Epstein-Barr virus (EBV) is associated with various benign and malignant diseases. Chronic active EBV disease (CAEBV) is a rare EBV-associated disease characterized by persistent EBV infection in natural killer (NK) and/or T cells, which is mainly prevalent in East Asia. While patients with CAEBV initially present with persistent or recurrent symptoms similar to ones present in infectious mononucleosis (IM), they often develop peripheral NK/T-cell lymphoma or multiple organ failure in the long run, and prognosis is dismal.

Numerous aspects of the molecular pathogenesis of CAEBV are unresolved; it is unclear whether this disease is a clonal disorder. Although several case reports indicate the presence of primary immunodeficiency (PID) in patients with persistent EBV disease, the prevalence of PID or other inherited diseases in patients with CAEBV remains unclear. To reveal the genetic background of CAEBV, we performed whole-exome and targeted sequencing.

Patients and Methods

We studied 83 patients with CAEBV. The diagnosis of CAEBV was based on the fulfillment of the following criteria: (1) persistent or recurrent symptoms similar to those present in IM, (2) increased quantities of EBV viral load in either affected tissues or peripheral blood, and (3) exclusion of other possible diagnoses. The Institutional Review Board of Nagoya University Graduate School of Medicine approved this study.

We performed whole-exome sequencing (WES) of T-, B-, and NK-cell subsets. Subsequently, we performed targeted gene sequencing of 605 genes associated with PID, inherited blood diseases, and hematological malignancies.


Regarding germline mutations, we identified two patients who carried nonsense mutations in EVC2, which led to the diagnosis of Weyers acrofacial dysostosis (WAD). We also identified a patient who carried POLH mutations, which led to the diagnosis of the variant type of xeroderma pigmentosum, and a patient who carried an APC mutation, which led to the diagnosis of familial adenomatous polyposis. Mutations leading to the diagnosis of PID were not identified in any patient.

WES led to the identification of at least one somatic mutation in 58% of patients. We identified DDX3X (n= 14; 16%), KMT2D (n= 4; 4.8%), BCOR / BCORL1 (n= 3; 3.6%), KDM6A (n= 3; 3.6%), and TET2 (n= 2; 2.4%) as undergoing recurrent mutations.

We analyzed multiple EBV-infected cell subsets within a patient. One patient (UPN125) developed CAEBV in which both T and NK cells were positive for EBV DNA. The patient's CD4+ and CD56+ cells carried 21 and 27 mutations, respectively, of which 8 were shared including a DDX3X p.Arg534Cys driver mutation. Similar to UPN125, CD3+ and CD56+ cells of UPN404 carried a shared DDX3X p.Thr198Pro mutation and private mutations. CD19+ cells of UPN404 also carried the identical DDX3X mutation.

In the survival analysis, patients carrying DDX3X mutations at diagnosis demonstrated a significantly worse OS compared with others (5-year OS: 40.6% vs. 78.6%; P= 0.007). Patients carrying KMT2D mutations also demonstrated poor prognosis (5-year OS: 50.0% vs. 73.9%; P= 0.02).


In the present study, we found that CAEBV was a malignant disease carrying somatic driver mutations, the spectrum of which showed similarity with that of extranodal NK/T-cell lymphoma (ENKTL) and Burkitt lymphoma. DDX3X mutations and other shared mutations, which were present in multiple cell subsets in a patient, strongly suggest the presence of common ancestry, possibly lymphoid progenitor cells, which acquired driver mutations.

We identified the presence of somatic DDX3X mutations as a marker of poor prognosis. A prospective study can be designed using genetic testing for DDX3X mutations and subsequent interventions, possibly chemotherapies used for ENKTL and early hematopoietic stem cell transplantation, to improve patient outcome.

Furthermore, we identified several germline mutations. WAD caused by EVC / EVC2 mutations is characterized by the malformation of mesoderm and ectoderm caused by decreased hedgehog signaling mediated via EVC/EVC2 . Together with POLH and APC mutations, EVC / EVC2 mutations may be a genetic predisposition to CAEBV.

In conclusion, we revealed the genetic background of CAEBV, which will lead to further clarification of its molecular pathogenesis and the development of therapeutic strategy. Clinical trials utilizing genetic diagnosis and stratified therapy are warranted.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.