Magnetic Resonance Guided Focused Ultrasound for targeted drug delivery in Diffuse Intrinsic Pontine Glioma

Background

Magnetic Resonance Image-guided Focused Ultrasound (MRgFUS) has been used to achieve transient BBB opening without tissue injury. Delivery of a targeted ultrasonic wave causes an interaction between administered microbubbles and the capillary bed resulting in enhanced vessel permeability. The use of MRgFUS in the brainstem has not previously been shown but could provide value in the treatment of tumours such as Diffuse Intrinsic Pontine Glioma (DIPG) where the intact BBB has contributed to the limited success of chemotherapy.
The focus of this study was to determine whether the use of MRgFUS in this eloquent brain region could be performed without histological injury and functional deficits. Thereafter, we wished to determine if we could enhance brainstem delivery of an effective chemotherapeutic against patient derived DIPG cell lines when combined with MRgFUS in vivo.
Further to this, we sought to develop two pre-clinical murine models of DIPG; a genetically engineered RCAS/Tv-a model (Nestin tva: RCAS PDGFB, RCAS H3.3K27M, p53KO) and a patient derived xenograft (SU-DIPG XVII) to compare their respective BBB permeability and thus determine their suitability for pre-clinical MRgFUS delivered therapeutics.
Lastly, recognising the need for multi-drug therapy, we set out to identify a highly synergistic drug combination therapy that could be tested in our chosen pre-clinical model.
Methods
Female Sprague Dawley rats were randomised to one of four groups: 1) Microbubble administration but no MRgFUS treatment; 2) MRgFUS only; 3) MRgFUS + microbubbles; and 4) MRgFUS + microbubbles + cisplatin. Physiological assessment was performed by monitoring of heart and respiratory rates. Motor function and co-ordination were evaluated by Rotarod and grip strength testing. Histological analysis for haemorrhage (H&E), neuronal nuclei (NeuN) and apoptosis (cleaved Caspase-3) was also performed. A drug screen of eight chemotherapy agents was conducted in three patient-derived DIPG cell lines (SU-DIPG IV, SU-DIPG XIII and SU-DIPG XVII). Doxorubicin was identified as an effective agent. NOD/SCID/GAMMA (NSG) mice were subsequently administered with 5mg/kg of intravenous doxorubicin at the time of one of the following: 1) Microbubbles but no MRgFUS; 2) MRgFUS only; 3) MRgFUS + microbubbles and 4) no intervention. Brain specimens were extracted at 2 hours and doxorubicin quantification was conducted using liquid chromatography mass spectrometry (LC/MS). BBB opening was confirmed by contrast enhancement on T1-weighted MR imaging and positive Evans blue staining of the brainstem.
Nestin-tv-a;p53fl/fl received brainstem injection of transfected DF-1 chicken fibroblast cells containing RCAS-PDGFB, RCAS Cre and RCAS H3.3K27M at postnatal day 2. NOD/SCID/GAMMA mice received brainstem injections of the SU-DIPG XVII cell line. Following tumour development, radiological and histological assessment was performed. A further measure of SU-DIPG XVII BBB permeability was performed by comparison of brainstem doxorubicin uptake with control mice.
A zero induced potency (ZIP) analysis of synergy was conducted between doxorubicin and several drugs classified as either PARP inhibitors, HDAC inhibitors or TK inhibitors in patient derived DIPG cell lines.
Results
Normal cardiorespiratory parameters were preserved across all cohorts of rats. Grip strength and Rotarod testing demonstrating no decline in performance across all groups. Histological analysis showed no evidence of haemorrhage, neuronal loss or increased apoptosis.
Doxorubicin demonstrated cytotoxicity against all three cell lines and is known to have poor BBB permeability. Quantities measured in the brainstem of NSG mice were highest in the group receiving MRgFUS and microbubbles (431.5 ng/g). This was significantly higher than in mice who received no intervention (7.6 ng/g).
On comparison of the Nestin-tva and SU-DIPG XVII mouse models, both demonstrated the key molecular markers of DIPG. However, the SU-DIPX XVII PDX model more faithfully recapitulated the disease due to a diffuse pattern of tumour growth seen histologically and the preservation of the BBB. In contrast, the Nestin-tva model resulted in a well circumscribed and avidly enhancing tumour.
Drug synergy analysis identified the PARP inhibitors Niraparib and Talazoparib as having potency and highly synergistic activity with doxorubicin in vitro. Tyrosine kinase inhibitors (TKI) and Histone Deacetylase inhibitors (HDACi) demonstrated a largely antagonistic action with doxorubicin.
Conclusion
Our data demonstrates both the preservation of histological and functional integrity of the brainstem following MRgFUS for BBB opening and the ability to significantly enhance drug delivery to the region. We have developed the SU-DIPG XVII PDX model which exhibits a diffuse brainstem tumour with an intact BBB as a potential pre-clinical model of DIPG for MRgFUS delivered therapeutics. In addition, we have identified a potent and synergistic in vitro drug combination therapy of doxorubicin and the PARP inhibitors Niraparib and Talazoparib to further assess in vivo.
Taken together, these findings give significant promise to the treatment of brainstem tumours using MRgFUS delivered therapeutics