Referee 3's review (Patrizia Zavattari):

• What are the main claims of the paper?

The authors performed an interesting study regarding the association of an X-chromosome gene region with SLE (Systemic lupus erythematosus). It was a pleasure to review this interesting and well written paper. Using a candidate gene association approach they investigate this association by genotyping 21 SNPs located within and around the gene MECP2 (Methyl-CpG-binding protein 2), which maps on chromosome Xq28. They initially performed the genotyping in 628 Korean female patients and 736 healthy female Korean controls, finding evidence of association (10-6). The authors replicated the association results in a second population (1080 patients and 1080 controls of European descent), raising up to a combined P value of around 10-8. The main disease associated alleles are much more frequent in Korean than in European-derived controls and the authors suggest that this could help to explain the higher frequency of SLE in Asians compared to Europeans. Interestingly this gene, well characterised, especially because of its involvement in Rett syndrome, seems to be involved in the regulation of methylation sensitive genes, such as CD70, known to be overexpressed in T cells from patients with SLE. This was also confirmed by functional experiments done by the authors and described in this paper. As the authors suggest, this genetic association and the epigenetic correlations could provide an explanation for the predominance of SLE in females and may also indicate an interesting pathway in the etiopathogenesis of other autoimmune diseases.

• Are the claims properly placed in the context of the previous literature?

Yes. There was a strong rationale to study the association of the very promising candidate gene MECP2 with SLE; in fact, as the authors reported, “the DNA methylation sensitive genes are overexpressed in SLE and MECP2 is critical in the transcriptional suppression of methylation sensitive genes”. Moreover, MECP2 is a gene encoding an HDAC (histone deacetylases)-binding protein. As reported in the literature, studies in SLE suggest a possible role for HDACs in skewed gene expression and disease pathogenesis. HDAC inhibitors also have widespread modulatory effects on gene expression within the immune system and have been used successfully in the lupus and rheumatoid arthritis autoimmune disease models [Nilamadhab M et al, J. Clin. Invest. 2003, 111:539–552; Gray SG and Dangond F, Epigenetics. 2006 Apr-Jun;1(2):67-75].

• Do the experimental data support the claims? If not, what other evidence is required?
The authors provide good experimental evidence, with a quite powerful sample set and they obtained statistically significant evidence of association.
Nonetheless I have some comments to address to them:

1) In order to guarantee a certain quality threshold of the genotype data, the authors made the choice of only showing results for SNPs with MAF > 5%. Did the authors genotype any marker with a second technology (i.e. Taqman) in order to confirm that the Illumina genotyping was actually well assigned? In my experience, it is always reassuring to have an independent control of the method; for example by typing one marker with another method and checking for consistency. It is, naturally, very unlikely to find so many SNPs associated with a disease in the same genomic region if the data are unreliable; furthermore different SNP alleles show a P value around 10-6 where a nominal P value of 10-5 evaluating for strong functional candidates across the genome, provide a low false positive rate (Dahlman I et al, Nat. Genet. 30: 149-150). But also given that, because of the importance of the results, I think it would be worth verifying at least one SNP with a second genotyping method.
2) In Tables 2, 4 and 5 the total number of fully typed individuals should be reported, to make it easier to reproduce calculations.
3) In Table 3 it would be useful to read not only frequencies but numbers of observations as well.
4) In the text, when the authors refer to Table 4, it would be helpful to point out that the frequencies given there correspond to frequencies of the homozygous genotypes.
5) In Table 5 the authors reported results of association analysis in the European-derived population. They clearly replicate the association evidence but could potentially improve the P values by applying a P one-tailed instead of two-tailed. In fact, they are no longer testing the probability that these alleles are associated with SLE or not, but are verifying whether the previous association results are reproducible in a different population. Unless they where not sure about which alleles could be associated for the different markers, i.e. because of the very different allele frequencies in the two populations.
6) In the methods where the authors describe the participants of the study, the participants do not come from the same family: is that nuclear family, or extended family. Please specify.
7) There is no evidence for any of the analysed SNPs located in MECP2 gene to be the etiologic one. Did the authors test any coding variants or SNPs potentially implicated in alternative splicing?
8) Because of the presence of another strong candidate gene in the same region, IRAK1, did the authors genotype any other SNPs located in this gene, in order to demonstrate that the main association evidence actually maps in MECP2 gene?

• In what further directions would it be useful to take the current research?

It will be very interesting to see the results of the next steps: resequencing of the coding and regulatory regions and intron-exon boundaries, in order to identify the etiological variants; functional studies demonstrating the effect of these variants in the etiopathogenesis of the disease. The authors have presented some interesting data concerning the role of MECP2 truncations in mice and demonstrated association of some polymorphisms in the orthologue in humans with SLE. Now the task will be to determine if you can find a mutation in MECP2 that has a functional correlation, i.e, aberrant methylation.

• Who would find this paper of interest? And why?

Apart from people working in the same field as the authors, groups working in other autoimmune diseases can be interested to test the association between those diseases and SNPs located in MECP2 gene.

• Is the manuscript written clearly enough that it is understandable to non-specialists? If not, how could it be improved?
This paper is very well written, allowing anybody to understand the rationale of the study, the performed analysis and the obtained results.

• Have the authors provided adequate proof for their claims without overselling them?

From my point of view the results of this work are very significant and deserve to be published.

**********
N.B. These are the comments made by the referee when reviewing an earlier version of this paper. Prior to publication the manuscript has been revised in light of these comments and to address other editorial requirements.