Familial Clustering of Hematological Malignancies: Harbingers of Wider Germline Cancer Susceptibility

Introduction. The Australian Familial Haematological Cancer Study (AFHCS) was founded in 2004 when only one familial haematological malignancy (FHM) gene was known, RUNX1 causing thrombocytopenia and predominantly an MDS/AML HM phenotype. Today, there are over a dozen known FHM genes, mostly characterized as predisposing to myeloid malignancies. Within our AFHCS cohort of over 200 families, this is also true with a significant proportion of families with predisposition to myeloid malignancies solved with our identification of germline mutations in genes such as GATA2, RUNX1,CEBPA and DDX41. In contrast the majority of unsolved families have a predominance of lymphoid malignancies, most commonly non-Hodgkin lymphoma and chronic lymphocytic leukaemia. In several of these families we have found known pathogenic mutations in genes involved in regulation of the homologous recombination (HR) DNA damage repair (DDR) pathway (e.g. BRCA2 Y1710fs* and PALB2 K142*) that are previously characterised with regard to their known solid tumour predisposition. In the AFHCS families with these mutations we observe a mixture of hematological tumors of mostly lymphoid origin as well as non-hematological tumours. This led us to hypothesise that there is a subset of families for which the predisposition to malignancy includes a spectrum of different malignancies, i.e. pan-cancer predisposition. Further, we hypothesise that these are associated with mutations in genes with roles in HR DNA repair. In light of this we undertook a review of the phenotypic and genetic characteristics of families in our AFHCS cohort.

Methods. Malignancy phenotype data, with regard to the different tumor types observed in each AFHCS family was annotated and aggregated. Covariance matrix analysis was then used to compare the number of times different tumor types were observed occurring together in a family pedigree. Genetic data: we curated whole exome sequence data for damaging mutations in genes with a known role in HR DNA damage repair.

Results. A phenotype review of our cohort found 52% of assessable FHM families also had multiple types of solid tumour and/or individuals with diagnosis of multiple malignancies, indicative of a 'pan-cancer' phenotype. Covariance analysis showed that Hodgkin and non-Hodgkin lymphoma most often occurred together and also co-occurred with non-HM tumors breast, melanoma, colorectal, prostate and ovarian subtypes, implying a possible shared genetic etiology (Figure 1A). Assessment of AFHCS families for mutations in DDR genes including DNA repair (e.g. FANC/BRCA pathway genes) and chromosome cohesion (RAD21 and related) identified 18 variants segregating with malignancy including 5 frameshift or stopgain mutations. These high impact mutations included genes previously characterised in the context of autosomal recessive predisposition to hematological malignancy such as PALB2/BRCA2 family (Fanconi Anemia, see Figure 1B), NBN1 (Nijmegen Breakage Syndrome), DDX11 (Warsaw Syndrome) and others. A further 18 variants segregated in families where pan-cancer status is still under investigation.

Conclusions. Phenotypic and genetic analysis of unsolved HM predisposed families have identified that familial clustering of lymphoma in actuality identifies families with pan-cancer germline genetic susceptibility, characterised by mutations in DNA damage response genes. Collectively these mutations may have a significant impact on our understanding of germline cancer predisposition, a finding supported by recent genetic population studies on sporadic cohorts, including pediatric cancer cohorts, opening the door for the application of rational therapies as evidenced by the BRCA/PARP inhibitor paradigm. However, as each individual gene mutation is rare, international collaborative studies will be essential for accurate measurements of risk and prognostication to assist individuals in the clinic.

View largeDownload slide

View largeDownload slide

Close modal