Focused Ultrasound Pre-Conditioning for Augmented Nanoparticle Penetration and Efficacy in the Central Nervous System

Abstract

Microbubble activation with focused ultrasound (FUS) facilitates the non-invasive and spatially-targeted delivery of systemically administered therapeutics across the blood-brain barrier (BBB). FUS also augments the penetration of nanoscale therapeutics through brain tissue; however, this secondary effect has not been leveraged. Here, we first identified 1 MHz FUS sequences that increase the volume of transfected brain tissue after convection-enhanced delivery of gene-vector “brain-penetrating” nanoparticles. Next, FUS pre-conditioning was applied prior to trans-BBB nanoparticle delivery, yielding up to a 5-fold increase in subsequent transgene expression. MRI analyses of tissue temperature and K_trans confirmed that augmented transfection occurs through modulation of parenchymal tissue with FUS. FUS pre-conditioning represents a simple and effective strategy for markedly improving the efficacy of gene vector nanoparticles in the central nervous system.

Gene therapy has the potential to slow or reverse numerous pathologies of the central nervous system (CNS), including Parkinson’s and Alzheimer’s Diseases ^[1]. However, despite successes in small animal models ^[2–4], gene therapy in the CNS has had limited effectiveness in clinical trials^[5]. It has been hypothesized that outcomes may be improved by enhancing delivery efficiency ^[6] and transfection volume ^[7], as well as by treating patients at an earlier (or prodromal) stage prior to the onset of irreversible pathology ^[8].

Achieving effective gene vector distribution in the brain is hampered by the presence of two physical barriers. The first is the dense and nanoporous extracellular matrix (ECM) ^[9], which consists of a lattice of electrostatically charged molecules, including proteoglycans, hyaluronan, and tenascins, that hamper diffusion of gene vectors via steric and/or adhesive interactions. To address the brain ECM barrier, our group has shown that engineering nanoparticles with high density coats of polyethylene glycol (PEG) [i.e. “brain-penetrating nanoparticles” (BPN)] improves their penetration and efficacy ^[2,10–13]. The second physical barrier is the blood-brain barrier (BBB), which prevents nearly 100% of systemically circulating molecules larger than ~400 Da from entering the brain ^[14]. Activating circulating microbubbles (MBs) with MR image-guided focused ultrasound (FUS) transiently opens the BBB in a spatially-targeted manner, with BBB integrity restored within 4-6 hours after treatment ^[15,16]. This strategy has been shown to be safe, and multiple clinical trials utilizing FUS and MBs for BBB opening in patients have either been completed or are underway^[17–20]. Our group has demonstrated that this approach facilitates the delivery of BPN, with diameters ranging from 40-65 nm, into the CNS ^{[2,10,13,21,22]}. Further, we have used this strategy to restore multiple indicators of neurodegeneration in a rat model of Parkinson’s disease after delivery of BPN bearing a gene for the glial cell-derived neurotrophic factor ^[2].

In addition to its ability to enable the delivery of therapeutic agents across the BBB, FUS also enhances the dispersion of therapeutic agents through brain tissue ^[23–25]. However, this secondary effect has yet to be leveraged therapeutically. To determine whether such secondary effects could be leveraged for improved BPN delivery, we first examined the extent to which FUS pre-conditioning enhances the volume of transfected brain tissue following convection-enhanced delivery (CED) of reporter gene-bearing BPN. CED is an intracranial administration method that involves continuous infusion of therapeutic solution into the brain parenchyma at a predetermined rate. The procedure establishes a pressure-driven bulk flow that convects infusate away from the site of administration, thereby enhancing the therapeutic distribution in the brain tissue. Sprague-Dawley rats were pre-conditioned in the right striatum with (i) FUS [1 MHz, 1.2 MPa peak-negative pressure (PNP), 1% duty cycle = 10 ms pulses with 1s interval, 4 minutes] or (ii) FUS + MB (1 MHz, 0.6 MPa PNP, 0.5% duty cycle = 10 ms pulses with 2s interval, 2 minutes, 1x10⁵ MBs/g i.v.) immediately prior to CED of ZsGreen-BPN (diameter = 44.7 ± 3.1 nm; ζ-potential = 0.56 ± 2.0 mV), wherein the ZSGreen reporter plasmid was driven by the ubiquitously active CMV promoter. Plasmids were condensed with a blend of polyethylenimine (PEI) and 5 kDa PEG-conjugated PEI (PEG_5k-PEI). The high density of PEG in these systems has been shown to greatly reduce, if not eliminate, toxicity caused by the cationic nature of PEI ^[13,26,27]. Nanoparticles containing similar size and charge characteristics have been shown to penetrate brain tissue when infused intracranially, providing uniform reporter gene expression ^[27–30]. ZsGreen-BPN transfected tissue volume was assessed at 48 h post-administration. Brain tissue pre-conditioning with 1.2 MPa FUS and 0.6 MPa FUS+MB led to, respectively, 44% and 142% increases in transfection volume when compared to CED of ZsGreen-BPN without FUS pre-conditioning (Figure 1). Note that a 1.2 MPa FUS+MB group was not tested here because activation of these MBs at PNPs well above 0.6 MPa generates petechiae in rats^[10]. ZsGreen expression was never observed in the contralateral striatum. Broadly speaking, these findings are in agreement with previous studies on the influence of ultrasound on therapeutic penetration in brain tissue ^[23–25].

