Specific Donor Human Leukocyte Antigen (HLA) Allotypes and CMV IgG Serology Status Predict the Risk of Cytomegalovirus-Related Disease in Acute Myeloid Leukemia Patients Who Received Allogeneic Hematopoietic Stem Cell Transplantation

Background

Cytomegalovirus (CMV) establishes lifelong latency after primary infection under the control of the immune system because of the numerous virus evasion strategies that interfere with the host immune response at many levels. Human leukocyte antigen (HLA)-restricted cytotoxic T lymphocytes (CTLs) are involved in the early immune response and are an important defense mechanism in CMV infections, reactivation, and related diseases. Furthermore, an assessment of the clonal diversity of T cell responses against CMV infection provides important insight into the molecular basis of T cell immunodominance. In this single-center study, we tried to demonstrate a specific correlation between the donor HLA genotype and cumulative incidence of CMV reactivation and disease.

Patients and methods

We retrospectively analyzed 613 donors and recipients diagnosed with acute myeloid leukemia (AML) who received allogeneic hematopoietic stem cell transplantation (allo-HSCT) from matched siblings (n=260), matched unrelated donors (n=167), or haploidentical family donors (n=186) from 2012 to 2017. The CMV-related disease was diagnosed with aggressive procedures in suspicious tissues such as the eyes, gastrointestinal tract, or respiratory tract. The cumulative incidence of overall CMV-related diseases was 12.3% (n=71; range, 9.8 - 15.2), and in each matched sibling, matched unrelated, and haploidentical family donor allo-HSCT group were 6.1% (range, 3.6-9.6), 14.4% (9.2-20.7), and 19.4% (14.0-25.5), respectively. Except for seven patients, all 64 patients developed CMV disease in the CMV reactivation state. We determined the genotypes of the HLA-A, B, C, and DRB1 alleles in 613 donors and recipients by sequencing method and further selected 560 (91.4%) CMV IgG seropositive donors to identify the genetic influence of donor HLA according to CMV infection.

Results

We first analyzed the relationship between entire donor HLA allotypes and the cumulative incidence of CMV-related disease, then subdivided the donor groups by CMV IgG seropositivity. In the CMV IgG seropositive donor group, we conducted subgroup analysis to identify any difference in CMV-related disease incidence according to types of allo-HSCT. As a result, an entire donor CMV serostatus, three genotype alleles, HLA A*3004 (OR 2.8; p-value 0.044), B*5101 (OR 2.3; p-value 0.003), and DRB1*0901 (OR 2.3; p-value 0.004), demonstrated a statistically significant odds ratio (OR) value with the proper number of patients. However, in the donor CMV IgG seropositive subgroup, two allotypes, HLA B*5101 (OR 2.0; p-value 0.003) and DRB1*0901 (OR 2.7; p-value 0.002), remained. Interestingly, the HLA DRB1*0901 allele showed a concrete association (OR 6.0; p-value <0.001, and p(c)-value 0.002) between CMV IgG seropositive donor HLA and the CMV-related disease incidence of the recipient, especially in the haploidentical allo-HSCT setting. The HLA-B*5101 allele showed a statistically significant association in the IgG seropositive donor subgroup with the matched unrelated allo-HSCT recipient and in the IgG seronegative donor subgroup.

HLA-DRB1*1302 showed a promising value as the protective marker (OR 0.2; p-value 0.041) only in the IgG seropositive donor subgroup with the matched unrelated allo-HSCT recipient category. HLA-A*2402 (OR 3.6; p-value 0.048) was only significant in the IgG seropositive donor subgroup with the matched sibling and haploidentical allo-HSCT recipient category. HLA-DR*1501 (OR 2.6; p-value 0.039) was only significant in the IgG seropositive donor subgroup with the matched sibling allo-HSCT recipient category.

Conclusion

This study demonstrated that certain donor alleles, donor CMV IgG serostatus, and types of allo-HSCT, especially the seropositive donor HLA-DR*0901 allele in the haploidentical allo-HSCT setting, significantly correlated with high CMV-related disease incidence and might be considered risk markers for suitable donor selection. Additionally, the specific donor HLA allele showed either protective or aggravated CMV-related disease incidence in a different allo-HSCT setting. For patients receiving various types of allo-HSCT, a strategic approach to donor selection with careful consideration of donor HLA allotype is important and intensive CMV reactivation monitoring may be required, especially in acute GVHD under active steroid pulse treatment.