We recently demonstrated that fibrosis determined by APRI or FIB4 cutoff scores is associated with CNS changes including higher T1-weighted signal intensities in basal ganglia structures associated with smaller volume selective to the globus pallidus (Kwong and Zahr 2023). Here we demonstrate that in people with APRI or FIB4 identified fibrosis relative to those without serum biomarkers of fibrosis, levels of striatal Cho and tCr and pontine mI are high. These findings are consistent with the literature demonstrating higher than control levels of Cho, tCr, and mI in various ROIs, including basal ganglia, in individuals with HCV (Forton et al. 2001; Forton et al. 2002; McAndrews et al. 2005; Reichardt et al. 2022). The current study, however, extends the literature by suggesting that liver fibrosis and not HCV per se is responsible for the elevation in striatal Cho, whereas elevations in mI and tCr may be better explained by the presence of HCV (cf., Table 5). This interpretation comports with a study that used APRI cutoffs to determine neuropsychological performance in those with HIV or HCV and concluded that liver fibrosis has a contribution to cognitive performance independent of HCV and HIV (Valcour et al. 2016). Here we suggest that liver fibrosis is associated with high levels of striatal Cho independent of HCV, HIV, or AUD. Longitudinal studies including people with liver disease from multiple etiologies and at earlier disease stages will be necessary to determine whether an increase in Cho is a feature of early, pre-cirrhotic liver disease whereas reduced Cho occurs in later liver disease stages (cf., Geissler et al. 1997; Lee et al. 1999; Ahluwalia et al. ; Mardini et al. 2011; Balata et al. 2003; Cordoba et al. 2001; Laubenberger et al. 1997; Ciecko-Michalska et al. 2012; Meng et al. 2015).
The current study also demonstrated that higher levels of striatal Cho are associated with a higher pallidal T1 signal. The hyperintense pallidal signal observed in cirrhosis is unusual in that it is present on T1- but not T2- weighted images and is not vulnerable to gadolinium enhancement (Pujol et al. 1993; Awada et al. 1995; Zeneroli et al. 1991; Brunberg et al. 1991). This signal was initially proposed to arise from the deposition of paramagnetic substances such as Mn (Inoue et al. 1991; Newland et al. 1989) or to lipid accumulation (Young 1984) because these are among the few naturally occurring substances that are known to reduce T1 relaxation times (Ginat and Meyers 2012) without affecting T2-weighted imaging (i.e., calcification or hemorrhaging would affect both T1 and T2 signaling). Due to converging lines of evidence, the mechanism involving Mn deposition as contributing to the altered pallidal T1-signal gained the greatest traction (Spahr et al. 1996; Krieger et al. 1995; Forton et al. 2004; Rose et al. 1999). Caveats regarding Mn deposition as the underlying mechanism of pallidal T1-weighted signal changes include the following considerations: systemic Mn levels are not related to pallidal signal intensity (Maffeo et al. 2014; Krieger et al. 1996); in iron-deficiency anemia, serum Mn levels are high but the pallidal signal is not present (Kim et al. 2005); and postmortem examination of tissue from patients with liver disease reveals Mn deposition beyond the pallidum in caudate, putamen, substantia nigra, and cerebellum (Klos et al. 2006; Krieger et al. 1995) and high levels of other metals such as copper in similar regions (Maeda et al. 1997). Further, animal exposure studies do not show preferential Mn loading to basal ganglia structures (Takeda, Sawashita, and Okada 1998; Finkelstein et al. 2008); notable Mn accumulation also occurs in olfactory bulb, frontal cortex, hippocampus, and cerebellum (Elder et al. 2006; Finkelstein et al. 2008; Ma et al. 2018; Fitsanakis et al. 2008; Pautler, Mongeau, and Jacobs 2003; Long et al. 2014). Indeed, an unresolved question with respect to the Mn hypothesis is a mechanistic explanation for preferential deposition in the pallidum.