Figure 1. FUS pre-conditioning increases volume of transfection after CED administration of ZsGreen-BPN.

(a) Confocal image slices from sections through the striatum after CED of ZsGreen-BPN. (b) Isosurface 3D images depicting volumetric distribution of ZsGreen reporter transgene expression, obtained by stacking multiple sequential confocal images. (c) Bar graph of volume of ZsGreen transgene expression after CED. n = 5 per group. Means ± S.D. *p<0.014 vs. CED. **p<0.003 vs. CED. Unpaired t-tests were used instead of a one-way ANOVA because 0.6 MPa FUS + MB is a distinct treatment when compared to 1.2 MPa FUS without MB.

Given the ability of FUS pre-conditioning to enhance BPN dispersion after CED administration (Figure 1), we hypothesized that pre-conditioning normal brain tissue with FUS before opening the BBB to facilitate the delivery of reporter gene bearing BPN could augment subsequent transgene expression. To address this hypothesis, we tested whether pre-conditioning brain tissue with pulsed FUS would enhance transgene expression generated by the delivery of Luc-BPN (diameter = 48.1 ± 3.9 nm; ζ-potential = 1.14 ± 2.7 mV) across the BBB. Here, reporter plasmids carried by the BPN and encoding luciferase were driven by the ubiquitously active human β-actin promoter. All treatments were operated under MR image guidance. T2-weighted MR images were acquired and used to plan 3 equally spaced sites in either the right or left striatum to be pre-conditioned with FUS. The FUS pre-conditioning parameters for one group were chosen to mimic the previous CED experiment (i.e. 1 MHz, 1.2 MPa PNP, 1% duty cycle = 10 ms pulses with 1 s interval, 4 minutes). However, because ZSGreen-BPN dispersion after CED was considerably (~3-fold) greater in the presence of acoustic amplifiers (MBs) in Figure 1, we generated 2 other treatment groups with much higher FUS pre-conditioning energy deposition levels. This was achieved by increasing PNP, duty-cycle, and total FUS application time (i.e. 1 MHz; 2 MPa or 4 MPa PNP; 2.25% duty cycle = 45 ms pulses with 2s interval; 10 min duration). After FUS pre-conditioning in the absence of MBs was applied on one side, BBB opening proceeded on both sides using FUS with i.v. MBs (1 MHz, 0.55 MPa PNP, 0.5% duty cycle = 10 ms pulses with 2s interval, 1x10⁵ MBs/g, 2 min duration) and the same spatial treatment plan (Figure 2a). We have previously shown, via T2* MRI^[2,10], H&E staining^[2,10,13], and immunohistochemistry^[13], that applying FUS with these parameters to MBs in rats safely opens the BBB.

Figure 2. FUS pre-conditioning increases transgene expression generated by the subsequent delivery of Luc-BPN across the BBB.

