Further insights into the causal link between immune cell phenotypes and autoimmune liver diseases

We thank Dr. Ren and colleagues for their interest in our article and for raising several important methodological considerations. In our study, we investigated the potential causal relationships between diverse immune cell phenotypes and autoimmune liver diseases (AILDs) using a bidirectional two-sample Mendelian randomization (MR) (1).

First, regarding population stratification, we acknowledge the difference between the Sardinian founder population used for immune cell phenotyping and the predominantly European-ancestry AILDs cohorts. However, both samples are of European descent. Furthermore, all instruments were harmonized using the European 1000 Genomes and conditional F-statistics were consistently high. Published benchmark studies have shown that such harmonization and strong instruments largely mitigate bias arising from modest linkage disequilibrium (LD) differences (2,3).

Second, horizontal pleiotropy and colocalization. Horizontal pleiotropy is an important concern when analyzing highly correlated immune traits. We applied multiple pleiotropy-robust methods [MR-Egger, weighted median, MR Pleiotropy Residual Sum and Outlier (MR-PRESSO)], which showed no evidence of systematic bias. At the time of analysis and submission, no publicly available cell-type-specific expression quantitative trait loci (eQTL) dataset with sufficient coverage of immune phenotypes existed to enable meaningful colocalization analysis.

Third, instrument strength and statistical power. Certain rare or highly specific immune subsets are associated with only a limited number of genome-wide significant single-nucleotide polymorphisms (SNPs) in the current largest flow-cytometry genome-wide association study (GWAS). This is a recognized feature of existing immunogenetic resources rather than a study-specific limitation. We followed current consensus guidelines (F>10 together with multiple sensitivity analyses) and observed consistent effect directions across methods (4). Most importantly, our strongest conclusions [e.g., regarding human leukocyte antigen-DR isotype (HLA-DR)⁺ natural killer (NK) absolute count (AC) and HLA-DR NK %NK, and other related cell types] were based on multiple SNPs and passed multiple-testing correction. Thus, although some analyses are underpowered, the key signals we highlight meet conventional thresholds.

Finally, tissue specificity and biological context. We agree that cytokines, autoantibodies, epigenetic modifications, gut microbiome, and environmental factors are important in AILDs; however, MR intentionally avoids non-genetic covariates to prevent confounding, which is a core strength of the method rather than a limitation. We fully concur that peripheral blood measurements do not fully reflect the hepatic immune microenvironment. Unfortunately, at the time of our analysis, no genome-wide genetic instruments derived from human liver-resident immune cells were publicly available. Until liver-specific immunogenetic datasets emerge, peripheral blood remains the only feasible source for large-scale studies of this kind. We look forward to future availability of such resources that will enable more tissue-specific analyses.

In summary, we appreciate the thoughtful critique and welcome future improvements as larger, multi-ancestry, and tissue-specific datasets become available. However, the concordance of our results with known immunopathogenic concepts (e.g., roles of HLA-DR alleles and CD27 B cells) and the fact that other MR analyses in immune contexts have used similar methods (5-10), give us confidence that our analyses were conducted according to current best practice standards and provide valid hypothesis-generating evidence for the role of specific immune subsets in AILDs susceptibility.

We are grateful to the commentators for highlighting these methodological points shared by many contemporary immunogenetic MR studies.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.