How does HLA-DQ typing inform celiac disease risk in IgA-deficient patients?

Celiac disease (CeD) is a chronic autoimmune disorder triggered by the ingestion of gluten in individuals with specific genetic risk factors, most notably human leukocyte antigens (HLA)-DQ2 and DQ8. The diagnostic evaluation of suspected CeD often involves serological testing for tissue transglutaminase (tTG) antibodies, primarily of the IgA class, and a confirmatory small intestinal biopsy [1]. In most large cohort studies, tTG-IgA shows the highest sensitivity and specificity of the available serology tests [2]. However, diagnosing CeD in IgA-deficient patients, who have a higher prevalence of this disease, can be challenging, as IgA-based markers may be unreliable due to the patient’s low or undetectable IgA concentrations. Instead, IgG-based serological tests for tTG-IgG and deamidated gliadin peptide (DGP) IgG are required to determine whether a confirmatory biopsy should be performed. Although the presence of HLA-DQ2 and DQ8 is necessary for disease development, their associated risk varies across different alleles. Based on serotype classification, HLA-DQ2 includes two heterodimers, HLA-DQ2.5 and HLA-DQ2.2. Previous studies characterized the genetic risk gradients of various HLA-DQ2 and DQ8 heterodimers and their association with positive for tTG-IgA in patients with sufficient total IgA concentrations [3]. However, the impact of both gene variants in IgA deficient patients remains uncertain. 

In our recent study, we analyzed data from 3,714 patients with insufficient total IgA who were being evaluated for CeD. These individuals had HLA-DQ genotyping, tTG-IgG and DGP-IgG testing performed at Mayo Clinic Laboratories. Among these patients, 56% expressed at least one HLA-DQ2 or DQ8 heterodimer. Homozygosity for HLA-DQ2.5 was associated with a significantly increased risk of tTG-IgG positivity (OR = 10.19, 95% CI: 6.22–16.69), demonstrating a strong genetic link to CeD. However, homozygosity for HLA-DQ2.2 or DQ8 alone did not significantly increase risk (OR < 1), suggesting limited correlation to tTG-IgG positivity. Although heterozygous HLA-DQ2.5 carriers displayed moderate risk (OR = 3.27, 95% CI: 2.28–4.68), the risk increased substantially in the presence of HLA-DQ2.2 (OR = 7.19, 95% CI: 4.38–11.83) or HLA-DQ8 (OR = 5.89, 95% CI: 3.5–10.21). DGP-IgG testing results further supported the higher odds ratio in those HLA-DQ genotypes. These observations emphasize that HLA-DQ2.5, particularly in homozygous or combined genotypes, confers the highest genetic risk for serology positivity. In contrast, patients carrying only HLA-DQ2.2 or HLA-DQ8 showed minimal risk associated with CeD, regardless of zygosity. 

Furthermore, tTG-IgG demonstrated a higher positivity rate across all genotypes compared to DGP-IgG. However, this is mitigated by the observation that, in patients without any HLA-DQ2 or DQ8 alleles, the positivity rate of DGP-IgG was 0.97% compared to 3.42% for tTG-IgG. This suggests that DGP-IgG may offer higher specificity, making it a valuable marker for minimizing false positives serology results. 

Interestingly, among patients with IgA deficiency (total IgA < 7 mg/dL), 71% carried at least one HLA-DQ2 or DQ8 allele. This finding raises the question as to whether this reflects a higher proportion of patients with CeD in patients with IgA deficiency or whether there is a direct genetic link between HLA-DQ2/DQ8, CeD and IgA deficiency. Regardless, the high prevalence of HLA-DQ2 or DQ8 among IgA-deficient patients is consistent with an increased risk of developing CeD. 

Overall, the findings from our study characterize the genetic risk gradient of HLA-DQ alleles for IgA-deficient patients. In comparison to individuals with sufficient IgA, relying on serological markers alone in patients with IgA deficiency may lead to ambiguous or inconclusive results, especially given the difference in specificity observed in our study. The questions now are: Could HLA typing be integrated with IgG-based serological markers for diagnosing CeD in IgA-deficient patients? How can we incorporate these findings into routine clinical practice to optimize patient care?

References

1. A. Rubio-Tapia, I.D. Hill, C. Semrad, C.P. Kelly, K.B. Greer, B.N. Limketkai, B. Lebwohl, American College of Gastroenterology guidelines update: diagnosis and management of celiac disease, Official journal of the American College of Gastroenterology| ACG 118(1) (2023) 59-76. 
2. A. Rubio-Tapia, I.D. Hill, C.P. Kelly, A.H. Calderwood, J.A. Murray, ACG clinical guidelines: diagnosis and management of celiac disease, Official journal of the American College of Gastroenterology| ACG 108(5) (2013) 656-676. 
3. J.R. Mills, M.R. Snyder, J.A. Murray, M.J. Gandhi, Celiac disease risk stratification based on HLA-DQ heterodimer (HLA-DQA1~ DQB1) typing in a large cohort of adults with suspected celiac disease, Human Immunology 81(2-3) (2020) 59-64.