Multiple Sclerosis Journal - October 2017 - 1466

Multiple Sclerosis Journal 23(11)
protective or detrimental effect of adiponectin in MS
cannot be definitively ruled out, and further studies
will be necessary to more clearly ascertain adiponectin's role. Notwithstanding, this study suggests that a
substantial, lifelong alteration in adiponectin levels
would be necessary to influence the risk of disease, if
adiponectin indeed plays a causal role therein.
Observational studies aiming to shed light on the
clinical relevance of adiponectin levels in MS have
yielded variable results.15-18,35 Observational studies such as these are susceptible to bias due to residual confounding, in addition to a number of other
factors that may bias observational studies. While
the potentially confounding effects of BMI were
accounted for in all of these studies, there are several related physiological effects which were not
likely controlled for through the use of BMI as a
measure of obesity and which could have influenced
the reported associations. For example, differences
in adipose tissue amount and location can influence
adiponectin concentrations, as production of adiponectin is differentially regulated in visceral and
subcutaneous adipocytes.36 These differences in
adipose distribution are not accounted for in BMI
calculations. Differential clearance through the
liver could also influence measurements of adiponectin in such studies.36 One strength of this
study is that it utilizes a method of analysis which
largely overcomes confounding, due to the random
assortment of alleles at conception.
As this study assessed the association between lifelong genetically increased adiponectin levels and the
odds of development of MS, the findings reported
here suggest that adiponectin is not an ideal preventative treatment target for MS. Adiponectin's therapeutic role in MS following disease onset, on the
other hand, cannot be ascertained based on the present findings. Interestingly, two of the adiponectinmodulating SNPs investigated in this study
(rs601339 and rs7955516) are located near genes
implicated in both the preventative and the therapeutic treatment of MS (GPR109A37 and PDE3A,38
respectively). In addition, adiponectin treatment following EAE induction in rodents has been shown to
attenuate the clinical course of EAE, findings suggestive of a potential therapeutic role of adiponectin
in MS following disease onset.13
Observational studies5-8 and MR analyses9 have indicated that increased body weight and BMI render
individuals more susceptible to MS. As an adipokine
with anti-inflammatory properties and which is
negatively correlated with BMI, adiponectin is a
1466

biological candidate of interest in the investigation of
the underlying causal pathway of MS. While the present findings cannot rule out the possibility of a protective or detrimental role for adiponectin in MS
etiology, they suggest that adiponectin's role in the
causal pathway of this disease is likely to be small.
Further studies will be necessary to ascertain which
biological factors drive the causal association between
BMI and MS.
This study has important limitations. The possibility
of residual pleiotropy biasing our estimates remains,
despite the sensitivity analyses conducted. MR-Egger
results can be biased when the effect on pleiotropic
pathways is proportional to its effect on adiponectin
level. Interestingly, genetic variation at adiponectinencoding gene ADIPOQ was marginally associated
with risk of MS (p = 0.08), and variation at this locus
is less likely to influence MS risk independent of adiponectin than the other SNPs investigated. In addition, it is impossible, using current methods, to
directly assess the extent to which canalization, or
developmental compensation, may have influenced
our results. While variation in adiponectin level
explained by ADIPOGen SNPs is relatively high
(~5%), MR relies on the assumption of linear dose-
response effects, which may not be suitable. It is also
possible that subtle population stratification of adiponectin levels across Europe biased our results. Yet,
no differences in adiponectin level across European
populations in ADIPOGen, a consortium measuring
adiponectin levels in 26 European or Europeandescent cohorts, were detected. Finally, as with any
null finding, the width of the 95% CIs gives a sense of
what effect sizes can be excluded, given the large
(two-SD) genetic increase in adiponectin levels.
In conclusion, using data from the largest genetic consortia for adiponectin and MS, we find that lifelong
exposure to a substantially (two-SD) genetically-elevated adiponectin level has no clinically-relevant
effects on MS susceptibility in individuals of
European descent. Adiponectin is therefore not likely
to be an ideal candidate target for MS prevention;
however, its therapeutic potential for MS following
disease onset remains to be determined. Additional
studies will be necessary to ascertain which biological
factors drive the causal association between body
weight and MS.
Acknowledgements
The authors wish to thank the ADIPOGen Consortium,
IMSGC, and the IMSGC/WTCCC2 study for access
to their data. ADIPOGen data are publicly available
and can be accessed at http://www.mcgill.ca/genepi/
journals.sagepub.com/home/msj
http://www.mcgill.ca/genepi/adipogen-consortium https://journals.sagepub.com/home/msj
Table of Contents for the Digital Edition of Multiple Sclerosis Journal - October 2017
