Using Mendelian randomization to understand and develop treatments for neurodegenerative disease

Methodological considerations for observational studies of treatment effectiveness in neurology: a clinician's guide

Data from cohorts, registries, randomised trials, electronic medical records and administrative claims databases have increasingly been used to inform the use of therapies for neurological diseases. While novel sophisticated methods are enabling us to use existing data to guide treatment decisions, the complexity of statistical methodology is making appraisal of clinical evidence increasingly demanding. In this narrative review, we provide a brief overview of the most commonly used methods for evaluation of treatment effectiveness in neurology. This primer discusses complementarity of randomised and non-randomised study designs, sources of observational data, different forms of bias and the appropriate mitigation strategies, statistical significance, Bayesian approaches and provides an overview of multivariable regression models, propensity score-based models, causal inference, mediation analysis and Mendelian randomisation.

A causal relationship between gut microbiota and subcortical brain structures contributes to the microbiota-gut-brain axis: a Mendelian randomization study

A correlation between gut microbiota and brain structure, referring to as a component of the gut-brain axis, has been observed in observational studies. However, the causality of this relationship and its specific bacterial taxa remains uncertain. To reveal the causal effects of gut microbiota on subcortical brain volume, we applied Mendelian randomization (MR) studies in this study. Genome-wide association study data were obtained from the MiBioGen Consortium (n = 18,340) and the Enhancing Neuro Imaging Genetics through Meta-Analysis Consortium (n = 13,170). The primary estimate was obtained utilizing the inverse-variance weighted, while heterogeneity and pleiotropy were assessed using the Cochrane Q statistic, MR Pleiotropy RESidual Sum and Outlier, and MR-Egger intercept. Our findings provide strong evidence that a higher abundance of the genus Parasutterella is causally correlated with a decrease in intracranial volume (β = -30,921.33, 95% CI -46,671.78 to -15,170.88, P = 1.19 × 10-4), and the genus FamilyXIIIUCG001 is associated with a decrease in thalamus volume (β = -141.96, 95% CI: -214.81 to -69.12, P = 1.0× 10-4). This MR study offers novel perspectives on the intricate interplay between the gut microbiota and subcortical brain volume, thereby lending some support to the existence of the microbiota-gut-brain axis.

Associations of cardiovascular risk factors and lifestyle behaviors with neurodegenerative disease: a Mendelian randomization study

Previous observational studies reported that midlife clustering of cardiovascular risk factors and lifestyle behaviors were associated with neurodegenerative disease; however, these findings might be biased by confounding and reverse causality. This study aimed to investigate the causal associations of cardiovascular risk factors and lifestyle behaviors with neurodegenerative disease, using the two-sample Mendelian randomization design. Genetic variants for the modifiable risk factors and neurodegenerative disease were extracted from large-scale genome-wide association studies. The inverse-variance weighted method was used as the main analysis method, and MR-Egger regression and leave-one-out analyses were performed to identify potential violations. Genetically predicted diastolic blood pressure (DBP: OR per 1 mmHg, 0.990 [0.979-1.000]), body mass index (BMI: OR per 1 SD, 0.880 [0.825-0.939]), and educational level (OR per 1 SD, 0.698 [0.602-0.810]) were associated with lower risk of late-onset Alzheimer's disease (LOAD), while genetically predicted low-density lipoprotein (LDL: OR per 1 SD, 1.302 [1.066-1.590]) might increase LOAD risk. Genetically predicted exposures (including LDL and BMI) applied to familial AD showed the same effect. The association of LDL was also found with Amyotrophic lateral sclerosis (ALS) (LDL: OR per 1 SD, 1.180 [1.080-1.289]). This MR analysis showed that LDL, BMI, BP, and educational level were causally related to AD; a significant association between LDL and ALS risk, as well as the potential effect of sleep duration on PD risk, were also revealed. Targeting these modifiable factors was a promising strategy of neurodegenerative disease prevention.

A cause-effect relationship between Graves' disease and the gut microbiome contributes to the thyroid-gut axis: A bidirectional two-sample Mendelian randomization study

Background

An association between Graves' disease (GD) and the gut microbiome has been identified, but the causal effect between them remains unclear.

Methods

Bidirectional two-sample Mendelian randomization (MR) analysis was used to detect the causal effect between GD and the gut microbiome. Gut microbiome data were derived from samples from a range of different ethnicities (18,340 samples) and data on GD were obtained from samples of Asian ethnicity (212,453 samples). Single nucleotide polymorphisms (SNPs) were selected as instrumental variables according to different criteria. They were used to evaluate the causal effect between exposures and outcomes through inverse-variance weighting (IVW), weighted median, weighted mode, MR-Egger, and simple mode methods. F-statistics and sensitivity analyses were performed to evaluate bias and reliability.

