transcardial perfusion with saline 0.9%. A blunt needle was introduced through the left ventricle and held in position when the needle head was visible in the aorta. The right atrium was punctured as an outlet for drainage of the perfusion fluid. Discoloration of the liver was used to confirm successful perfusion. After the perfusion, the brains were coronally sectioned using a stainless-steel brain matrix (Stoelting). Coronal sections (2 mm thick) were placed on a glass mount which allowed brains to be scanned dorsally and ventrally using an Epson Perfection V39 Scanner. The percent area of EB detected in the brain in proportion to the total brain area was calculated using ImageJ software (United States National Institutes of Health, Bethesda, MD, United States⁴).

Additionally, saline perfused brains were also coronally sectioned at 20 μm thickness using a Leica CM1850 microtome. EB (620 nm excitation/680 nm emission) extravasation to brain parenchyma was detected with a Keyence BZ-X800 fluorescence microscope after mounting slides with mounting media containing DAPI (Saria and Lundberg, 1983; Wang and Lai, 2015).

The radiolabeled probes used to detect differences in transcytosis in NG291-injected animals and controls were prepared by dissolving 2 mg of bovine serum albumin (BSA) (Sigma A-7030) and 250 μg of Stannous tartrate (MP Biomedicals: ICN221948) in 1 ml of deionized water and adjusting to a pH of 3.0. Subsequently, one millicurie of ^99mTc-NaOH₄ (GE Healthcare, Piscataway, NJ, United States) was added to the solution and incubated for 20 min at room temperature. Then, the ^99mTc-albumin was purified using a G-10 Sephadex (GE Healthcare) column in 0.1 ml fractions of phosphate buffer (0.25 M). The reaction efficiency was evaluated by acid precipitation with 30% trichloroacetic acid. Reaction efficiencies under 87% were discarded (Logsdon et al., 2018). The prepared ^99mTc-albumin [10 × 10⁶ counts per minute (cpm)/mouse] was combined with 0.75 μCi/mouse (1.665 × 10⁶ dpm/mouse) of C¹⁴-sucrose (Perkin Elmer, Waltman, MA, United States) and dissolved in lactated Ringer’s solution containing 1% BSA (at the final volume of 0.2 ml/mouse). Either 50 or 100 μg/kg of NG291 were administered to assess changes in radiolabeled probe accumulation in brain parenchyma.

A tritiated 2-PAM probe was used to verify whether 2-PAM would access the brain parenchyma following NG291 administration. To prepare the solution, 0.1 μCi/mouse of ³H-pralidoxime (³H-2-PAM, American Radiolabeled Chemicals, ART 2162-50 μCi; specific activity: 10 Ci/mmol) was constituted with 100 μg/kg of NG291 in lactated Ringer’s solution containing 1% BSA (final volume of 0.2 ml/mouse).

CD-1 Mice were anesthetized with urethane (4 g/kg; 0.2 ml i.p.). Once the animals had been sedated, the jugular veins were exposed, and intravenous co-injection of NG291 with the radiolabeled tracers was administrated. The injectate circulated for 15 min then the blood was collected from the severed inferior carotid artery. Immediately, transcardial perfusion was performed with 20 ml of Lactated Ringer’s solution. The brain was harvested and weighed before the counting of retained radioactivity. Blood was centrifuged for 10 min at 3500 × g and 4°C, and 20 μl of serum was collected for counting. At the end of the study, 20 μl of the injectate was collected as an “injection check,” which is used as a standard to determine the amount of tracer present in the provided injectate. Samples were counted in a gamma counter for 3 min to determine cpm per tissue weight for the brain samples or per 20 μl in the case of serum samples and injection check. Solvable™ (1.5 ml, Perkin Elmer, catalog 6NE9100) was then added to each sample tube and sealed. Samples were then stored until the brain tissue was fully dissolved and technetium decayed. After dissolution, they were counted in a beta counter to quantify ¹⁴C or ³H concentrations in terms of dpm/g for brain samples or dpm/μl for serum samples and injection checks. Radiolabeled probe accumulation in brain parenchyma is presented in terms of brain to serum ratios (μl/g).

