Two Cases of Possible Familial Chronic Myeloid Leukemia in a Family with Extensive History of Cancer

Unlike other myeloid malignancies, CML is a largely homogeneous disease almost exclusively characterized by the presence of a single oncogenic gene fusion. The encoded protein, BCR-ABL1, is the result of a translocation between chromosomes 9 and 22, a fusion commonly referred to as the Philadelphia chromosome [1, 2]. The identification of this abnormality has led to the development and widespread use of tyrosine kinase inhibitors (TKIs), resulting in dramatically improved outcomes for nearly all patients with CML such that life expectancy is now on par with age-matched controls without CML [3, 4]. In fact, newer studies suggest that prolonged use of TKIs has resulted in cures for a subset of patients, with the option of TKI withdrawal resulting in sustained treatment-free remissions in the majority of cases [5-9].

Despite this remarkable therapeutic progress, it is still unclear how or why the Philadelphia chromosome first appears in a given patient and whether heritable aspects or environmental exposures play a substantial role in the etiology of this disease. Unlike some other hematologic disorders, such as myeloproliferative neoplasm, where family members are thought to have an increased risk of hematologic malignancy, no such risk relationship has been found for CML [10, 11]. However, several cases of potentially familial CML have been reported [12-19]. Here, we describe a case of CML in a young woman diagnosed shortly before her maternal great aunt was also diagnosed with CML. We describe these 2 cases, their treatment and disease characteristics, and propose potential explanations for familial cases of CML.

In July 2017, a 24-year-old woman (patient 1) presented to her local emergency department with severe abdominal pain and a several week history of increased fatigue, low-grade fevers, and night sweats. At this time, her WBC count was 128 × 10⁹/L, hemoglobin (Hgb) 10.1 g/dL, and platelets (Plt) 573 × 10⁹/L. A computed tomography scan of the abdomen and pelvis showed a markedly enlarged spleen measuring 22 × 19 × 11 cm. Hydroxyurea was started. A bone marrow biopsy showed chronic phase CML with BCR-ABL1 rearrangement detected by PCR, without increase in blast percentage or reticulin fibrosis. Conventional cytogenetics were not reported. Two weeks later, she was referred to our leukemia center for further treatment.

Peripheral blood (PB) PCR demonstrated 87.43% BCR-ABL1 to ABL1 transcripts (co-expression of b3a2 and b2a2 transcripts) by the International Scale (BCR-ABL1^IS). The patient was started on a clinical trial of low-dose dasatinib (50 mg/day) [20]. At 1 month follow-up, her spleen was significantly reduced in size and her PB counts normalized. PB PCR demonstrated a decrease in BCR-ABL1^IS to 0.25% at 3 months and 0.02% at 6 months. Bone marrow aspiration at 6 months demonstrated cellular bone marrow with tri-lineage hematopoiesis, no increase in blasts, and diploid (46,XX) cytogenetics.

At a follow-up visit, the patient disclosed that her maternal great aunt had recently received a CML diagnosis. Due to this new family history of hematologic malignancy, a targeted next-generation sequencing-based 81-gene panel was performed on bone marrow at this time to evaluate for any germline mutations that could underlie a putative genetic predisposition to CML. As she was in a deep molecular remission at the time of sequencing, no somatic mutations were detected, though 2 variants of probable germline origin were identified in KDM6A 2177C>A (T726K) and TET2 1088C>T (P363L). Approximately 2.5 years after initial diagnosis, PB PCR began to trend upward, at which time ABL1 was sequenced and no kinase domain mutations were detected. Given the lack of ABL1 kinase domain mutations, dasatinib was increased to 100 mg per day. After 3 months at this higher dasatinib dose, PCR continued to increase, reaching 8.84% in bone marrow, at which time treatment was switched to bosutinib 400 mg daily. After approximately 2 months of bosutinib treatment, PB PCR had decreased to 0.4%. Patient remains well and will continue to follow in our clinic every 2 months with PCR monitoring.

