HLH and TET2 Mutation Presenting after First Cycle of CLL Treatment

• Chronic B-cell
• Hemophagocytic
• Lymphocytic leukemia
• Lymphohistiocytosis
• TET2 protein

A male patient, aged 75 years, with a history of chronic lymphocytic leukemia (CLL), diagnosed in 2002 and observed without treatment for decades, began to show declining white blood cell (WBC) and platelet count in May, 2022. On May 19, 2022, bone marrow biopsy showed 60% cellularity with nodular and interstitial CLL involvement occupying a third of sampled marrow. Fluorescence in situ hybridization (FISH) analysis revealed 13q14 deletion; this abnormality was not detected in conventional chromosome analysis. The patient received one cycle of bendamustine+rituximab, along with pegfilgrastim, but he developed severe thrombocytopenia, worsening anemia, and fevers with no infectious source identified. Corticosteroids did not improve blood counts, and cyclosporine was initiated for presumed marrow aplasia. Romiplostim was also administered on June 15 and June 21, each time at a dose of 1 microgram per kilogram, with no significant improvement in platelet count. Follow-up bone marrow was performed due to high clinical suspicion for hemophagocytic syndrome, with laboratory evidence supporting the possibility of hemophagocytic lymphohistiocytosis (HLH). Bone marrow biopsy on July 1, 2022 did not show a significant increase in histiocytes compared to the prior specimen, and evidence of hemophagocytosis was scant by morphology, but there was significant newly developed reticulin fibrosis and megakaryocytic hyperplasia with atypia, along with persistent CLL involvement. Chromosome analysis continued to show normal male karyotype, but next generation sequencing revealed TET2 mutation (L1101S) at 32% burden.

Laboratory values from before and throughout this time period are shown in Table 1. The highly elevated soluble CD25 (sCD25) , CXCL9, and ferritin values appeared to support the clinical diagnosis of HLH, which had been suspected due to pancytopenia, unexplained new fever, low fibrinogen, high triglycerides, and splenomegaly. On July 7, 2022, etoposide and high-dose dexamethasone were initiated per HLH protocol, which resulted in clinical stabilization within one week. However, due to persistent neutropenia, etoposide was held, and filgrastim support was added to dexamethasone treatment, resulting in eventual improvement in platelet and neutrophil counts, and gradual improvement in ferritin levels.

Figure 1 shows May 19 and July 1, 2022 bone marrow samples, including hematoxylin and eosin stains, silver stains for reticulin fibrosis, and CD68 stains for histiocytes. A pertinent negative was lack of robust evidence for hemophagocytosis in the latter specimen, with no appreciable difference in CD68 pattern between the two dates; however, bone marrow biopsy is considered to be an unreliable predictor of the probability of HLH.¹ HLH is a hyperinflammatory immune syndrome that can be seen in younger patients due to underlying genetics, but in adults is most often seen as a consequence of a precipitating condition such as infection or underlying hematolymphoid malignancy.² There are eight diagnostic criteria, of which at least five must be present for HLH diagnosis.³ In this case, the patient exhibited (1) fever, (2) splenomegaly, (3) cytopenias, (4) hypertriglyceridemia and hypofibrinogenemia (June 24, 2022 triglyceride level was 454 mg/dL), and (5) hyperferritinemia. The patient did not show (6) detectable hemophagocytosis in bone marrow, spleen, or lymph nodes by morphology (only bone marrow was sampled) or (7) low/absent natural killer (NK)-cell functional activity (not assessed). The eighth criterion, elevated sCD25, was not present at the cutoff considered positive for HLH (>2400 U/mL, which converts to 21,239 pg/mL); our patient’s sCD25 value was 11,936 pg/mL, which is significantly elevated above normal range but not at HLH threshold. The incidence of HLH has been estimated at 1-2 per million annually, but the syndrome is likely underreported, given its co-existence with other highly morbid conditions such as cancer and autoimmune disease.⁴

• Download figure
• Open in new tab
• Download PowerPoint

Figure 1.

Bone marrow trephine core biopsies from May 19, 2022 and July 1, 2022. Partial involvement by CLL can be seen in both specimens (circled), but an atypical megakaryocytic proliferation developed within the six intervening weeks (square). This was accompanied by development of moderate to severe reticulin fibrosis (triangle). Despite laboratory and clinical evidence for HLH, there was no appreciable increase in CD68+ histiocytes, and histiocytes did not show a hemophagocytic pattern (no engulfment of hematopoietic elements).

