Identifying Imaging Biomarkers for Manganese Toxicity in Occupationally Exposed Welders

2019-06-10T18:31:45Z (GMT) by David A Edmondson

Manganese (Mn) is an essential element and, at high doses, a neurotoxin that many workers are exposed to daily. Increased Mn body burden due to occupational exposures leads to a parkinsonian disorder that features symptoms such as mood disturbances, cognition deficits, and motor dysfunction. To understand exposed workers’ risk, biomarkers of exposure have been developed using blood, hair, bone, and toenails. None of these biomarkers take into account how much Mn is in the brain and instead rely on the assumption that Mn uptake in these materials is proportional and related to the levels in the brain. One way to measure Mn in the brain is through neuroimaging modalities, such as magnetic resonance imaging and positron emission tomography, however there remains a need to establish reliable neuroimaging biomarkers for Mn exposure and its toxicological effects. This thesis addresses this need.

First, we hypothesized that changes in Mn exposure would be reflected by changes in the MRI relaxation rate R1 and thalamic γ-aminobutyric acid (GABA). As part of a prospective cohort study, 17 welders were recruited and imaged on two separate occasions approximately two years apart. MRI relaxometry was used to assess changes of Mn accumulation in the brain. Additionally, GABA was measured using magnetic resonance spectroscopy (MRS) in the thalamic and striatal regions of the brain. Air Mn exposure ([Mn]air) and cumulative exposure indexes of Mn (Mn-CEI) for the past three months (Mn-CEI3M), past year (Mn-CEI12M), and lifetime (Mn-CEILife) were calculated using personal air sampling and a comprehensive work history, while toenails were collected for analysis of internal Mn body burden. Finally, welders’ motor function was examined using the Unified Parkinson’s Disease Rating Scale (UPDRS). Mn-CEI12M decreased significantly between the first and second scan (Wilcoxon Signed Rank, p = 0.02). ΔMn-CEI3M were correlated with R1 in the substantia nigra (spearman partial correlation, ρ = 0.50, p = 0.036) and thalamic GABA (ρ = 0.66, p = 0.001), but only GABA significantly decreased linearly with Mn-CEI3M (quantile regression, β = 15.22, p = 0.02). Finally, Mn-CEILife influences the change in R1 in the substantia nigra with Δ[Mn]Air, where higher Mn-CEILife lessened the ΔR1 per Δ[Mn]Air (F-test, p = 0.005). While R1 and GABA changed with Mn exposure, UPDRS was unaffected.

Secondly, we hypothesized that occupational exposure to Mn would lead to disturbances in dopamine release (DA), as measured with PET. Excess exposure to manganese (Mn) can lead to symptoms similar to Parkinson’s disease (PD). While symptoms of PD are due to loss of nigrostriatal dopaminergic neurons, there is no DA neuron loss with Mn toxicity. To assess how DA release may be affected by Mn exposure, 6 subjects (3 welders, 3 controls) were scanned with positron emission tomography and [11C]raclopride (a DA D2/D3 receptor antagonist displaceable by endogenous DA) at baseline and during an amphetamine challenge. There were no apparent differences in amphetamine-induced striatal DA release between groups. However, UPDRS motor scores were positively linearly related to [11C]raclopride binding potential (BPND) in the putamen, whereas Mn-CEILife was negatively related to baseline pre-commissural caudate and ventral striatum BPND. The pilot results suggest that [11C]raclopride PET might delineate the cause of mood and motor dysfunction in subjects exposed to Mn.

Third, we hypothesized that advanced data analysis techniques, such as machine learning, would increase our power in finding differences between groups of welders and controls based on exposure and biological outcomes. This study used data from previous studies in occupationally exposed welders and controls. Whole brain relaxometry using MRI measuring the relaxation rate R1 was acquired in 52 welders and 37 controls. Because measures of R1 in selected regions of the brain have been previously found to be proportional to Mn, we hypothesized that an advanced model taking into account the whole brain might be more predictive for Mn exposure. Additionally, because R1 is proportional to Mn in the region, we used a biokinetic model to estimate the amount of excess Mn in the brain from occupational exposures. Support vector machines (SVM) with a linear kernel were trained using leave-one-out cross-validation. Results indicated that models had recall and accuracy better than chance targeting air Mn exposure, years welding, age, and thalamic GABA. In comparison to all models, R1 appears to reliably predict thalamic GABA across all thresholds, which was previously shown to change with increased Mn exposure. This suggests that while R1 may be proportional to Mn, some Mn may not be free to contribute to signal, and instead thalamic GABA might better reflect the overall amount of free Mn in the brain.

Collectively, this thesis is a successful step towards establishing neuroimaging biomarkers of effect from occupational Mn exposure. The MRI relaxation rate R1, with adequate modeling, could eventually be used to measure total Mn brain burden while thalamic GABA might represent a better metric for measuring the neurochemical effects from recent exposures. However, future research should incorporate more endpoints, such as motor tests, mood assessments, and behavior assessments.