Sprague Dawley rat treated with saline (control), acute NG291 (100 μg/kg NG291 3 days prior to saline perfusion), and chronic NG291 (100 μg/kg NG291 per day for 3 days for three doses) with brain perfusion conducted 1 h after the final dose was sectioned 20 μm thick. Brain sections were stained for F-actin (phalloidin, Alexa Fluor™ 488, green) and nuclei (DAPI, blue). Stained slides were observed under a Keyence BZ-X800 fluorescence microscope and imaged.

CD-1 mice were anesthetized with urethane (4 g/kg; 0.2 ml; i.p.). Anesthetized mice were co-injected with ³H-2-PAM (0.1 μCi/mouse) and ^99mTc-albumin (10 × 10⁶ counts/injection) in a volume of 0.1 ml in the jugular vein allowed to circulate at 1, 2, 5, 10, and 15 min. At which point, blood was collected from the descending abdominal aorta. Immediately after blood collection, transcardial perfusion with lactated Ringer’s solution was performed before collecting brain, lung, liver, and kidney. Serum and tissue samples were then gamma- and beta-radiation scintillation counted.

CD-1 mice were subjected to 100 μg/kg NG291 i.v. or Ringer’s lactate alone (control) co-injected with ³H-2-PAM (0.1 μCi/injection, 1 μCi/μl) and ^99mTc-albumin (10 × 10⁶ counts/injection). Injectate circulated for 10 min, at which point blood was collected from the descending carotid artery. The samples were counted in a gamma counter for 3 min to determine ^99mTc-albumin cpm per tissue weight for the brain samples or per 20 μl in the case of injection checks and serum samples. Transcardial perfusion was done before the brain collection. The samples were then counted in a beta-scintillation counter to quantify ³H-2-PAM radioactivity in terms of dpm/g for brain samples or dpm/μl for injection checks and serum samples.

CD-1 mice were anesthetized with 0.1–0.2 ml of 40% urethane i.p. Anesthetized mice were placed in the supine position, and the right and left jugular vein and carotid arteries were exposed. The thorax was opened from the epigastric region of the abdomen up to the sternal notch, cutting through the sternum. Jugular veins were severed bilaterally, and carotid arteries were left intact. The descending thoracic aorta was clamped with a hemostat, and a butterfly needle was introduced into the left ventricle of the heart. The syringe pump was activated, perfusing 2 ml/min for 2 min (Smith and Allen, 2003). Perfusate solution used in control groups and in NG291 studies was prepared by diluting 50 × 10³ dpm/ml of ³H-verapamil (0.09 μCi/mouse) and 10 × 10⁶ counts/ml of ^99mTc-albumin in Zlokovic buffer [7.19 g/L NaCl, 0.3 g/L KCl, 0.37 g/L CaCl₂, 2.1 g/L NaHCO₃, 0.16 g/L KH₂PO₄, 0.17 MgCl₂, 0.99 g/L D-(++)-glucose, 10 g/L BSA, diluted in di H₂O, adjusted to pH of 7.4]. In NG291 studies, 100 μg/kg of NG291 was administered i.v. 10 min prior to in situ brain perfusion. For pralidoxime studies, 10 μg/ml of pralidoxime (not radiolabeled, Sigma), diluted in the perfusate preparation, was used in control groups. At the end of the perfusion, the brain was collected along with perfusate samples for posterior gamma- and beta-scintillation counting.

The equation above was used to obtain the brain to perfusion ratio for both ^99mTc-albumin and for ³H-verapamil. The ratio for ^99mTc-albumin was then subtracted from the ratio for ³H-verapamil to yield the ratio for ³H-verapamil taken up by the brain. The activity of the P-gp efflux system is inversely related to this delta ³H-verapamil ratio.