Results

In total, 1,560 instrumental variables were extracted from the gut microbiome data (p< 1 × 10⁵). The classes Deltaproteobacteria [odds ratio (OR) = 3.603] and Mollicutes, as well as the genera Ruminococcus torques group, Oxalobacter, and Ruminococcaceae UCG 011 were identified as risk factors for GD. The family Peptococcaceae and the genus Anaerostipes (OR = 0.489) were protective factors for GD. In addition, 13 instrumental variables were extracted from GD (p< 1 × 10^-8), causing one family and eight genera to be regulated. The genus Clostridium innocuum group (p = 0.024, OR = 0.918) and Anaerofilum (p = 0.049, OR = 1.584) had the greatest probability of being regulated. Significant bias, heterogeneity, and horizontal pleiotropy were not detected.

Conclusion

A causal effect relationship exists between GD and the gut microbiome, demonstrating regulatory activity and interactions, and thus providing evidence supporting the involvement of a thyroid-gut axis.

Causal inference on neuroimaging data with Mendelian randomisation

While population-scale neuroimaging studies offer the promise of discovery and characterisation of subtle risk factors, massive sample sizes increase the power for both meaningful associations and those attributable to confounds. This motivates the need for causal modelling of observational data that goes beyond statements of association and towards deeper understanding of complex relationships between individual traits and phenotypes, clinical biomarkers, genetic variation, and brain-related measures of health. Mendelian randomisation (MR) presents a way to obtain causal inference on the basis of genetic data and explicit assumptions about the relationship between genetic variables, exposure and outcome. In this work, we provide an introduction to and overview of causal inference methods based on Mendelian randomisation, with examples involving imaging-derived phenotypes from UK Biobank to make these methods accessible to neuroimaging researchers. We motivate the use of MR techniques, lay out the underlying assumptions, introduce common MR methods and focus on several scenarios in which modelling assumptions are potentially violated, resulting in biased effect estimates. Importantly, we give a detailed account of necessary steps to increase the reliability of MR results with rigorous sensitivity analyses.

A review of Mendelian randomization in amyotrophic lateral sclerosis

Amyotrophic lateral sclerosis is a relatively common and rapidly progressive neurodegenerative disease that, in the majority of cases, is thought to be determined by a complex gene-environment interaction. Exponential growth in the number of performed genome-wide association studies combined with the advent of Mendelian randomization is opening significant new opportunities to identify environmental exposures that increase or decrease the risk of amyotrophic lateral sclerosis. Each of these discoveries has the potential to shape new therapeutic interventions. However, to do so, rigorous methodological standards must be applied in the performance of Mendelian randomization. We have reviewed Mendelian randomization studies performed in amyotrophic lateral sclerosis to date. We identified 20 Mendelian randomization studies, including evaluation of physical exercise, adiposity, cognitive performance, immune function, blood lipids, sleep behaviours, educational attainment, alcohol consumption, smoking and type 2 diabetes mellitus. We have evaluated each study using gold standard methodology supported by the Mendelian randomization literature and the STROBE-Mendelian randomization checklist. Where discrepancies exist between Mendelian randomization studies, we suggest the underlying reasons. A number of studies conclude that there is a causal link between blood lipids and risk of amyotrophic lateral sclerosis; replication across different datasets and even different populations adds confidence. For other putative risk factors, such as smoking and immune function, Mendelian randomization studies have provided cause for doubt. We highlight the use of positive control analyses in choosing exposure single nucleotide polymorphisms (SNPs) to make up the Mendelian randomization instrument, use of SNP clumping to avoid false positive results due to SNPs in linkage and the importance of multiple testing correction. We discuss the implications of survival bias for study of late age of onset diseases such as amyotrophic lateral sclerosis and make recommendations to mitigate this potentially important confounder. For Mendelian randomization to be useful to the amyotrophic lateral sclerosis field, high methodological standards must be applied to ensure reproducibility. Mendelian randomization is already an impactful tool, but poor-quality studies will lead to incorrect interpretations by a field that includes non-statisticians, wasted resources and missed opportunities.

Growth Factors and Their Roles in Multiple Sclerosis Risk

Background

Previous studies have suggested essential roles of growth factors on the risk of Multiple Sclerosis (MS), but it remains undefined whether the effects are causal.

