Attack of the Clones: CHIP in the Clinic

Charles O. Mattick (“Chip”) is a 68-year-old man with a history of coronary artery disease and recent placement of drug eluting stents in his left anterior descending and left circumflex arteries. His cardiologist sent blood tests following the procedure, which revealed a normocytic anemia with a hemoglobin of 11.4 g/dL, mean corpuscular volume of 90 fL, and red blood cell distribution width (RDW) that was elevated at 17.1 percent. He also had mild iron deficiency. As he had been placed on clopidogrel and an aspirin for the stents and had a family history of “leukemia” in his father, he underwent endoscopic evaluation of his gastrointestinal tract and had a myeloid next generation sequencing (NGS) panel assessment performed. While myeloid NGS panels have become more widely available, their use to identify germline cancer susceptibility has not been clearly demonstrated. An esophagogastroduodenoscopy (EGD) revealed a bleeding ulcer, which was sclerosed, and the NGS panel showed a DNMT3A abnormality with a variant allelic frequency of 6 percent. His hemoglobin recovered to a normal range during the next two months, as did his RDW and iron levels. He is referred to a hematologist for follow-up after being told that the NGS report stated that DNMT3A is associated with myeloid malignancies such as myelodysplastic syndromes (MDS), myeloproliferative neoplasms, and acute myeloid leukemia (AML). A bone marrow evaluation shows normal hematopoiesis and cellularity for his age, with a normal karyotype.

While at first it might have seemed that Chip O. Mattick had a clonal cytopenia of undetermined significance (CCUS),¹  his anemia was actually caused by bleeding from a gastric ulcer. For conditions that primarily affect older adults, like myeloid malignancies, other causes of cytopenias must be identified. “Response” to agents used to treat conditions such as MDS may be due to resolution of other underlying conditions causing blood loss, as evidenced by measurable response rates in MDS patients enrolled to placebo arms of randomized studies. In a recent trial in which lower-risk MDS patients with ring sideroblasts were randomized to receive luspatercept or placebo, those on the luspatercept had a transfusion independence response rate of 38 percent, while those on the placebo arm had a 13 percent response rate.²  This patient, as his name implies, has clonal hematopoiesis of indeterminate potential (CHIP), a term applied to individuals with hematologic malignancy-associated somatic mutations in the blood or bone marrow, but without other diagnostic criteria for a hematologic malignancy and with preserved blood counts.¹  While Chip O. Mattick has a name for his condition (and even an acronym), this may serve only to increase his anxiety about a future diagnosis that may never even come to be, for which he has been ill prepared.

Somatic exonic mutations are acquired at a rate of approximately 1.3 per hematopoietic stem cell per decade, and are rarely detected in people under age 40 years.^3,4  Clonal hematopoiesis can be found in approximately 10 percent of people without known hematologic malignancies older than 65 years, in almost 12 percent of those aged 80 to 89 years, and in more than 18 percent of those older than 90 years. Those with a detectable mutation are at an 11-fold higher risk for developing a hematologic malignancy than those without such mutations, and a nearly 50-fold risk if the mutation has a variant allele frequency (VAF) of 10 percent or higher. Still, the absolute risk is small: 4 percent with a median follow-up of almost eight years.

Carrying a mutation is associated with increased all-cause mortality (hazard ratio, 1.4), much of which is due to cardiovascular causes, which occurs at twice the rate in people with CHIP. Presumably, cardiovascular events occur through a shared effect on inflammation, localizing to cells of the monocyte-macrophage lineage, and propagation of atherosclerotic plaques.⁵  Identifying people at increased risk may be helpful, but only if interventions that would consequently increase that risk can also be identified –– an area of active investigation.

