Hereditary acute myeloid leukemia associated with C-terminal CEBPA germline variants

Acute myeloid leukemia with germline CEBPA mutation is a subtype of acute myeloid leukemia that is associated with a favorable prognosis. Most of the reported cases of acute myeloid leukemia with CEBPA germline variants involve a germline variant in the N-terminus and a somatic variant in the C-terminus. There are only a few reported cases where the CEBPA germline variant has been identified in the C-terminus and the somatic variant in the N-terminus. This case report and review of the literature illustrates that, although acute myeloid leukemia with CEBPA N- or C-terminal germline variants have certain similarities such as atypically young age at diagnosis, frequent relapse, and favourable overall prognosis, there are also significant differences such as lower life-time penetrance of acute myeloid leukemia and shorter time to relapse for germline C-terminal cases. These findings add important information on the natural history and clinical outcomes of acute myeloid leukemia with germline CEBPA C-terminal variants and these findings should be considered in the management of patients and their family members.


Introduction
The CCAAT enhancer binding protein alpha (CEBPA) is a single exon gene located on chromosome 19 and consists of two N-terminal transactivating domains, a basic DNA binding domain and a C-terminus bZIP leucine-zipper dimerization domain [1]. The protein product, C/EBP-α, is an important transcription factor that plays a role in the differentiation of myeloid cells and regulates the expression of granulocyte specific genes [2]. Likely pathogenic and pathogenic germline CEBPA variants predisposing to acute myeloid leukemia (AML) are recognized as a distinct subtype of AML [2][3][4]. Germline CEBPA variants predisposing to AML account for approximately 0.65% of all new AML cases and 7-11% of AML with bi-allelic CEBPA mutations, and are associated with a favorable prognosis, despite frequent relapses, compared to AML with bi-allelic CEBPA mutations without a germline predisposing variant [2,5,6]. In the majority of described AML cases with germline CEBPA variants, a germline variant was found on the N-terminus of CEBPA and a second, somatic, variant found on the C-terminus [6]. The penetrance for N-terminal CEBPA germline variants is nearly 100% for the development of AML [7]. There are only a few reported cases of hereditary AML associated with CEBPA C-terminal germline variants and therefore less is known about the prognosis for these patients. In this case report and review of the literature, we aim to add to the literature of germline CEBPA C-terminal variants predisposing to AML by describing a family from Atlantic Canada with a CEBPA C-terminus germline variant (c.932A > C, p.Gln311Pro) and summarize the clinical characteristics of all published CEBPA C-terminus germline variants to help further advance our understanding of this AML subtype and its implications for the management of patients and their families.

Germline CEBPA variant testing
Germline genetic testing for the 40-year-old male with AML was conducted using a commercial 41-gene Hereditary Leukemia Next Generation Sequencing (NGS) Panel from Blueprint Genetics. DNA was extracted from cultured skin fibroblasts and a heterozygous germline CEBPA c.932A > C, p.Gln311Pro was identified at a variant allele frequency (VAF) of 48%. Targeted familial variant testing for the germline CEBPA c.932A > C, p.Gln311Pro variant was also done through Blueprint Genetics. DNA for the familial variant testing was obtained from a peripheral blood sample and from saliva for the sister and parents, respectively.

CEBPA variant search and review of the literature
A search for published cases of AML with CEBPA germline variants in the C-terminal was performed using PubMed and Google Scholar. The human genomic variant search engines VarSome®, ClinVar, and Mastermind® were also used to search for reported variants in the CEBPA (NM_004364.5) C-terminus, where the C-terminus was defined as amino acids 278-358. Each reported CEBPA C-terminus variant was inputted into the Mastermind® search engine to identify any published articles citing the particular variant. Each article was reviewed to determine if the CEBPA C-terminus variant was (1) germline origin and (2) associated with AML.

Informed consent
All individuals in this case report provided written consent to be a part of an Inherited Predisposition to Hematologic Malignancies research study at our institution. This study was reviewed and approved by the Nova Scotia Health research and ethics board. Consent for publication of clinical information in medical journals for research purposes was obtained at the time of enrolment in the study.