Iron induced, oxidative damage to lipids is a proposed mechanism of liver disease pathophysiology (Britton 1996; Britton, Subramaniam, and Crawford 2016; Ahmed, Latham, and Oates 2012). A similar mechanism may explain pallidal susceptibility to chronic liver disease. The globus pallidus is unique among brain regions in having high iron and high myelin content: in healthy basal ganglia, high iron accounts for a hypointense pallidum on T2-weighted images, whereas the caudate and putamen are isointense relative to cortical gray matter (Hegde et al. 2011); a high pallidal T1-signal relative to other basal ganglia regions is associated with its greater myelin content (Henkelman, Watts, and Kucharczyk 1991; Van Cauter et al. 2020; Zaitout et al. 2014). Iron-induced damage of pallidal myelin may result in the accumulation of lipids (i.e., damaged biological membranes) which can accentuate water proton relaxation and contribute to increasing the pallidal T1-weighted signal {Brunberg, 1991 #9741;Warakaulle, 2003 #9792;Lai, 1999 #9786;Young, 1984 #9638;Gupta, 2017 #9789; Kucharczyk, 1990 #9874;Ginat, 2012 #11087}. As the Cho signal arises from choline-containing compounds such as membrane phospholipids and as Cho elevations are consistent with cell membrane turnover (Adalsteinsson, Sullivan, and Pfefferbaum 2002; Mader et al. 2008), a significant positive relationship between striatal Cho and the pallidal T1-weighted signal may signify impaired membrane homeostasis that may precede demyelination in chronic liver disease {Hathout, 2015 #11127}. Indeed, moderate myelin loss and lipidic droplets were observed in postmortem pallidal tissue in those who exhibited altered pallidal signals in vivo {Kulisevsky, 1992 #9731}. Further support for the hypothesis that disruption of myelin homeostasis represented by high levels of Cho may underlie pallidal T1 changes in liver disease will require larger, longitudinal neuroimaging studies across the liver disease time course.
High pallidal signal intensities in cirrhosis have been associated with excessive postural body sway (Kim et al. 2007), tremor of the hands (Pujol et al. 1993), and impaired performance on grooved pegboard (Chang et al. 2009; Chang et al. 2010; Kulisevsky et al. 1992), but consistent functional consequences of T1-signal alterations have not been forthcoming (cf., Shin and Park 2017). In our recent study, higher pallidal signal intensity was associated with greater postural instability in both eyes open and closed conditions (Kwong and Zahr 2023). Here, consistent with our own previous finding and the literature, higher pallidal signal intensities were associated with ataxia with eyes closed.
Together, previous MRS results demonstrating changes to tCr in HCV (Bladowska et al. 2013; McAndrews et al. 2005), 31phosphoros MRS experiments demonstrating bioenergetic abnormalities (e.g., increase in phosphocreatine) in patients with minimal HE or stable overt chronic HE (Patel et al. 2000), and the current results suggest that tCr should not be used as a referent in MRS studies of liver disease.
Limitations of the current study include the relatively small number of individuals with APRI-defined fibrosis (n = 13) and the inclusion of those with liver fibrosis due to several etiologies (i.e., AUD, HIV, HCV). Indeed, there is some evidence suggesting that CNS changes could be unique to liver disease etiology (e.g., alcohol vs. non-alcohol related cirrhosis) (Miese et al. 2006). Further, although we demonstrate that the effects of fibrosis on striatal Cho are greater than the effects of HCV per se, the relatively high representation of HCV in the APRI and FIB4 groups precludes a clear dissociation. Finally, elevated levels of Cho have been interpreted as reflecting energy failure (Zahr et al. 2014; Brooks et al. 2000; Callot et al. 2008; Bizzi et al. 2001; Fischer-Smith et al. 2004) and may not reflect membrane degradation or demyelination as suggested here.
In conclusion, using serum biomarkers of liver fibrosis, higher than control values of Cho were identified in striatum and related to the pallidal T1-weighted signal providing initial evidence for an alternative mechanism to Mn deposition to explain focal signal brightening in liver disease that may instead arise from accumulation of lipids related to myelin disruption.