Unraveling family ties in Hodgkin lymphoma

In this issue of Blood, Flerlage et al¹ expand our understanding of genetic factors which contribute to a predisposition to the development of Hodgkin lymphoma (HL). Comprehensive genetic analysis with whole genome sequencing was performed on individuals with and without HL from families in which at least 2 or more first-degree relatives had experienced the disease. Risk variants for HL were identified and the description of these variants by Flerlage et al provides the rationale and a starting point for further interrogation of these variants in other cohorts of patients with HL and their families.

In recent years, many advances have been made in understanding genetic predisposition to some forms of myeloid malignancy, particularly myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML). Although there remain aspects of familial MDS and AML that are understood incompletely, sufficient evidence has been obtained to permit inclusion for the first time, in the 2017 World Health Organization Classification of Tumors of Hematopoietic and Lymphoid Tissues,² a category entitled “Myeloid neoplasms with germline predisposition.” Consideration of whether one of these monogenic predisposition conditions exists is now deemed to be an essential part of new acute leukemia evaluation, regardless of the age of the patient. Since the 2017 iteration of the World Health Organization Classification, more evidence in the setting of predisposition to lymphoid malignancy has been obtained, leading to an alteration in the International Consensus Classification category title from “myeloid neoplasm predisposition” to “hematologic neoplasms with germline predisposition.”³ However, these recent developments have largely been in the area of predisposition to lymphoblastic leukemia, and the current classifications do not specifically describe a germline predisposition to HL.

Although we remain unable to include a HL predisposition within contemporary hematologic malignancy predisposition classifications, clustering of HL within some families has long been recognized.⁴ In 1959, Razis et al⁵ commented that “we do not know whether the reported cases of familial Hodgkin's disease in the medical literature signify medical curiosities, or whether they carry weighty environmental and genetic implications.” Evidence obtained since this time indicates that the latter was correct, namely, that multiple extrinsic and intrinsic factors are likely to contribute to HL development (see figure). There is an increased risk of approximately 3-fold described in first-degree relatives of patients with HL compared with the general population risk and siblings (as opposed to parents or offspring) experience a higher risk.⁶ Although some rare monogenic causes have been suggested, the majority of HL pedigrees have remained largely unsolved, with current thinking that the observation of HL clustering in some families is due to a combination of polygenic and environmental factors.⁷ A well-described association between specific human leukocyte antigen alleles and risk of both Epstein-Barr virus (EBV)-positive and EBV-negative HL development provides evidence of a genetic susceptibility to this disease.⁸ 

Flerlage et al analyzed individuals from 36 HL pedigrees. To increase the probability of analyzing a family with an underlying germline predisposition, the pedigree selection criteria included that at least 1 patient from the pedigree had onset of HL before the age of 22 years. The EBV status of patients and tumor tissue in this cohort was unknown. This factor is a limitation acknowledged by the authors and was partly mitigated by the deliberate selection of kindred in which a young person had developed HL (the group with lowest rates of EBV-associated HL). The authors described 44 HL risk variants among 28 of the HL pedigrees analyzed, including both coding and noncoding risk variants. Variants were considered recurrent if present in more than 1 pedigree. However, of the 4 variants deemed recurrent, they were observed in only 2 or 3 families. Of these 4 recurrent risk variants, 3 were noncoding, including one variant in the 5' untranslated region of KLHDC8B and 2 were intronic variants (in GATA3 and PAX5), highlighting the importance of undertaking a comprehensive analysis with whole genome sequencing to permit examination of the DNA outside of the coding regions. GATA3 and PAX5 were considered novel predisposing loci, whereas KDR and KLHDC8B had been described previously. Most remaining risk variants identified were unique to a single pedigree, and so constitute private, rare variants.

Given the spectrum and rarity of postulated HL risk variants and the probable polygenic manner in which most will confer a HL predisposition, we are currently unable to neatly categorize a genetic predisposition to HL in the same way we can for several forms of hereditary MDS and AML. This factor presents challenges for how we counsel, test, and manage patients and families with a suspected familial HL predisposition. Flerlage et al have suggested 44 HL risk variants that may now be further evaluated in functional and translational research studies. The publication of these findings also permits the interrogation of other HL cohorts for the presence of these postulated HL risk variants. There is potential for this work alongside other datasets and future research efforts, to permit design of a multifaceted risk score for HL predisposition or development. Flerlage et al state that the genomic landscape of familial HL remains incompletely characterized. This publication represents an important step toward understanding the familial clustering that may occur with this disease.

Conflict-of-interest disclosure: The author declares no competing financial interests.