Case description
A 40-year-old Caucasian male of Scottish ancestry from Prince Edward Island, Canada, whose past medical history was significant for vitamin B12 deficiency, presented to medical attention with progressive fatigue and complete blood count (CBC) revealed normocytic anemia (hemoglobin (HGB) 83 g/L), thrombocytopenia (platelets (PLT) 93 × 10 9 g/L) and leukopenia (white blood cell count (WBC) 2.3 × 10 9 g/L) with 45% circulating blasts suggestive of acute leukemia. Bone marrow aspirate and biopsy confirmed the diagnosis of AML and ancillary testing identified normal cytogenetics (46 XY), four CEBPA variants, and a TET2 variant (Table 1). Germline testing using DNA extracted from cultured skin fibroblasts was done given the patient's young age at diagnosis and identified that one of the four CEBPA variants [CEBPA (NM_004364.5) c.932A > C, p.Gln311Pro] was germline in origin. The CEBPA (NM_004364.5) c.932A > C, p.Gln311Pro variant, located on the C-terminus of CEBPA, was present in a heterozygous state at variant allele frequency (VAF) of 48%. The patient received "3 + 7" (Daunorubicin-Cytarabine) induction chemotherapy and achieved a complete remission (CR), which was then consolidated with three cycles of high dose cytarabine (HiDAC) chemotherapy.
The patient's family history was significant for a maternal first cousin once removed (II.9) who was diagnosed with breast cancer in her late 50's followed by acute leukemia in her early 60's as well as a paternal grandfather with prostate cancer (I.1) (Fig. 1). The patient's younger sister (III.2), mother (II.4) and father (II.3) all consented to undergoing variant specific testing. At ages 70 and 40 years, respectively, the mother (II.4) and sister (III.2) were found to be Seven months after achieving first CR, the patient's AML relapsed. Next generation sequencing at the time of relapse revealed persistence of the germline CEBPA c.932A > C, p.Gln311Pro variant as well as one of the previous three somatic variants, CEBPA c.233_234dup, p.A79fs; disappearance of the other two somatic CEBPA variants; and a new TET2 variant (Table 1). Re-induction chemotherapy with fludarabine, cytarabine and filgrastim (FLAG) was given and a second CR was achieved. This was followed by a myeloablative matched-unrelated allogeneic hematopoietic stem cell transplant (HSCT) using peripheral blood stem cells. Although the patient's sister was a 10/10 human leukocyte antigen match, an unrelated donor was selected to reduce the risk of a donor derived leukemia, as the sister was known to carry the same deleterious germline CEBPA variant as the patient. At 17 months post HSCT, the patient remains in a CR.

