CEST MRI for Monitoring Early Treatment Responses in Brain Tumors and Characterizing Molecular Changes in Atherosclerosis

Abstract

Chemical exchange saturation transfer (CEST) magnetic resonance imaging (MRI) is a non-invasive molecular imaging technique, which probes exchangeable solute protons such as amide, amine or lipids, using radiofrequency (RF) irradiation. Monitoring of tissue compositional or microenvironmental changes in the brain allows CEST MRI to detect various brain diseases, including tumor, stroke, Alzheimer’s disease (AD) or Parkinson’s disease (PD). CEST has been widely explored for tumor grading and differentiating between tumor recurrence and necrosis using amide proton transfer (APT) signal, a type of CEST contrasts. APT reflects pH and relative concentrations of amide protons of mobile peptides or proteins. A few studies have assessed early molecular alterations in tumors in response to treatment. Moreover, apart from brain diseases, atherosclerosis also affects the brain molecular exchange environment due to cholesterol dysfunction and inflammation. CEST MRI can be used to study the complex interplay between the atherosclerosis development and the altered exchange environment, especially using another type of CEST contrasts, i.e. relayed nuclear Overhauser effect (rNOE) signals. rNOE indicates the aliphatic protons of lipids and protein, which is related to the cellularity in tumors. In this thesis, we developed two approaches to treat brain tumor and observed early tumor treatment response at a molecular level, prior to a decrease in tumor volume. We also investigated and characterized brain molecular changes associated with atherosclerosis progression. Our findings aim to demonstrate the potential of CEST MRI as a versatile tool for studying brain diseases or treatments in depth.

First, we designed a liposomal hydrogel to treat glioblastoma (GBM), one of the most aggressive brain tumors. Conventional treatment for GBM mostly includes surgical resection followed by chemotherapy, but the high rate (90%) of tumor recurrence limits long-term survival and requires the urgent need for improved treatment monitoring strategies. Here, we developed a soft and self-healing hydrogel encapsulated with liposome, allowing for continuous and direct delivery of high concentrations of chemotherapeutic drugs into the tumors to achieve local treatment. We then applied CEST MRI to monitor drug and liposome release using 2.4 and –3.5 ppm, as well as tumor treatment response using APT and rNOE signals. Within days of hydrogel implantation, we observed sustainable drug release over 5 days and continuous APT signal and tumor volume decrease using CEST MRI and T2-weighted images. The decrease in APT signal of 21.45% over 10 days, along with the decrease in tumor volume of 56.20% over 10 days, could be attributed to a decrease in endogenous proteins or peptides, and the major mechanisms of the chemotherapeutic drugs led to apoptosis, which altered the tumor microenvironment. Critically, these APT signal changes were detected before any observable changes in tumor volume (p = 0.0249), indicating early indicators of treatment response. Based on these findings, we further found out that the APT signal could capture the spatial heterogeneity of the treatment response, where the tumor region near the hydrogel showed the highest number of apoptotic cells of 78.00 ± 1.53% and the lowest APT signal of 5.39 ± 0.65% correlated with lowest number of proliferative cells of 9.32 ± 5.94%. These tumor regional information upon treatment using CEST MRI provides deeper information for enhancing treatment efficacy in a dynamic and non-invasive manner.

Second, photodynamic therapy (PDT) is another efficient treatment for brain tumor. CEST MRI was further utilized to assess the tumor response during PDT in a murine GBM model. Compared to the previous study, this study differed in the type of treatment (local treatment vs. PDT) and the focus of the evaluation (early and regional response vs. overall efficacy). 5-aminolevulinic acid (ALA), a prodrug, generates reactive oxygen species (ROS) to destroy tumor vasculature and kill cancer cells under PDT, yet the treatment of repeated PDT (rePDT) using conventional MRI has rarely been recorded clinically and might not be sufficient to observe the treatment outcome at a molecular level. We proposed to evaluate the treatment efficacy of rePDT using CEST MRI to monitor APT and rNOE signals at 3T over 7 days. rePDT resulted in continuous tumor suppression, yielding the smallest tumor volume at day 7. APT and rNOE signals in the rePDT group were significantly lower (p < 0.05) than those in the control group from day 5, indicating reduced protein content and cellularity. A 17.9% and 11.3% decrease in APT and rNOE were observed in the rePDT group, respectively. Furthermore, we explored the relationship between these signals and cellularity, proliferation and apoptosis following PDT treatment, where histological analysis confirmed moderate correlations between APT and cell proliferation (R = 0.730, p < 0.01) and apoptosis (R = 0.715, p < 0.05), and rNOE and cellularity (R = 0.796, p < 0.01). The specific molecular and microenvironmental changes that are detectable by CEST MRI depend on the type of treatment.

CEST is a robust approach to assess many molecular changes in vivo. As demonstrated in our theranostic applications in brain tumors, we further investigated the molecular events related to the neuropathology in the ApoE –/– mouse model of atherosclerosis. This specific rodent model enable us to study the specific changes that are relevance to neurodegenerative conditions such as cognitive decline, were characterized using CEST MRI. Specifically, we investigated neuroinflammation and alterations in lipid metabolism using APT and rNOE signals. Our results revealed continuously decreasing rNOE signals at –1.6 ppm in the mouse brain in ApoE –/– mice and ApoE –/– mice fed with high fat diet (HFD) compared to age-matched wildtype, indicating the potential changes in lipid metabolism and accumulation, where histological analysis demonstrated highest lipid droplet (LD) density of 19.77 (/mm²) in hippocampus region in ApoE –/– + HFD group. Positive correlations were observed between rNOE signal intensities at –1.6 and –3.5 ppm with LD density (R = 0.7942 at –1.6 and R = 0.4713 at –3.5 ppm). Our study suggests that CEST MRI can detect metabolic changes associated with atherosclerosis in the brain, providing a potential biomarker for monitoring disease progression and evaluation therapeutic interventions.

Date of Award	11 Sept 2025
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Wai Yan Kannie CHAN (Supervisor)