Background The natural disease course for glioblastoma (GB) entails invariably grim outcomes for patients. Phagocytic immunotherapies, such as CD47 blockade (e.g. mCD47), have recently demonstrated promise for GB therapy. However, their efficacy is challenged by presence of the blood brain and tumor barriers (BBB/BTB). Transient disruption of the BBB/BTB via focused ultrasound (FUS) and circulating microbubbles (MB) holds promise for improving therapeutic outcomes in the context of mCD47. However, critical questions regarding the optimal protocol for therapeutic antibody delivery with FUS remain. We herein leverage immuno-PET imaging to spatiotemporally map [89Zr]-mCD47 delivery across the BBB/BTB with FUS in an orthotopic GB model. We then use these insights to design a combinatorial paradigm for mCD47 delivery with repeat FUS BBB/BTB-D.

Methods MRI-guided FUS BBB/BTB-D was performed in the presence of systemically circulating MBs in mice with orthotopically implanted GL261 tumors. Mice received i.v. [89Zr]-mCD47 either without FUS, immediately prior to FUS [FUSPRE] or following FUS [FUSPOST]. Subsequently, mice underwent serial static PET/CT imaging followed by terminal ex vivo assessment of antibody biodistribution. A therapeutic paradigm was then executed, wherein GL261-bearing mice received i.v. mCD47 (8 mg/kg) either as monotherapy or in combination with FUS BBB/BTB-D over three sessions spaced three days apart. Overall survival was monitored and tumor outgrowth was tracked via serial contrast-enhanced MRI.

Results Contrast-enhanced MRI confirmed BBB/BTB-D in GL261 tumors (figure 1A). However, PET/CT imaging revealed a lack of tumor-preferential [89Zr]-mCD47 uptake with or without FUSPRE, suggesting that neither condition improved antibody penetrance over that in naïve brain (figure 1B-C). Remarkably, FUSPOST conferred superlative [89Zr]-mCD47 uptake at the site of BBB/BTB-D, boasting between 4.3- to 6.7-fold more uptake relative to other groups (figure 1C). This elevation in uptake was sustained over the time points assessed (0–72 hours post-FUS) (figure 1C-D). Using these insights, we evaluated a rational paradigm (figure 2A) combining mCD47 with repeat FUSPOST BBB/BTB-D (figure 2B-C) for glioma therapy. FUS-mediated delivery of mCD47 across the BBB/BTB significantly constrained tumor outgrowth (figure 2D-E) and enhanced survival (figure 2F) in GL261-bearing mice.

• Download figure
• Open in new tab
• Download powerpoint

Abstract 472 Figure 1

Immuno-PET monitoring of [89Zr]-mCD47 delivery(A) Representative contrast-enhanced T1-weighted MR images of GL261 tumor-bearing brain pre- and post-FUS. Enhancement in the right cerebral hemisphere on pre-FUS imaging indicates baseline barrier disruption induced by the presence of a brain tumor. Expansion of the enhanced region on post-FUS imaging reflects effective FUS-mediated BBB/BTB disruption. (B) Representative axial decay-corrected PET/CT images for each experimental group on Day 14. White arrows denote region of visibly elevated radioactivity at the tumor site targeted with FUS. (C) Whole brain standardized [89Zr]-mCD47 uptake values (SUVs) extracted from serial static PET/CT images obtained between days 14 and 16 post-implantation (0 to 2 days post-[89Zr]-mCD47 injection).%ID/mL =% injected dose per mL. **p<0.01, ****p<0.0001 vs. group(s) indicated. Significance assessed by RM mixed effects model implementing restricted maximum likelihood method, followed by Tukey multiple comparison correction. (D) Tumor-drug exposure for [89Zr]-mCD47 in naïve brain or GL261 tumors, based on integration of SUVs from decay-corrected PET/CT images collected between 0 and 48 hours after BBB/BTB-D and/or [89Zr]-mCD47 injection. Significance assessed by one-way ANOVA followed by Tukey multiple comparison correction. ****p<0.0001 vs. all other groups

• Download figure
• Open in new tab
• Download powerpoint

Abstract 472 Figure 2

Therapeutic impact of FUS-mediated mCD47 delivery(A) Overview of experimental design for evaluating mCD47 delivery to orthotopically implanted GL261 tumors in the context of repeat BBB/BTB-D with FUS. Mice received i.v. mCD47 (8 mg/kg) either alone (FUS-) or following BBB/BTB-D at 0.4 MPa (FUS+). (B) Axial contrast-enhanced T1-weighted MR images of murine GL261-tumors pre- and post-FUS. (C) Mean greyscale intensity (MGI) of contrast enhancement pre- and post-FUS over three separate BBB/BTB-D sessions conducted every three days. Calculated as fold change over contralateral brain. Mean ± SD. **p=0.0023. Significance assessed by RM 2-way ANOVA followed by Sidak’s multiple comparison test. (D) Contrast-enhanced T1-weighted MR images of GL261 tumors on days 14, 17 and 20 post-implantation. ? = image excluded due to poor quality. (E) GL261 tumor outgrowth quantified based on serial MR imaging. Mean ± SD. *p=0.0010. Significance assessed by RM mixed-effects model implementing restricted maximum likelihood method, followed by Sidak’s multiple comparison test. (F) Kaplan-Meier curve depicting overall survival of GL261-bearing mice. n=5–6 mice per group. *p=0.0008. Significance assessed by log-rank (Mantel-Cox) test

Conclusions Taken together, our findings suggest that mCD47 delivery with FUS BBB/BTB-D is a promising therapeutic strategy for GB. For myriad ongoing pre-clinical and clinical evaluations of FUS-mediated immunotherapy delivery, these findings generate timely and compelling insights regarding impact of injection timing on antibody penetrance in brain tumors. This study underscores the outstanding potential role of immuno-PET imaging for rational design and monitoring of response to FUS immunotherapy approaches.

Acknowledgements This study was supported by NIH R01CA197111, R01EB020147, R21NS118278 and the Schiff Foundation (R.J.P.). Additional support from NCI F99/K00 Predoctoral to Postdoctoral Fellow Transition Award (F99CA234954), NSF Graduate Research Fellowship and the Robert R. Wagner Fellowship (N.D.S.).

Ethics Approval This study was prospectively reviewed and approved by the University of Virginia Animal Care and Use Committee.

Consent N/A