Objective

We applied Mendelian randomization (MR) approaches to disentangle the causal relationship between genetically predicted circulating levels of growth factors and the risk of MS.

Methods

Genetic instrumental variables for fibroblast growth factor (FGF) 23, growth differentiation factor 15 (GDF15), insulin growth factor 1 (IGF1), insulin-like growth factor binding proteins 3 (IGFBP3) and vascular endothelial growth factor (VEGF) were obtained from up-to-date genome-wide association studies (GWAS). Summary-level statistics of MS were obtained from the International Multiple Sclerosis Genetics Consortium, incorporating 14,802 subjects with MS and 26,703 healthy controls of European ancestry. Inverse-variance weighted (IVW) MR was used as the primary method and multiple sensitivity analyses were employed in this study.

Results

Genetically predicted circulating levels of FGF23 were associated with risk of MS. The odds ratio (OR) of IVW was 0.63 (95% confidence interval [CI], 0.49-0.82; p < 0.001) per one standard deviation increase in circulating FGF23 levels. Weighted median estimators also suggested FGF23 associated with lower MS risk (OR = 0.67; 95% CI, 0.51-0.87; p = 0.003). While MR-Egger approach provided no evidence of horizontal pleiotropy (intercept = -0.003, p = 0.95). Results of IVW methods provided no evidence for causal roles of GDF1, IGF1, IGFBP3 and VEGF on MS risks, and additional sensitivity analyses confirmed the robustness of these null findings.

Conclusion

Our results implied a causal relationship between FGF23 and the risk of MS. Further studies are warranted to confirm FGF23 as a genetically valid target for MS.

On the Use of TMS to Investigate the Pathophysiology of Neurodegenerative Diseases

Neurodegenerative diseases are a collection of disorders that result in the progressive degeneration and death of neurons. They are clinically heterogenous and can present as deficits in movement, cognition, executive function, memory, visuospatial awareness and language. Transcranial magnetic stimulation (TMS) is a non-invasive brain stimulation tool that allows for the assessment of cortical function in vivo. We review how TMS has been used for the investigation of three neurodegenerative diseases that differ in their neuroanatomical axes: (1) Motor cortex-corticospinal tract (motor neuron diseases), (2) Non-motor cortical areas (dementias), and (3) Subcortical structures (parkinsonisms). We also make four recommendations that we hope will benefit the use of TMS in neurodegenerative diseases. Firstly, TMS has traditionally been limited by the lack of an objective output and so has been confined to stimulation of the motor cortex; this limitation can be overcome by the use of concurrent neuroimaging methods such as EEG. Given that neurodegenerative diseases progress over time, TMS measures should aim to track longitudinal changes, especially when the aim of the study is to look at disease progression and symptomatology. The lack of gold-standard diagnostic confirmation undermines the validity of findings in clinical populations. Consequently, diagnostic certainty should be maximized through a variety of methods including multiple, independent clinical assessments, imaging and fluids biomarkers, and post-mortem pathological confirmation where possible. There is great interest in understanding the mechanisms by which symptoms arise in neurodegenerative disorders. However, TMS assessments in patients are usually carried out during resting conditions, when the brain network engaged during these symptoms is not expressed. Rather, a context-appropriate form of TMS would be more suitable in probing the physiology driving clinical symptoms. In all, we hope that the recommendations made here will help to further understand the pathophysiology of neurodegenerative diseases.

Intricacies of aetiology in intrafamilial degenerative disease

The genetic underpinnings of late-onset degenerative disease have typically been determined by screening families for the segregation of genetic variants with the disease trait in affected, but not unaffected, individuals. However, instances of intrafamilial etiological heterogeneity, where pathogenic variants in a culprit gene are not shared among all affected family members, continue to emerge and confound gene-discovery and genetic counselling efforts. Discordant intrafamilial cases lacking a mutation shared by other affected family members are described as disease phenocopies. This description often results in an over-simplified acceptance of an environmental cause of disease in the phenocopy cases, while the role of intrafamilial genetic heterogeneity, shared de novo mutations or epigenetic aberrations in such families is often ignored. On a related note, it is now evident that the same disease-associated variant can be present in individuals exhibiting clinically distinct phenotypes, thereby genetically uniting seemingly unrelated syndromes to form a spectrum of disease. Herein, we discuss the intricacies of determining complex degenerative disease aetiology and suggest alternative mechanisms of disease transmission that may account for the apparent missing heritability of disease.