Design and evaluation of novel PET strategies to evaluate immunotherapy response

Improving cancer treatment outcomes requires early and accurate evaluation of therapeutic efficacy. Existing approaches, such as monitoring late-stage morphological or metabolic changes, performing invasive biopsies, or using liquid biopsies that lack spatial information, are often insufficient. This underscores the urgent need for novel tools to evaluate cancer therapy response. In this work, two novel positron emission tomography (PET) based methods were developed to this purpose. In a first strategy, cancer cell apoptosis was measured using caspase-3-selective activity-based probes (ABPs). A library of six 18F-labeled ABPs was developed with fast binding kinetics to caspase-3 and high selectivity over other proteases. Among the library, the tracer containing a 6-aminohexanoic acid linker ([18F]MICA-316) was the least hydrophilic (LogD = -0.42 ± 0.18) and was selected for further evaluation. [18F]MICA-316 selectively labeled active caspase-3 in apoptotic cell lysates and tumor homogenates and showed increased uptake in apoptotic cells in vitro. Its in vivo performance was tested in a colorectal tumor bearing mouse model treated with the death receptor agonist conatumumab. No significant differences in tracer uptake were observed between conatumumab-treated and IgG1 isotype control-treated groups (0.55 ± 0.33 %ID/cc vs. 0.52 ± 0.12 %ID/cc at 60 minutes post-injection). These findings underscore the challenges of achieving sufficient tumor uptake and retention with this class of tracers. In a second strategy, a novel method for tracking CAR T cells was developed and evaluated using metabolic glycoengineering to introduce azide groups on the cell surface, enabling bioorthogonal labeling with strained alkynes. The radiotracer [18F]DBCO, a lipophilic compound (LogD = 1.40 ± 0.014) with 83.1% plasma protein binding was used for bioorthogonal labeling. Metabolic glycoengineering of CAR T cells was achieved by incubating cells for 48 hours with 25 µM Ac4ManNAz, which did not impair cell viability, expansion, or functionality. The engineered cells were efficiently labeled with DBCO-Cy5, confirming surface azide presentation. Following infusion of azide-tagged CD19 CAR T cells into CD19+ tumor bearing mice, higher tumor uptake and tumor-to-background ratios were observed compared to CD19− tumor-bearing controls at days 1, 3, 4, and 7 post-infusion. These findings demonstrate the feasibility of using metabolic glycoengineering and bioorthogonal chemistry for CAR T-cell imaging and highlight its potential as a surrogate biomarker for therapeutic response. Overall, this work presents the design, synthesis, characterization, and evaluation of novel molecular imaging strategies to improve the early and non-invasive assessment of cancer therapy efficacy.