Sprague Dawley rats were anesthetized with a mixture of 1.5% isoflurane/70% nitrous oxide/30% oxygen. Anesthetized rats were treated i.v. with saline, 50 μg/kg NG291, or 100 μg/kg NG291 and were euthanized 24 h after the injection. The brains were harvested, and the weight was recorded. The brains were then left to dehydrate in an oven at 52°C for 72 h or until there was no change in weight after two consecutive readings. Percentage of brain water was calculated with the following equation as used by Elliott and Jasper (1949):

Sprague Dawley rats were anesthetized with a mixture of 1.5% isoflurane/70% nitrous oxide/30% oxygen. In this study, three groups were included: saline group (negative control), a group provided an acute 100 μg/kg NG291 on the first day, and a “chronic” group that was given 100 μg/kg NG291 per day for 3 days. Rats were perfused with 0.9% saline solution, and the brains were stored at −70°C on the third day after the treatment or an hour after the third dose of 100 μg/kg NG291 in the chronic group. Brains were sectioned with 20 μm thickness using a Leica CM1850 cryostat. We stained the sections with GFAP-Cy3 (Sigma C9205) and DAPI to label activated astrocytes and nuclei, respectively. Stained brain sections were observed under a Keyence BZ-X800 fluorescence microscope.

Sprague Dawley rats were anesthetized with a mixture of 1.5% isoflurane/70% nitrous oxide/30% oxygen. A group of animals was subjected to middle carotid artery occlusion (MCAO) (Rupadevi et al., 2011), which served as our positive control. In this study, three groups were included: saline group (negative control), a group provided an acute 100 μg/kg NG291 on the first day, and a “chronic” group that was given 100 μg/kg NG291 per day for 3 days. Rats were perfused with 0.9% saline solution, and the brains were stored at −70°C on the third day after the treatment or an hour after the third dose of 100 μg/kg NG291 in the chronic group. Brains were sectioned with 20 μm thickness using a Leica CM1850 cryostat. We stained the sections with the Fluoro-Jade C (FJC) Ready to Dilute Staining Protocol by Biosensis (catalog number TR-100-FJ) for non-paraffin embedded sections (Ferchmin et al., 2014; Deshpande and Phillips, 2016; Krishnan et al., 2016). Stained slides were observed under a Keyence BZ-X800 fluorescence microscope for positive FJC staining.

One−way analysis of variance (ANOVA) was used for analysis when the data had a normal distribution. Statistical analyses were performed using GraphPad Prism^® (Version 8.0.2, GraphPad Software, Inc., La Jolla, CA, United States). Bar graphs and scatterplots exhibit the mean value of the group, with error bars indicating the standard error of mean (SEM). The presented results were not tested for outliers.

ADMET profiles of bradykinin and NG291 were predicted using the pkCSM tool (see text footnote 1) as described by Pires et al. (2015) and Oladele et al. (2021). The results obtained by running the pkCSM tool for bradykinin and NG291 were summarized and compared in Table 1. Neither bradykinin nor NG291 was predicted to cross Caco-2 monolayers since highly permeable compounds should have Papp values above 8 × 10^–6 cm/s. BK and NG291 have been predicted to have a low volume of distribution (Vd = 0.208449088; 0.086896043, respectively) in humans, unable to penetrate BBB (logBB = −2.137, −2.592, respectively) and unable to penetrate the CNS (logPS = −6.197; 6.465, respectively) (Pires et al., 2015). This can be expected as bradykinin or NG291 do not comply with Lipinski’s rule of five in terms of rotatable bonds, the number of proton donors and acceptors, and molecular weight (Lipinski et al., 2001). According to Table 1, BK and NG291 are predicted to be CYP3A4 substrates but do not play a role in its inhibition. NG291 is predicted to inhibit hERG II (Table 1), which would lead to a prolonged QT interval and potential Torsades de pointes in susceptible patients. Additionally, NG291 is predicted to be a P-glycoprotein efflux transporter substrate but not inhibit P-glycoprotein activity (Table 1).