(a) T2 MR images showing treatment planning. The striatum of one hemisphere was pre-conditioned with pulsed FUS. The BBB in both hemispheres was then opened with FUS+MB. Red spots correspond to sonication locations. (b-d) Ex vivo bioluminescence images and bar graphs of total flux. Pre-conditioning was performed at 1.2 MPa (1.0% duty-cycle; 4 min duration) (b), 2.0 MPa (2.25% duty-cycle for 10 min) (c), and 4.0 MPa (2.25% duty-cycle for 10 min) (d). Bar graphs show fold-change over FUS Pre⁻ and represent means ± SD. n=7 (1.2 MPa), n=7 (2.0 MPa), and n=10 (4.0 MPa) per group. *p=0.01 (paired t-test). **p=0.007 (paired t-test).

Luc-BPN were i.v. administered immediately before BBB opening. Of note, in previous studies, we have shown that i.v. administration of gene-bearing BPN to rats has no effect on weight gain over a 12 week period^[2] and does not elicit detectable off-target transfection^[2,13]. Three days post-treatment, bioluminescence imaging showed that transgene expression was localized to FUS treated striata (Figure 2b–d). Despite enhancing the volume of transfection after CED of ZSGreen BPN (Figure 1a and 1b), FUS pre-conditioning at 1.2 MPa had no appreciable effect on the magnitude of transfection generated by the delivery of Luc-BPN across the BBB (Figure 2b). However, increasing duty-cycle to 2.25%, application time to 10 min, and PNP to 2 MPa (Figure 2c) or 4 MPa (Figure 2d) led to, respectively, 1.8- and 4.9-fold increases in gene expression when compared to the contralateral striatum without pre-conditioning.

After establishing that FUS pre-conditioning at 2 MPa and 4 MPa led to enhanced transfection magnitude, we used both MRI and acoustics emissions monitoring to test whether this augmentation could have been due to the prolonging and/or enhancing of subsequent BBB opening with FUS and MBs. Static T1 contrast MR images were acquired at 0, 2, and 4 hours after opening the BBB, both with and without 4 MPa FUS pre-conditioning (Figure 3a). As expected, grayscale image intensities decreased significantly with time after treatment, consistent with the transient nature of the BBB opening response. Image intensity was unchanged with 4 MPa FUS pre-conditioning at all time points, indicating FUS pre-conditioning does not impact the rate of BBB closure (Figure 3b). In separate experiments, dynamic contrast enhanced (DCE) MRI was performed to calculate transfer coefficients (K_trans) immediately after FUS + MB BBB opening, both with and without 2 MPa and 4 MPa FUS pre-conditioning. K_trans was unaffected by FUS pre-conditioning (Figure 3c and 3d), indicating there was no effect on the magnitude of subsequent BBB opening. K_trans values in the 4 MPa group (Figure 3d) appeared to trend lower than the 2 MPa group (Figure 3c). We believe this may be attributed, at least in part, to increased acoustic attenuation by the skull^[31], as rats were slightly heavier in the 4 MPa group for this particular experiment [214.7 ± 1.8 g (2 MPa) vs. 223.6 ± 2.4 g (4 MPa); p<0.014 by unpaired t-test]. We recorded acoustic emissions via a listening hydrophone during all BBB opening procedures for Luc-BPN delivery, and both stable (SCD) and inertial (ICD) cavitation doses were calculated. At both the 2 MPa (Figure 3e) and 4 MPa (Figure 3f) pre-conditioning levels, SCD and ICD were unchanged from contralateral striata, wherein the BBB was opened without FUS pre-conditioning.

Figure 3. FUS pre-conditioning does not affect subsequent BBB opening.

(a) T1 contrast MR images showing BBB opening (3 sonication locations per hemisphere), both with (yellow circles) and without 4 MPa FUS pre-conditioning, taken at 0, 2, and 4 hours after BBB opening. Neither the magnitude nor duration of BBB opening appeared to be affected by FUS pre-conditioning. (b) Bar graph of R.O.I. mean grayscale intensity, taken as a percentage over T1 baseline, from static T1 contrast MR images at 0, 2, and 4 hours after BBB opening. n=3 per group. Mean ± S.D. *p<0.01 vs. same group at 0 h. Two-way RM-ANOVA and Tukey’s pairwise comparisons. (c, d) Bar graphs of transfer coefficients (K_trans) calculated from DCE MRI during BBB opening, both with and without 2 MPa (c) and 4 MPa (d) FUS pre-conditioning. n=6 (2 MPa) and n=5 (4 MPa) per group. Means ± S.D. Paired t-tests. (e, f) Bar graphs of stable and inertial cavitation doses collected during BBB opening, both with and without 2 MPa (e) and 4 MPa (f) FUS pre-conditioning. n=7 (2 MPa) and n=10 (4 MPa) per group. Means ± S.D. Paired t-tests.

