Can we apply the Mendelian randomization methodology without considering epigenetic effects?

Abstract

Introduction: Instrumental variable (IV) methods have been used in econometrics for several decades now, but have only recently been introduced into the epidemiologic research frameworks. Similarly, Mendelian randomization studies, which use the IV methodology for analysis and inference in epidemiology, were introduced into the epidemiologist's toolbox only in the last decade.

Analysis: Mendelian randomization studies using instrumental variables (IVs) have the potential to avoid some of the limitations of observational epidemiology (confounding, reverse causality, regression dilution bias) for making causal inferences. Certain limitations of randomized controlled trials, such as problems with generalizability, feasibility and ethics for some exposures, and high costs, also make the use of Mendelian randomization in observational studies attractive. Unlike conventional randomized controlled trials (RCTs), Mendelian randomization studies can be conducted in a representative sample without imposing any exclusion criteria or requiring volunteers to be amenable to random treatment allocation.Within the last decade, epigenetics has gained recognition as an independent field of study, and appears to be the new direction for future research into the genetics of complex diseases. Although previous articles have addressed some of the limitations of Mendelian randomization (such as the lack of suitable genetic variants, unreliable associations, population stratification, linkage disequilibrium (LD), pleiotropy, developmental canalization, the need for large sample sizes and some potential problems with binary outcomes), none has directly characterized the impact of epigenetics on Mendelian randomization. The possibility of epigenetic effects (non-Mendelian, heritable changes in gene expression not accompanied by alterations in DNA sequence) could alter the core instrumental variable assumptions of Mendelian randomization.This paper applies conceptual considerations, algebraic derivations and data simulations to question the appropriateness of Mendelian randomization methods when epigenetic modifications are present.

Conclusion: Given an inheritance of gene expression from parents, Mendelian randomization studies not only need to assume a random distribution of alleles in the offspring, but also a random distribution of epigenetic changes (e.g. gene expression) at conception, in order for the core assumptions of the Mendelian randomization methodology to remain valid. As an increasing number of epidemiologists employ Mendelian randomization methods in their research, caution is therefore needed in drawing conclusions from these studies if these assumptions are not met.

Figure 1

Conceptual analogies between a randomized controlled trial (left graph) and Mendelian randomization approach (right graph). Adapted from Nitsch D, et al. [4].

Figure 2

Directed Acyclic Graph (DAG) specifying the core conditions for an instrumental variable with (2B) and without (2A) the presence of a mediator variable in Linkage Disequilibrium (LD) with the instrumental variable of interest.

Figure 3

Three important assumptions in Mendelian Randomization are: a) and b): G associated with X and independent of U. Thus, the effect of G on X is not affected by U. c) Given X and U, G is independent of Y. Thus, the effect of G on Y can be fully assessed by the effect of G on X and then the effect of X on Y, after adjusting for confounders U; i.e. β^ITT = β^XGβ^IV as we have shown in Analysis section. (Adapted from Nitsch D, et al. [4]).

Figure 4

Epigenetic effect present due to inherited altered gene expression (β^XG + β^EG = β^XG*). Compared to Figure 3, when epigenetic effect is present, the core assumptions of MR are violated. Let E denote an environmental factor, which interacts with G at the maternal level. In the presence of this interaction, G and E × G are clearly dependent. Thus, the association between G and X is affected by the E × G interaction term. This violates core assumptions (a) and (b) above. Thus, Mendelian Randomization should be applied with caution if the possibility of epigenetic effects exists. Further, as shown in Analysis section, β^XG* = β^XG+ β^EG and β^ITT = β^XG*β^IV'. Therefore, β^ITT ≠ β^XGβ^IV'. Thus, when Mendelian Randomization is violated, there is a tendency to contravene the stated relationship between β^ITT, β^XG, and β^IV as given by equation (5). From the randomized controlled trial analogy, β^ITT is the intention-to-treat effect; β^IV(biologic effect of received treatment); and β^XG (the effect of G on X).

Figure 5

Schematic representation of possible scenarios of the effects of epigenetics on Mendelian randomization (IL13: Inter-leukin-13 gene).