Patient 2 is a 73-year-old woman (and maternal great aunt of patient 1) who presented to her local emergency department in January 2017 with back pain and a 6-month history of progressive weight loss, night sweats, and increased back pain. At this time, her WBC was 130 × 10⁹/L. Bone marrow biopsy showed chronic phase CML. Cytogenetics included 46,XX, t(9;22) (q34;q11.2). She was started on hydroxyurea and transitioned to dasatinib 100 mg/day. She came to our institution for a second opinion.

Upon presentation to our leukemia center, WBC was 16 × 10⁹/L. Bone marrow biopsy revealed hypercellular bone marrow with granulocytic hyperplasia and mild reticulin fibrosis, and FISH was positive for BCR-ABL1 rearrangement. A targeted 81-gene panel was performed, and a mutation in NOTCH1 7484A>G (D2495G) was identified. Additionally, 4 variants of probable germline origin were found: CBLC 1303C>T (P435S), NF1 528T>A (D176E), NOTCH1 6733 G>A (G2245R), and STAG1 3396G>C (Q1132H). PB PCR demonstrated BCR-ABL1^IS (b3a2 transcript) of 91.77%. The patient was started on the same protocol as patient 1 using lower dose dasatinib (50 mg/day) [20]. Approximately 18 months after therapy initiation, patient 2 developed a pleural effusion attributed to dasatinib. Following thoracentesis and 2 weeks of dasatinib discontinuation, dasatinib dose was reduced to 20 mg/day. She continues on this regimen, with BCR-ABL1^IS detectable via PB PCR at 0.1% 46 months from the time of her initial diagnosis.

Patient 2 previously had localized uterine cancer that was surgically resected approximately 35 years prior, with no exposure to chemotherapy. Given the relationship to patient 1, we gathered a detailed family oncologic history (Fig. 1). Patient 2’s sister was diagnosed with liver cancer at age 50, and all 3 of that sister’s children were diagnosed with breast cancer between 30 and 40 years of age. A brother with substantial smoking history had lung cancer at age 70. Another sister had stomach cancer at age 62 and was the grandmother of patient 1. No members of this family had undergone any genetic testing to identify cancer predisposition.

Whereas patient 1 had no history of smoking or occupational chemical exposures, patient 2 reported a 15 pack-year smoking history, quitting 37 years prior to her CML diagnosis. Patient 2 denied any occupational exposures but did report that during her childhood in the Mississippi River Delta, she and her siblings had extensive exposure to dichlorodiphenyltrichloroethane (DDT) that was sprayed by crop dusters in fields near their home.

We have described an instance of third-degree relatives sharing a CML diagnosis. CML is not generally considered to harbor a familial pattern although rare cases of possible familial CML have been described [12, 14-19] (Table 1). In a systematic review of chronic myeloproliferative neoplasms occurring in families, almost 10% of cases involved CML [21]. Most of these reported cases have been described in first-degree relatives, leaving minimal information on cases described in more distantly related individuals. Indeed, fewer than 8,500 cases of CML are estimated to occur in the USA in 2020 [22]. The rarity of this disease argues against, but does not exclude, coincidental cases of CML in the same family. In the case presented here, the family had a significant exposure history to genotoxic chemicals in earlier generations as well as a compelling history of cancer. This history raises the potential for both genetic and environmental factors contributing to the development of CML in these 2 patients.

Compared to myelodysplastic syndrome and acute myeloid leukemia, CML is not genetically diverse. The BCR-ABL1 oncogene is usually the sole identifiable genetic lesion in this disease, contrasting with the variety and often numerous aberrations found in other myeloid malignancies. However, recent studies demonstrate additional chromosomal abnormalities within or aside from the BCR/ABL1 clone may be present in a small subset of cases and portend poor prognosis [23, 24]. Furthermore, the staggering conservation of chromosomal breakpoints in chromosome 9q34, where BCR is found, and 22q11, where ABL1 is found, would suggest that CML should be a rather uncommon entity in the absence of specific inciting factors [25]. To date, only few of these factors have been convincingly identified, and yet these still explain only a subset of CML cases diagnosed each year. These potential contributing factors to CML pathogenesis include genetic predisposition, prior malignancy, or toxic exposures.

