Multi-parametric MRI of the physiology and optics of the in-vivo lens

Abstract

The majority of cases of age-related vision degeneration in the world today are associated with the lens pathologies presbyopia and cataract. Although effective treatments for presbyopia and cataract exist, the pathophysiological mechanisms responsible for these two lens pathologies are not fully understood due to the lack of effective non-invasively methods to study the physiological optics of the lens under in-vivo conditions. The main parameters that link the cellular physiology of the lens to its optics are its geometry (shape) and gradient of refractive index (water to protein ratio). In a series of previous studies, the Molecular Vision Laboratory (MVL) has used an MRI-based optical modelling approach to show in organ cultured bovine lenses, that the lens actively maintains its optical properties by regulating lens water transport (Donaldson et al., 2017). In this thesis, I have optimised the MRI-based optical modelling approach developed to study the lens in-vitro, to enable it to be successively applied in-vivo to human subjects across a range of ages, and transgenic mouse models of lens cataract. I initially developed and validated in-vivo MRI protocols to extract parameters (T2 and geometry) from the young human lens, which when combined with biometric measures of the eye and optical modelling (ZEMAX), enabled the physiological optics of the human lens to be investigated (Pan et al., 2019). Quantitative measurements of lens geometry and T2 showed high inter-day repeatability, and the resultant optical models produced accurate refractive error and lens power estimations. To extend this platform to older subjects, I optimised the modelling platform by implementing equations (Navarro et al., 2007) to capture the GRIN profile change that develops with age more effectively. Then, I incorporated an additional “age parameter” into the T2-n calibration to improve the ability of the model to predict each subject’s refractive error. These improvements to the modelling platform we able to correctly reproduce the decrease in lens power that is observed clinically with advancing age, known as the lens paradox (Lie et al., 2020). To measure lens water, I also optimised in-vivo imaging protocols to perform T1 & PD mapping of the lens by developing a scan protocol that allowed free water (T1) and total water (PD) content to be rapidly and reliably estimated. The validity of this protocol was first established using a phantom, before demonstrating its feasibility for in-vivo assessment of relative contributions of free and total water in the human lens using a cohort of young subjects. In this cohort, I found that while the T1 profiles of all subjects had the same parabolic shape, the magnitudes of their T1 values were highly variable. In contrast, the water content of all participants was relatively constant across all regions of the lens. Since this finding was repeatable, I concluded that the observed subject to subject variability of T1 was not due to technical issues, but inherent biological variability in free water content between subjects. In a second project, I had also developed and optimised in-vivo MRI protocols to study the physiological optics of the mouse lens. This project was conducted at SUNY, Stony Brook, New York, using a pre-clinical 7T MRI scanner equipped with a customised coil for the mouse eye. Experiments were performed under the guidance of Dr Eric Muir and utilised transgenic mouse lines supplied by Professor Thomas White. I helped collect the images and subsequently developed the post-processing protocols to extract parameters from the acquired images and perform quantitative analysis. To convert the acquired T2 data to refractive index (n), I utilised published X-ray interferometry measurements of n to generate a unique T2-n calibration for the mouse lens (Muir et al., 2020). By combining MRI measurements of ocular biometry and lens geometry and GRIN into ZEMAX, I created customised eye models for mouse, from which the optical parameters of the lens can be extracted. This platform was applied to wildtype C57BL/6 mice and two mouse lines in which the gap junction coupling was either decreased (Cx46Het) or increased (Cx50KI46) by genetically modifying the expression of Cx46. These proof of principles experiments showed that reducing Cx46 mediated gap junction coupling had no significant effects on in either water content, GRIN or geometry in Cx46Het mice relative to the WT mice. In contrast, increasing gap junction coupling led to a significant increase in water content and decrease in refractive index, respectively, in the nucleus of Cx50KI46 mice, which together reduced the optical power of the lens. In summary, the MRI-based optical modelling protocols I have developed in this thesis have the potential to allow non-invasive imaging of the human lens in-vivo, which can facilitate longitudinal studies into how age affects the physiological optics of the lens. In addition, the application of this platform to study a variety of transgenic mice models will increase our understanding of the link between the molecular and cellular physiology of the lens and the overall optical properties of the lens, which will provide new insights into the mechanisms of lens ageing that result in presbyopia in middle age and cataract in the elderly.