Lucy Mia and Jeannie Mendelson are sisters in their early 60s with no personal or family history of hematologic disorders. Lucy Mia was diagnosed with anemia and mild leukopenia two years ago when a routine CBC noted her hemoglobin of 10.7 g/dL, a mean corpuscular volume of 98 fL, and a white blood cell count of 3.2 × 10⁹/L. She was asymptomatic, but absent a ready explanation for her cytopenias, a bone marrow biopsy was performed. This was notable for mild hypocellularity (30%), a lack of dysplasia, no blasts, and a normal karyotype. Her hematologist concluded that she had an idiopathic cytopenia of undetermined significance (ICUS) and ordered an NGS sequencing panel of her peripheral blood. The report listed three variants: an IDH2 R140Q mutation with a VAF of 22 percent, a frameshift in ASXL1 with a VAF of 7 percent, and a DDX41 K381* nonsense mutation with a 53 percent VAF. While mutations in these genes are typical of myeloid malignancies, they are not considered diagnostic and Ms. Mia’s condition was reclassified as a clonal cytopenia of undetermined significance (CCUS),¹  which puts her into a substantially different risk group than our previous example patient.

Patients with ICUS have a high rate of clonal hematopoiesis (~30-40%) indicative of CCUS and early studies suggest that they have a high risk for progression to a frank hematologic malignancy.^6,7  Certain genetic profiles have the greatest risk (10-20% per year) of such somatic mutations in more than one myeloid malignancy driver gene or isolated mutations of a splicing factor gene, JAK2, or RUNX1.⁸  A lone mutation in one of the three most common CHIP genes, DNMT3A, TET2, or ASXL1, has a more moderate risk of 5 to 10 percent per year. Patients who go on to develop AML often have preceding clonal hematopoiesis even in the absence of abnormal blood counts. Somatic mutations of IDH1, IDH2, and TP53 may be particularly likely to progress to AML, although in noncytopenic patients, the latency can last many years.⁹ 

After a period of relatively stable counts, Lucy Mia develops more profound cytopenias and 5 percent circulating blasts. A bone marrow biopsy is notable for a hypercellular marrow with 40 percent blasts, trisomy 8, and the acquisition of an NRAS G12D mutation and a second DDX41 mutation (L87F), both with a VAF of 33 percent. She is slated for induction chemotherapy for what is now AML, with a plan to proceed to allogeneic stem cell transplantation. Her sister Jeannie Mendelson is a perfect HLA match, but due to occasional mild leukopenia, she has an NGS sequencing panel performed. This reveals the same DDX41 K381* mutation Lucy Mia has with a comparable VAF of 49 percent, strongly suggesting that this mutation is congenital in both sisters and likely predisposed Lucy to develop AML.¹⁰ An alternative stem cell donor is recommended to avoid the potential risk of developing a donor-derived malignancy.

Several somatically mutated genes can be congenitally mutated in some cases. They may not cause early onset of disease, syndromic features, or a recognizable hematologic prodrome. Due to congenital de novo mutations and variants with incomplete penetrance, there may be no suspicious family history as in the case of Lucy and Jeannie. It is important to identify these germline variants as they have implications for family members, especially if they are being considered as potential stem cell donors.¹¹  If instead of the DDX41 variant, Jeannie had been found to carry the same low-abundance, somatic DNMT3A mutation as our first patient Chip, the need for an alternative donor would be less clear. Early data suggest that typical CHIP mutations in donor cells are not necessarily associated with adverse outcomes.¹²  However, larger studies that examine the type and abundance of clonal mutations are needed to better characterize this phenomenon.

For both Charles O. Mattick and Lucy Mia, continued monitoring of blood counts is necessary, with more intensive monitoring for patients who have germline or somatic abnormalities more characteristic of myeloid malignancies, and/or for those with multiple abnormalities. Whether interventions in those with molecular abnormalities typical of myeloid malignancies can modify risk of disease evolution (such as Vitamin C for those with TET2 lesions) has not yet been determined.¹³  Prospective studies that track clone size, molecular evolution, and clinical outcomes will require large numbers of patients followed over long periods. Early intervention studies might instead focus on patients with the highest risk (like Jeannie) who carry a predisposition allele, CCUS patients, or individuals with CHIP after cytotoxic therapy. Utlimately, we will need to develop approaches to attack the highest risk clones before they attack us.