The presence of CML in a younger patient with no known toxic exposures, personal history of cancer, or smoking history, but with a strong family history of cancer and with a close relative with CML, raises the possibility that there may be a genetic predisposition to this disease in this particular case. Despite detection of germline variants in both patients, these variants were not over-lapping. A recent case-control study described poly-morphisms in CYP26B1, PPARD, PSMB2, PSMB10, and TNFRSF10D that were associated with a predisposition to CML in a cohort of French patients [26]. While it would be of interest to evaluate if our patients harbored the same haplotypes containing the rare alleles for their predictive genes, these genes have not been heavily associated with myeloid leukemia previously and, thus, are not included on our targeted 81-gene panel. This same study demonstrated that younger patients had higher numbers of risk alleles than older patients, suggesting that CML in a younger patient may be linked to an inherited predisposition [26]. To this end, patient 1 had 2 germline variants (TET2 P363L and KDM6A T726K), though neither of these have been linked to an increased risk for hematologic malignancies. Patient 2 had germline variants in several leukemia-associated genes (CBLC P435S, NF1 D167E, NOTCH1 G2245R, and STAG1 Q1132H). Of these, only CBLC P435S has been described in vitro to have no effect on proliferation rates of 32D cells [27], while the others have not been functionally tested to our knowledge.

Another recent Swedish case-control study has described an increased prevalence of prior malignancy in CML patients [28]. Patient 2 had a remote history of surgically resected uterine cancer. Interestingly, a previous history of gynecologic cancers did not yield a statistically significant odds ratio for malignancy prior to CML diagnosis in this Swedish study [28]. Notably, patient 1 did not have a cancer diagnosis prior to CML.

The exposure history of patient 2 is also of interest. Per her report, she and her siblings had frequent exposure to DDT, as they would frequently chase crop dusters spraying the compound in rural Mississippi. The carcinogenicity of DDT has been a controversial subject over the past several decades, with little conclusive evidence regarding its potential carcinogenic properties. There has been no association between DDT exposure and leukemia/lymphoma, hepatocellular carcinoma, or endometrial, colorectal, prostate, or testicular cancers [29, 30]. Controversy still exists between DDT exposure and breast cancer risk – particularly in the context of in utero DDT exposure [31-33]. However, data from our institution demonstrated an association between exposure to agricultural chemicals and CML, raising the possibility that environmental factors contributed to the development of patient 2’s CML [34]. High-dose radiation and, to a lesser extent, smoking, have also been associated with the development of CML [35-38]. While patient 2 did report this exposure history, she was diagnosed with CML at the age of 74 – well above the median age of a CML diagnosis in the USA [22].

This report is limited by the lack of comprehensive genetic sequencing. Targeted sequencing of 81 leukemia-associated genes was performed in both cases, though no overlapping mutations or variants between patients 1 and 2 were observed. Patient 1 was in a deep molecular remission at the time of sequencing, which limited the ability to assess leukemia-associated mutations although notably no overlapping germline mutations were detected. Without more comprehensive genetic sequencing, we cannot completely exclude the possibility that these 2 cases do share genetic variants. Furthermore, it would also be of interest to perform a systematic genetic analysis on other family members affected by cancer to further evaluate the possibility of an underlying cancer predisposition.

In this report, we described CML occurring in 2 third-degree relatives with an extensive family history of various cancers. While both patients had germline variants in genes associated with leukemia development, the same variants were not found in each patient. It is possible that childhood exposure history in patient 2 may have contributed to CML predisposition. The possibility remains that these diagnoses were merely coincidental; however, a heritable contribution to CML development in these family members cannot be excluded. In light of other reports of CML cases arising in close relatives, future studies into potential genetic or environmental factors that contribute to CML pathogenesis are warranted.

Written informed consent was obtained from the patients for publication of this case.

The authors have no conflicts of interest to declare.

This work was supported in part by the Cancer Center Support Grant from the NIH/NCI, P30 CA016672.

M.J.L.A. compiled data and wrote the manuscript; C.B.B. critically revised the manuscript; N.J.S. critically revised the manuscript and oversaw the project. All the authors reviewed and approved the final version of the manuscript.