Monoallelic Pathogenic Variants in Hemophagocytic Lymphohistiocytosis Genes Are Uncommon and Not Associated with Hematopoietic Cell Transplantation Outcomes in Severe Aplastic Anemia

Introduction: Severe aplastic anemia (SAA) is primarily an immune-mediated disease, but a subset of patients have a germline genetic predisposition. Germline genetic testing for patients with SAA is recommended to guide patient treatment, including adjustment for hematopoietic cell transplantation (HCT) modalities. Hemophagocytic lymphohistiocytosis (HLH) is a life threatening hyperinflammatory condition often associated with cytopenias. HLH etiology is genetically variable with X-linked or autosomal recessive (XLR or AR) modes of inheritance. The role of monoallelic pathogenic/likely pathogenic (P/LP) variants in recessive HLH-associated genes in SAA etiology or outcomes is unclear. In this study, we determined the frequency of monoallelic HLH-associated gene variants in a large cohort of patients with acquired SAA and evaluated their association with HCT outcomes.

Methods: The Transplant Outcomes in Aplastic Anemia (TOAA) project is a collaboration between the NCI and the Center for International Blood and Marrow Transplant Research and consists of genomic and clinical data from 824 patients with SAA who received HCT between 1989-2015. We excluded 140 patients with inherited bone marrow failure syndrome (92 with a clinical diagnosis and 48 genetically identified; McReynolds, Blood, 2022). We used exome sequencing data from the remaining 684 non-inherited SAA patients and used ACMG/AMP criteria to curate variants in 14 HLH-associated genes (AR: PRF1, UNC13D, STX11, STXBP2, ITK, CD27, LYST, RAB27A, AP3B1, BLOC1S6, SLC7A7; and XLR: XIAP, SH2D1A, MAGT1). Variants were deemed deleterious variants of uncertain significance (del-VUS) if there were damaging predictions in >3/5 in-silico meta-predictors (BayesDel, REVEL, CADD, MetaSVM and EIGEN). Follow-up started at the day of HCT and ended at outcome of interest, death, loss to follow-up, or end of study period (August 2017). We used the Kaplan-Meier estimator to calculate the probability of overall survival (OS) after HCT, using a log-rank test for statistical comparisons. For other HCT outcomes, we used cumulative incidence calculator accounting for competing risk of death.

Results: There were 57 HLH variants in 49 of the 684 (7.2%) patients: 17 variants present in 19 patients (2.8%) were pathogenic or likely pathogenic (P/LP) and 40 were del-VUS. Eight of 17 P/LP variants were loss of function (3 frameshift, 3 splice site loss, and 2 nonsense). Among the 19 patients with P/LP variants, 16 (84%) had monoallelic variants in genes with AR inheritance, and PRF1 was the most frequently affected (in 8/19 patients). The median age at HCT was comparable in patients with (median=20 years; range=5-62) and without (median=20 years; range=0.2-77) P/LP HLH variants. We did not observe statistically significant differences in other demographics or transplant-related factors in the two groups, except for a higher frequency of male in patients with P/LP variants (79% vs. 55%, in those with and without HLH variants, respectively; p=0.04). All patients with P/LP HLH variants achieved neutrophil engraftment by day 30 compared with 92% of patients without HLH variants. The cumulative incidence of neutrophil engraftment at day 20 was 89%, 95% CI=70-99% for patients with HLH variants and 64%, 95% CI=60-67% in those without variants (p=0.01) (Figure 1A). No statistically significant associations between HLH variants and other key HCT outcomes were noted (p>0.05). The 5-year OS probability in all patients with P/LP variants was 89% (95% CI=72-99) compared to 66% (95% CI=63-70) in those without variants (p=0.06) (Figure 1B). When restricting the analysis to patients receiving reduced intensity/non-myeloablative regimen, the 5-year OS was 92% (95% CI=72-100%) in those with P/LP HLH variants versus 75% (95% CI=70-79%) in those without variants.

Conclusions: We found that 7.2% of patients undergoing HCT for acquired SAA carried potentially deleterious variants in HLH genes, with 2.8% carrying a P/LP variant. HCT outcomes in patients with monoallelic P/LP variants was similar to those without P/LP HLH variants. Our data suggest that identification of monoallelic P/LP variants in HLH genes does not influence HCT outcomes, and no special treatment consideration, other than that used for SAA, is warranted.