Several methods can be used to evaluate BBB disruption. A simple, low-cost, and fast method widely used is EB dye extravasation, which can determine BBB disruption following NG291 administration without any sophisticated equipment (Saunders et al., 2015). Here, SD rats were injected i.v. with NG291 followed by i.v. EB injection. In Figure 1, NG291 increases EB extravasation in regions surrounding blood vessels in the cortex and near lateral ventricles. To explore how long NG291 can maintain increased BBB permeability, we administered EB at various time points following an acute i.v. injection of 50 or 100 μg/kg NG291. Using this procedure, we confirmed that NG291-mediated BBB disruption is dose-dependent. Observe in Figure 2A that BBB integrity is recovered by the 2 h following 50 μg/kg of NG291 time point (0.09453 ± 0.02517; p = 0.7195). Doubling the concentration of NG291 (Figure 2B) causes the BBB to remain disrupted for a longer period, with detected EB extravasation returning to basal levels by the 4 h time point (0.3484 ± 0.04128; p = 0.0230). Figure 2C reveals that NG291 promotes similar effects with EB extravasation returning to control conditions by the 4-h time point (0.5815 ± 0.08710; p = 0.1102). Additionally, we noticed in Figure 2C that the BBB of female rats is more permeable to EB extravasation compared to male counterparts, as shown in Figure 2B. Post hoc power analysis in these studies was above 80%, with alpha = 0.05.

Evans blue extravasation into the brain confirmed that NG291 transiently disrupted the BBB. For further verification of transport enhancement by NG291, we used radiolabeled probes for assessing BBB permeability. ¹⁴C-sucrose and ^99mTc-albumin are commonly used radiolabeled markers that remain in vascular compartments under normal conditions because of their incapability to effectively cross the BBB (Levin et al., 1970; Blasberg et al., 1983; Patlak et al., 1983). ¹⁴C-sucrose is used as the small molecular weight marker (∼342 Da) while ^99mTc-albumin is the high molecular weight marker (∼66.44 KDa) to probe BBB disruption (Ziylan et al., 1984; Mayhan and Heistad, 1985; Armstrong et al., 1987; Robinson and Rapoport, 1987). CD-1 mice were used in the study instead of rats to reduce excess exposure to radioactive material handled throughout the investigations. Studies had to be conducted quickly after preparing ^99mTc-albumin because discrepancies resulting from radioactive decay can influence the results [technetium (^99mTc) half-life is 6 h]. Due to the large number of studies that needed to be completed on the same day for obtaining statistically significant data, we limited the time of probe circulation to 15 min between NG291 administration and radiolabeled probe administration. As shown in Figure 3A, the brain to serum ratio of ^99mTc-albumin significantly increased 15 min after either 50 μg/kg (1.271 ± 0.370; n = 16; p = 0.0083) or 100 μg/kg (1.398 ± 0.533; n = 24; p = 0.0001) NG291 administration. In Figure 3B, ¹⁴C-sucrose accumulation was only statistically significant, compared to controls, 15 min after the injection of 100 μg/kg NG291 (6.357 ± 2.199; n = 14; p = 0.0128).

To aid in the detection of paracellular leakage, Supplementary Figure 1 shows representative images of CA1 hippocampal brain sections (interaural 5.20 mm, Bregma −3.80 mm) of SD rats treated with saline (i.v.) or 100 μg/kg of NG291. Brain sections were stained for F-actin expression (phalloidin, Alexa Fluor™ 488, green) and nuclei (DAPI, blue). F-actin expression in rat CA1 hippocampal sections was reduced in either acute or chronic NG291 administration.