Discussion
Most of what is known about the clinical features and outcomes of AML with germline CEBPA variants is based on cases of N-terminal germline variants. Our case report and comprehensive review of the literature highlights that, although there are some similarities between AML with germline N-and C-terminal CEBPA variants, there are also important differences. To date, a total of nine unique germline C-terminal CEBPA variants associated with AML have been reported in eleven unrelated families (Fig. 2, Table 2) [1,5,[8][9][10][11][12]. Of these, the CEBPA c.932A > C, p.Gln311Pro variant that was identified in our case is one of only two variants to have been reported in separate, unrelated families. It is classified as a likely pathogenic variant according to the American College of Medical Genetics/ Association for Molecular Pathology (ACMG/AMP) criteria [13]. Of the other eight variants, three are classified as either pathogenic or likely pathogenic and five as a variant of uncertain significance ( Table 2, Fig. 2) [1,5,[8][9][10][11][12][13].
In terms of similarities, deleterious germline variants in both N-and C-terminal regions of CEBPA predispose to de novo AML, without any preceding dysplastic or cytopenic phase, and compared to sporadic AML, which has a median age at diagnosis of 68 years, the affected individuals typically develop AML at a relatively young age, often less than age 50 years [, , 2, 6, 7]. The median age at AML diagnosis for individuals with a CEBPA C-terminal germline variant, based on our case and those reported in the literature to date, is 30 years, with a range of 9-60 years old (Table 2). Excluding the individuals with a CEBPA germline C-terminal VUS, the median age of AML onset remains similar  [2,6,7]. While it is challenging to determine the prognosis of AML associated with a CEBPA C-terminal germline variant based on the limited information in the small number of published cases, it appears that, just as in AML with CEBPA N-terminal germline variant, the prognosis is favorable. The ten-year overall survival (OS) for AML with CEBPA N-terminal germline variants is 67% [7]. In the cases of AML with CEBPA C-terminal germline variants, the follow up data was variable. However, of the eight patients with follow up data, five (62.5%) were still alive at time of last follow up ( Table 2).
The prognosis for AML with germline CEBPA mutation is favorable but there are high rates of AML relapse, including both relapse of original disease or occurrence of a second de novo AML. There is an estimated 56% incidence of AML relapse at 10 years in individuals with a CEBPA N-terminal variant [7]. Although the number of reported cases is low, from our literature review, the rate of AML relapse in individuals with a CEBPA C-terminal variant was lower, at 27%, but with a shorter median duration of follow up of 2 years (range 7 months to 11 years) ( Table 2). Another notable difference was the time to first relapse; with C-terminal germline variant patients having a shorter time to first relapse, with a median of 9 months (range 7-12 months), compared to those with an N-terminal variant, with a median of 27 months [7].
The most distinctive difference, however, between AML associated with CEBPA C-versus N-terminal germline variants is the degree of penetrance of the variants. CEBPA N-terminal germline variants are highly penetrant, with a reported life-time incidence of AML between 90 and 100% [7]. Conversely, CEBPA C-terminal germline variants appear to have incomplete penetrance [6]. A penetrance of 46% (i.e. 12 of 26 confirmed germline carriers of a CEBPA germline C-terminal variant developed AML), was found based on data from our case and all other currently published cases ( Table 2). Excluding the individuals with a CEBPA germline C-terminal VUS, the penetrance would be 50%. Another difference is the types of variants themselves. Germline CEBPA C-terminal variants are mostly missense variants (67%) compared to N-terminal variants, which are most frequently frameshift variants ( Fig. 2; Table 2) [6].
Identification of germline predisposition variants in AML has important clinical implications for the prognosis and management of patients. There should be a high degree of suspicion with pursuit of clinical germline genetic testing for individuals with AML in which molecular testing reveals two or more CEBPA variants in the leukemic clone and/or have persistence of a CEBPA variant at a VAF of ~ 50% at the end of induction therapy despite achieving a CR. When the presence of a germline CEBPA variant is suspected, the patient should ideally be referred to a centre with expertise in germline predisposition syndromes for appropriate genetic counselling and germline testing [14].
Our case and literature review highlight the importance of early consideration of HSCT in first CR for patients with AML with CEBPA C-terminal germline variants. While this subtype of AML appears to be sensitive to chemotherapy and has a favorable prognosis, there remains a high risk of early relapse, as occurred in our patient's case [1]. An HSCT in CR is the only potential therapy that can rid the bone marrow of the predisposing germline variant, thereby decreasing the risk of future relapse and the need for further intensive induction chemotherapy [14]. Given the short interval between CR and first relapse observed in individuals with AML and CEBPA C-terminal germline variants, early identification of the germline variant and  1 3 testing any potential related stem cells for presence of the same variant is of critical importance to avoid delays in donor selection and ultimately delay in HSCT. Donorderived leukemias have been reported with the use of stem cells from donors carrying a deleterious germline CEBPA variant [15]. It is therefore recommended that such donors be excluded and, if no suitable related donor is available, that a matched unrelated donor is selected [14]. There are also important considerations for family members of individuals with AML and a CEBPA C-terminal germline variant. These individuals should be referred to a centre with expertise in germline predisposition syndromes to receive genetic counselling and consideration for genetic testing [14]. Follow up and surveillance of CEBPA C-terminal variant carriers is based on available data and expert opinion and consists of routine monitoring of CBCs for early detection of a possible myeloid malignancy as well as early bone marrow aspirate and biopsy if any signs of new unexplained cytopenia(s) arise [6][7][8][9][10][11][12][13][14].
Our case and review of the literature sheds new light on the natural history and clinical outcomes of AML with CEBPA C-terminal germline variants. A limitation to these findings is the heterogeneity of available information from the published cases of AML with CEBPA C-terminal germline variants. As outlined in Table 2, there were gaps in the information relating to clinical outcomes, testing of other family members and presence of other genetic variants in the leukemic clone. It is uncertain how these missing data would impact the reported findings. Another limitation is the potential presence of other co-occurring variants in genes known to be associated with germline predisposition to AML. This is seen in Family 1, where there is a germline RUNX1 c.421T > G, p.Ser141Ala variant that was found in conjunction with the likely pathogenic CEBPA C-terminal germline variant (Table 2). This RUNX1 variant is classified as a VUS according to the ACMG/AMP criteria and thus it's contribution to the development of AML in the two individuals from Family 1 cannot be determined.
In summary, AML with CEBPA N-and C-terminal germline variants have certain overlapping features such as atypically young age at diagnosis, early and frequent relapse, and favourable overall prognosis. However, there appear to be significant differences with a lower life-time penetrance of AML and shorter time to relapse for C-terminal cases that must be considered when managing patients with AML with a CEBPA C-terminal germline variant and counselling their family members.
Author contributions AT identified the cases and obtained written informed consent for the case report as well as supervised and revised the manuscript. AH performed the literature review, wrote the