We next aimed to determine how FUS pre-conditioning affected tissue heating. To characterize spatial temperature distributions and estimate thermal dose (cumulative equivalent minutes at 43°C; CEM43°C) ^[32], MR thermometry was performed during FUS pre-conditioning at both 2 MPa (n=1) and 4 MPa (n=6). From the 4 MPa group, a representative maximum temperature plot as a function of time is provided (Figure 4a). Mean maximum temperature per rat (Figure 4b) was calculated by averaging maximum temperatures across the 3 FUS sonication locations. CEM43°C values were calculated for each rat (Figure 4c), yielding a mean of 1.02 min with 4 MPa FUS pre-conditioning. In previous studies, the minimum reported CEM43°C at which brain tissue begins to show damage is 7.5 min ^[32], which is multi-fold higher than the CEM43°C values measured here. In another study, when heating brain tissue to 42°C, a full 60 min of exposure was needed to damage cortical neurons ^[33]. Assuming an absolute baseline temperature of 37°C, this corresponds to a 5°C temperature increase. In our study, wherein FUS exposures lasted only 10 min, a >5°C temperature rise was observed in only one rat, corresponding to a CEM43°C of 3.84 min. In yet another study, brain tissue damage was not observed histologically with FUS application until temperatures >50°C were achieved ^[34]. In that study, FUS was applied for 10 s, corresponding to a CEM43°C of >20 min..

Figure 4. FUS pre-conditioning elicits neither significant tissue heating nor overt evidence of tissue irritation.

(a) Representative plot of mean maximum temperature as a function of time for a 4 MPa FUS pre-conditioning procedure. (b) Maximum temperature increases, averaged over 3 sonication spots, for 2 MPa (n=1) and 4 MPa (n=6) FUS pre-conditioning. Mean ± S.D. (c) CEM43°C values for 2 MPa (n=1) and 4 MPa (n=6) FUS pre-conditioning. Median and interquartile range. (d) Confocal images of Iba1 and GFAP immunofluorescence in cross-sections through ipsilateral and contralateral striata. (e) Bar graph of GFAP grayscale intensity in all groups. No statistically significant differences were observed, indicating that neither the FUS treatments, nor Luc-BPN delivery, elicited astrogliosis. n=6 (Untreated), n=6 (2 MPa groups), or n=8 (4 MPa groups). Means ± S.D.

Finally, we examined Luc-BPN transfected normal brain tissue for overt signs of ischemic injury, astrogliosis, and/or microgliosis by immunolabeling for astrocytes [glial fibrillary acidic protein (GFAP)] and microglia [ionized calcium binding adaptor molecule 1 (Iba1)]. Three days after treatment to deliver Luc-BPN across the BBB, both with and without FUS pre-conditioning, and immediately after ex vivo imaging for luciferase transgene expression, brains were fixed, sectioned, and immunolabeled. GFAP and Iba1 stained sections (Figure 4d) were examined by a board certified neuropathologist (JWM) blinded to the treatment conditions. No evidence of neural damage or ischemic injury was observed. Microglial structure was similar across all groups. Comparisons of grayscale intensity of GFAP stained images revealed no changes in GFAP staining intensity amongst any groups, including completely untreated animals (Figure 4e).

In summary, the primary goal of the current study was to advance new pulsed FUS pre-conditioning schemes for amplifying the dispersion and efficacy of non-viral BPN formulations in the brain delivered across the BBB. To this end, we first demonstrated the ability of selected FUS pulsing sequences to increase the dispersion of BPN in brain tissue following CED. Next, in a clinically-operable treatment paradigm guided entirely by MR imaging, we demonstrated that pre-conditioning brain tissue with pulsed FUS before opening the BBB to permit BPN delivery facilitates multi-fold enhancements in transgene expression. The powerful FUS pre-conditioning strategy presented here may be appended to already-established protocols to markedly augment the efficacy of drug or gene therapy strategies in the CNS.