Pralidoxime is an FDA-approved medication used to recover acetylcholinesterase activity before covalently binding to organophosphates. Such effects are induced by some insecticides or nerve agents employed in warfare (Cannard, 2006; Buckley et al., 2011; Ferchmin et al., 2015). The inability of 2-PAM to recover acetylcholinesterase activity in the CNS has restricted the overall success of the drug (Okuno et al., 2008). This is because 2-PAM is unable to reach therapeutically relevant concentrations in the CNS (Sakurada et al., 2003, 2015). 2-PAM is a small positively charged drug that should have similar kinetics as the ¹⁴C-sucrose marker used in Figure 3B. In Figure 3B, the 50 μg/kg dose of NG291 was unable to significantly increase ¹⁴C-sucrose accumulation in the brain. Following suit, we opted to use the higher dose (100 μg/kg) of NG291 to see if we can achieve a significant increase in 2-PAM concentrations in the brain. We first determined the baseline pharmacokinetics of ³H-2-PAM in the liver, kidney, lung, heart, and brain in proportion to serum (Figures 4A–E). Longer exposure times lead to the accumulation of ³H-2-PAM in the liver and kidney (Figures 4A,B), whereas lung and heart tissue results in a reduction in ³H-2-PAM (Figures 4C,D). In Figure 4E, we observe signs of ³H-2-PAM accumulation in the brain with longer exposure times. With this observation, we sought out to determine if NG291 mediated increase in BBB permeability facilitates ³H-2-PAM penetration in the CNS. However, there was no significantly increased accumulation of ³H-2-PAM (482.5 ± 127.3 μl/g; p = 0.2269) in CD-1 mice treated with ³H-2-PAM and 100 μg/kg NG291 (Figure 4F).

Verapamil is a P-gp substrate. Increased ³H-verapamil concentrations in the brain indicate the presence of additional substrates interacting with P-gp transporters. Decreased ³H-verapamil concentrations in the brain indicate enhanced P-gp transporter activity. No changes in Verapamil concentrations indicate the absence of interactions with P-gp transporters. NG291 induced P-gp transporter activity as it significantly decreased ³H-verapamil accumulation in the brain (Figure 5A, p = 0.0028), suggesting that NG291 plays a role in increased P-gp transport activity. However, 2-PAM displaced ³H-verapamil from binding to P-gp transporters (Figure 5B, p = 0.0067), suggesting that 2-PAM is a P-gp substrate. This may also explain why there was no increased 2-PAM entry into the brain when NG291 induced BBB leakage.

One of the main risks with BBB disruption has been associated with changes in the brain water content (Abbott, 2000; Marceau et al., 2020). The principle is that a compromised BBB allows access to low and high molecular weight molecules would result in vasogenic edema. In this study, we quantify brain water content to ensure that the risks associated with this proposed drug delivery strategy do not outweigh the potential benefits. The data were obtained by following an established protocol to determine the brain water content (Mao et al., 2015). As shown in Figures 6A,B, there is no statistically significant change in brain water content in male and female SD rats treated with NG291.

In order to test the hypothesis (Siracusa et al., 2019), SD rats were treated with: saline; acute 100 μg/kg NG291 treatment as well as 100 μg/kg of NG291 administered every 24 h for 3 days (chronic). Following imaging of brain sections (interaural 5.20 mm, Bregma −3.80 mm) in Figure 7A, GFAP-Cy3 positive activated astrocytes in CA1 hippocampus in the different experimental conditions are presented. In Figure 7B, GFAP-Cy3 fluorescent intensity was calculated via ImageJ throughout a brain cross-section (interaural 5.20 mm, Bregma −3.80 mm). One-way ANOVA revealed p-values above 0.05 when comparing GFAP expression in control conditions with that of acute or chronic 100 μg/kg NG291-treated rats. Post hoc power analysis was calculated with an alpha of 0.05 at 80.6%. n = 5 in all conditions.

For NG291 mediated BBB disruption to represent a viable strategy to facilitate therapeutic delivery to the brain, the benefits must outweigh the risks. FJC staining was employed to detect signs of neurodegeneration following NG291 treatment. FJC stains all degenerating neurons, regardless of specific insult or mechanism of cell death (Schmued et al., 2005; Li et al., 2011). Brain sections of saline-treated rats are presented in Figure 8A. As a positive control, rats underwent an MCAO procedure to provide significant FJC positive cell staining for neurodegeneration (Figure 8B). Positive FJC staining was absent in rats treated with either a single 100 μg/kg NG291 dose 3 days before harvesting the brain (Figure 8C) or in rats treated daily with 100 μg/kg NG291 for 3 days (Figure 8D).