Reply to Sawalha: Evidence for IRAK1 is not evidence against MECP2

Sawalha's letter (1) questions our interpretation of the results reported in our recent article (2). Sawalha states “because of high LD in this locus, both IRAK1 and MECP2 should be considered candidate associations,” echoing what we have explicitly discussed in our article. Indeed, we acknowledge and reference Sawalha's paper from 2008 (3) suggesting that MECP2 is associated with systemic lupus erythematosus (SLE) and clearly state that “given the physical proximity of IRAK1 and MECP2 on Xq28, it is plausible that they are in linkage disequilibrium, and the 2 independent studies possibly describe the same genetic association.” In the same vein, we would also urge Sawalha to consider the possibility that the MECP2 associations he reports could very well be due to genetic alterations in IRAK1, a reference and a discussion that is absent in his papers (3, 4), despite our description of the association to IRAK1 as early as 2007 (5).

Sawalha's letter implies that the IRAK1 association with SLE is secondary to the association with MECP2 and states that “indeed, IRAK1 expression is regulated by MECP2.” He bases this inference on 2 papers that show IRAK1 as one of the many overexpressed genes in the brain of MECP2-deficient mice [refs. 3 and 4 in Sawalha's letter (1)]. Nonetheless, this observation is contradicted by his recent report (4) showing no difference in IRAK1 expression between SLE subjects homozygous for the lupus-associated MECP2 risk haplotype as compared to lupus patients homozygous for the MECP2-protective haplotype, based on gene-expression profiling (4). Intriguingly, Sawalha's letter fails to mention this finding, which is highly relevant to his argument. In any event, whether IRAK1 expression is regulated by MECP2 is irrelevant to the question of whether IRAK1 is a culprit gene. Just because IRAK1 can be impacted by gene methylation (possibly due to MECP2 or other genes) does not preclude it from being a candidate gene in SLE.

We were cautious to avoid extrapolating from the IRAK1^−/− mice to humans. Nowhere in our article do we state that “the abrogation of lupus manifestation in an IRAK1-deficient lupus mouse model localizes the lupus susceptibility gene in the Xq28 locus to IRAK1,” as attributed to us by Sawalha.

However, we have to consider the functional properties of these 2 genes. IRAK1 has been convincingly documented to have an integral role in inflammation, and inflammation drives end-organ pathology in SLE. In contrast, how MECP2 might be functionally linked to SLE remains a black box. Ample evidence suggests that MECP2 mutations cause the neurodevelopmental disorder, Rett syndrome (RTT). These mutations, including the premature stop-codon-associated MECP2-null genotype, induce a host of neurobehavioral abnormalities characteristic of RTT (6). Furthermore, MECP2 gene duplication also elicits features overlapping those of RTT (7). Surprisingly, neither gain nor loss of MECP2 function in humans has been associated with any features of autoimmunity. Thus, from the functional perspective, IRAK1 emerges much more attractive as a potential pathogenic player, compared to MECP2.