Chemical exchange saturation transfer (CEST) is an emerging contrast mechanism of magnetic resonance imaging (MRI) to detect molecular changes in diseases. Based on the chemical exchange process between the solute and water protons, CEST can detect molecular changes down to the millimolar range. Although CEST has been translated from the preclinical to the clinical applications, the clinical application has been focused on brain cancer imaging. Other biomedical applications could be challenging, and yet rewarding due to its unique contrast mechanism. First, CEST is susceptible to in vivo microenvironmental changes, such as pH, temperature, and concentration of exchangeable protons. The confounding signal origins can make the analysis challenging. Second, different postprocessing methods could result in different interpretations of the data. Amide-based CEST with magnetization transfer ratio with asymmetric analysis (MTRasym) is acquired for brain tumor diagnosis and prognosis in the clinic. Recent studies with other postprocessing methods, such as Lorentzian fitting and apparent exchange dependent-relaxation (AREX), suggested that other signals of the CEST spectrum (Z-spectrum) from the aliphatic protons and amine protons could provide additional information about the underlying pathology. Third, the application of CEST in diagnosis of diseases other than brain tumor is relatively young, thus further preclinical studies could reveal molecular alterations in various neuropathology. This thesis focuses on the CEST application on two neurological disorders, hemorrhagic stroke and Parkinson’s disease (PD), at preclinical 3T MRI. By monitoring these diseases in a spatiotemporal manner during disease progression and upon treatment in mouse models, valuable information can be provided to reveal the underlying molecular changes to facilitate the clinical application of CEST in these neurological disorders. The experimental design and the key findings are summarized in the following paragraphs.

Chapter 1 introduced the principles of molecular MRI approaches and the current status of imaging stroke and Parkinson’s disease. Chapter 2 studied the possibility of CEST in imaging hemorrhagic stroke under iron chelation treatment. Stroke imaging has been a thriving research field in CEST after discovering CEST detectability in the pH-related hallmark in the penumbra of ischemic stroke. However, hemorrhagic stroke, which is another category of stroke, does not share the same pathology, and the CEST-related research is much less. One of the critical hurdles comes from iron pathology, which can affect CEST detection by distorting the magnetic field. Our study proposed to address this issue by using AREX for data processing. Our phantom data showed that AREX could suppress the influence of longitudinal relaxation (T1 contribution) from iron, and the CEST signal would not be affected with iron concentration up to 10 mM. With the optimized protocol, we studied the spatiotemporal changes of the hematoma of an intracerebral hemorrhage (ICH) mouse model for two weeks. Two distinctive trends could be observed by extracting the signals at -3.5 ppm and 3.5 ppm from the Z-spectrum, which indicate the relayed nuclear Overhauser effect (rNOE) and amide proton transfer (APT). rNOE could identify the core and peri-hematoma on day 3 and day 14 and by APT on day 1, day 7, and day 14. To further study the related neuropathology in ICH, immunohistochemistry was performed. We found that rNOE primarily correlated with the myelin pathology (P<0.05), and APT could reflect the cellularity increase at hematoma up to day 7. To assess the theranostic property of CEST for disease prognosis, the mice were given deferoxamine (DFX), a clinically available iron chelation drug, or a placebo. We observed a 25.7% reduction (P<0.05) in rNOE on day 3 in the treatment group, while the treatment effect was not observable in APT. Our multiparametric CEST readouts could reflect the iron chelation treatment response and regeneration in the hematoma based on cellularity and lipid-related events, providing valuable information for ICH assessment to monitor treatment and outcome.

The latter section of this thesis has studied the neuropathology of PD in relation to the glymphatic system by using an advanced CEST technique, Magnetization transfer Indirect Spin Labeling (MISL).

MISL is an emerging contrast mechanism to quantify the cerebrospinal fluid (CSF)-tissue water exchange. Reveal molecular alterations could enable interactive interventions in PD. In Chapter 3, we first explored the connection of MISL with the glymphatic system, which is a newly discovered waste clearance system of the brain. We studied the aquaporin 4 (AQP4) inhibition effect on MISL. AQP4 is the major channel protein in the brain to mediate water movement in the glymphatic system. N-(1,3,4-thiadiazol-2-yl) pyridine-3-carboxamide dihydrochloride (TGN-020) was applied to transiently inhibit the AQP4 function of the wildtype mice during a dynamic MISL acquisition. We observed that the MISL signal dropped to a minimum of 28 minutes after TGN-020 administration (15.2%, P<0.05) and persisted up to 60 minutes (14.7%, P<0.05). Such change was not observable in magnetization transfer (MT), implying that MISL had the sensitivity to detect AQP4 changes in the glymphatic system.

In Chapter 4, we applied MISL and CEST to study the changes in a 1-methyl-4-phenyl-2,3-dihydropyridi-nium (MPTP)-induced PD mouse model. The development of the mouse model was confirmed by the depletion in tyrosine hydroxylase (TH) immunoreactivity at the substantia nigra pars compacta. After that, the changes of three consecutive brain slices covering the nigrostriatal pathway were monitored for up to 28 days. MISL detected a prolonged signal reduction after the onset of PD. The MISL signal gradually dropped and became stable after day 7. Significance was observed on day 7 (-19.0% in MISL anterior slice, P<0.05), day 14 (-13.8% in MISL posterior slice, P<0.05), and day 28 (-17.6% in MISL anterior slice, P<0.05). The AQP4 function of PD mice was further assessed using TGN-020, and a significant MISL drop was observed after the administration. Together with immunohistochemical validation, the MISL decrease in PD mice was explained by the regional change of the AQP4 expression but not the functional change. This result suggests that AQP4 could be a potential biomarker for PD diagnosis by MISL. For CEST, we observed regional signal changes over time. The rNOE signal of the three brain slices increased on day 7 (1.49% in CEST middle slice, P<0.05). Structures along the nigrostriatal pathway, including the thalamus, hippocampus, and substantia nigra, showed an average of 2% increase (P<0.05). After that, the overall rNOE signal dropped (-0.98% in CEST anterior slice, P<0.01), and no significance was observed on day 28 compared with the control. The overall APT signal showed a trend of a decrease on day 7. Nevertheless, the changes were not significant. The rNOE signal could indicate the regional change of cellularity and the transition of the disease. Since MISL and CEST at 3T are clinically translatable, our findings can provide a non-invasive glymphatic function assessment for early PD detection.

To conclude, the first half of the thesis has demonstrated the feasibility of rNOE and APT to reveal the spatiotemporal changes of lipid and protein, respectively, in ICH at 3T. This is the first study to demonstrate that rNOE can detect the DFX treatment effect on ICH on day 3, which was further validated to be the myelin-lipid change. In the second half of the thesis, we have shown that MISL and CEST could be used to study the early molecular changes in a PD mouse model. MISL was further validated with the sensitivity in detecting AQP4 changes, which could reflect the glymphatic-related pathology in PD. Nonetheless, the multiparametric CEST readouts, such as APT, have disclosed other confounding molecular changes, which will need further validations. In general, the findings here have demonstrated the use of CEST in disease detection and prognosis, which has high potential to be translated to the clinical level in the future.