Although a myeloid neoplasm was suspected due to the altered appearance of the July 1 bone marrow, the TET2 result was not reported by the reference laboratory until July 26, at which time the presence of a second hematolymphoid malignancy could be considered as confirmed. At an allele burden of 32%, the TET2 finding seems to exceed the threshold for being an incidental discovery of clonal hematopoiesis of indeterminate potential (CHiP), and would therefore appear to be directly related to the observed megakaryocyte morphology and fibrosis, and also to the decline in blood counts prior to the May 19 marrow, despite the relatively bland appearance of that marrow. Applying precise World Health Organization nomenclature to the patient’s TET2+ myeloid malignancy is difficult with just the July 1 marrow specimen, given the patient’s acute illness and recent treatments with marrow stimulating agents including romiplostim and g-CSF. However, the availability of targeted sequencing analysis was key in achieving diagnostic clarity, as it would have been difficult to definitively attribute marrow findings to myeloid neoplasia without evidence of clonal myelopoiesis.

It is curious that the patient’s rapid deterioration happened shortly after administration of bendamustine and rituximab, given others have reported successful management of CLL-related HLH with rituximab therapy.⁵ Whereas the TET2-mutated myeloid process almost certainly predated administration of chemotherapy, the abrupt change in complete blood count (charted in Table 1) likely corresponds to development of the inflammatory (HLH) component to this patient’s complex medical picture. Others have proposed a causal relationship between TET2 mutation and adult-onset HLH.² An interesting pathologic observation in this case is the unmasking of neoplastic myeloid morphology within a 6-week period between the two bone marrow biopsies (Figure 1). The second biopsy was performed only 2 days after HLH markers were ordered and drawn on peripheral blood; sCD25 and CXCL9 results were not yet available when the decision was made to further evaluate the patient’s status through bone marrow assessment. This was fortuitous: given the imperfect reliability of bone marrow morphology for predicting HLH, the elevations in sCD25 and CXCL9 may have been taken as adequate HLH confirmation on their own, and suspicion for myeloid malignancy – triggered by morphology and resulting in submission of marrow for sequencing analysis – would have been absent.

It would have been interesting to see how the TET2-mutated myeloid neoplasm behaved in this patient after full recovery from acute pancytopenia and systemic illness. The patient’s initial steady improvement, both in laboratory values and clinical status, was notable given the high mortality rate of HLH. However, the patient then showed declining performance status and was admitted to the hospital with possible infection in September, 2022. Although he was subsequently discharged after doing well, he went to the emergency room a week later with fever and potential infection, declining further treatment. From that facility (not part of the authors’ institution, so records are not available), the patient chose to be discharged home to hospice care and passed away shortly afterwards. A recent multicenter retrospective study showed very poor outcome in HLH associated with hematologic malignancies, with much better overall survival in other adult-onset HLH subtypes including autoimmune and infection-associated HLH.⁴ In addition, the authors of that study showed that sCD25 offers both diagnostic and prognostic significance: in their analysis, there was a statistically significant decrease in survival among patients with sCD25 >7500 U/mL, which converts to 66,372 pg/mL.⁴ Our patient’s value of 11,946 pg/mL falls well below that threshold, perhaps correlating with his ability to initially recover from acute illness during the month of August.

Administration of romiplostim raises consideration of the reported connection between thrombopoietin analogs and fibrosis. A published retrospective review of 100 primary immune thrombocytopenia patients treated with eltrombopag or romiplostim revealed development of grade 3 bone marrow fibrosis in two patients (2%).⁶ In those cases, bone marrow biopsy was performed due to loss of platelet count response (to thrombopoietin analog and rescue steroid treatment). In both cases, fibrosis had not been detected previously, and other conditions causing fibrosis had been excluded. Fibrosis reverted to grade 1 at 6 months after drug cessation, with normalization of lactate dehydrogenase and hemoglobin levels.⁶ As such, the two patients in that study had not developed a clonal malignancy causing marrow fibrosis, but a transient non-neoplastic fibrotic drug reaction. Two observations argue against our patient’s marrow fibrosis resulting directly from two doses of romiplostim, given 10 and 16 days prior to bone marrow biopsy: (1) the time course seems very short for development of moderate to severe reticulin fibrosis, and (2) the role of TET2 mutation – which at 32% allele burden, almost certainly predated romiplostim administration – would be called into question. Romiplostim may have contributed to unmasking of TET2-driven myeloproliferation, but it seems unlikely to be the primary cause for altered marrow morphology, especially when considering that worsening cytopenia predated its administration.

• Received September 22, 2022.
• Revision received March 23, 2023.
• Accepted December 12, 2023.