Alzheimer’s and inflammatory proteins: genetic clues without robust confirmation

A genetic investigation of 91 proteins

The researchers studied associations between 91 circulating inflammatory proteins and Alzheimer’s disease. Their method, two-sample Mendelian randomization, uses genetic variants associated with protein levels to explore their possible contribution to disease risk. This is not a trial in which a treatment alters these proteins.

The discovery stage relied on genetic data on proteins from 14,824 participants. The validation analysis used data from 54,219 UK Biobank participants. The Alzheimer’s dataset included 85,934 cases and 401,577 controls. The supplied abstract does not specify a follow-up duration.

Signals that do not pass the statistical filter

In the initial analysis, four proteins showed what is known as a nominal association with Alzheimer’s: AXIN1, NGF, CXCL11 and TNFRSF9. But none of these associations withstood correction for multiple comparisons. This distinction matters: when many proteins are tested, some results may appear significant by chance.

In the validation stage, only TNFRSF9 reached the nominal threshold, but it did not withstand this correction either. For AXIN1, the direction of the association was even reversed compared with the initial analysis.

Pooling the two stages yielded suggestive estimates for CXCL11, TNFRSF9 and NGF. However, the abstract specifies that these combined results should not be considered independent confirmation.

Research hypotheses, not an intervention to adopt

These results do not demonstrate that altering any of these proteins would reduce the risk of Alzheimer’s. A genetic association can inform a causal hypothesis, but it does not directly measure the effect of a drug or supplement.

The researchers also searched a database for interactions between the genes involved and various drugs or compounds. Identifying these interactions does not amount to testing efficacy in patients.

The useful conclusion therefore remains limited: some proteins warrant further investigation, but this study does not establish a robust link after statistical correction. It supports neither recommending an anti-inflammatory intervention against Alzheimer’s nor ruling out a role for inflammation in the disease.

What we can take away

This study generates hypotheses about inflammatory proteins and Alzheimer’s without confirming a target that could be acted on to prevent or treat the disease.