Hemophagocytic lymphohistiocytosis and myelodysplastic syndrome: a case report and review of the literature

Hemophagocytic lymphohistiocytosis (HLH) is characterized by deregulated immunity and end organ damage. Primary HLH is linked with autosomal recessive and X-linked mutations. Secondary HLH results from predisposing conditions. Both primary and secondary HLH may be activated by an infections and malignancies. [1] The association of HLH with hematologic malignancies is accepted. However, HLH initiated by myelodysplastic syndrome (MDS) is exceedingly unusual. Although dysplasia is observed in patients with hemophagocytosis, a hyperinflammatory entity characterized by dysplastic cytopenias, MDS like karyotypic abnormality and hemophagocytosis suggests a different spectrum of disease, in which HLH originates from a clonal disorder. In this report, we present an elderly female who fulfilled HLH 2004 criteria exhibiting erythroid and megakaryocytic dysplasia associated with trisomy 8 [+ 8]. Additionally, we examine published cases in English literature to identify clinical, laboratory and cytogenetic features observed in MDS patients fulfilling HLH 2004 criteria.

An 83-year-old African American female presented with lethargy, temperature of 102.9° F and tachycardia. After 9 days of broad-spectrum antibiotics, she developed hemodynamic instability requiring vasopressors, worsening liver function (peak bilirubin of 12.5 mg/dL, Aspartate aminotransferase (AST) of 689 IU/L, Alanine aminotransferase (ALT) of 239 IU/L) and leukocytosis of 16,000/uL. Her hemoglobin was 8.9/uL and platelets had fallen to 26,000/uL. Disseminated intravascular coagulation (DIC) was considered given progressive severe thrombocytopenia of 13,000 U/L, fibrinogen < 70 mg/dL and d-dimer of 12.4 mg/L. Her ferritin was 9479 ng/mL and fasting triglycerides (TAG) were 321 mg/dL. Soluble interleukin-2Rα (CD25) was < 38 pg/dL. Her human immunodeficiency virus (HIV), rapid influenza A/B, hepatitis B/C serologies were all negative. Epstein-barr virus (EBV) viral load was negative. Given concern autoimmune hepatitis, a liver biopsy showed Kupffer cell hypertrophy with hemophagocytosis. Bone marrow biopsy demonstrated hemophagocytosis (Fig 1a, b). In addition, significant erythroid nuclear fragmentation and karyorrhexis were observed (Fig 1c). Marrow cytogenetics showed 47, XX +8 [6], 46, XX [5]. Next generation sequencing (NGS) including CSFR1, SF3B1, SRSF2, U2AF1, NRAS, KRAS, FLT3, JAK2, KIT, PHF6, PDGFRA, CDKN2A, IDH1, IDH2, TET2, EZH2, CEBPA, EP300, PTPN11, P53, CREBBP, IKZF1, IKZF3, NOTCH1, RUNX1, WT1 and NPM1 showed DNMT3A p.Arg736His (c.2207G>A) and DNMT3A p.Leu859Ter (c.2576T>A) at allele frequencies of 2.7 and 2.4 %, respectively. In view of her hemophagocytosis, cytopenias, high temperature, abnormal liver function test, low fibrinogen and elevated fasting triglycerides and ferritin, she fulfilled 5/8 HLH 2004 criteria. HLH-94 regimen was initiated with dexamethasone and etoposide. She developed neutropenic sepsis and etoposide was stopped. Blood cultures were positive for Escherichia coli. Patient expired after developing hemodynamic instability.

In addition to our case, three additional previously published reports were included in our reviewed cases. Clinical, laboratory and karyotypic data was analyzed to investigate potential features commonly observed in patients presenting with hyperinflammation and MDS. Data aggregation from four cases presenting with HLH and MDS allowed identification of clinical outcome in patients receiving HLH directed therapy or alternative regimens.

Efforts to elucidate pathogenesis of HLH demonstrate that expansion of CD8⁺ cytotoxic T cells, low Treg frequencies and cytokine storm are frequently observed [7]. In MDS, similar clonal T cell expansion and decreased Treg frequencies results in stem cell/progenitor apoptosis in low-risk disease. Rather than hyperinflammation normally found in HLH patients, chronic low-grade inflammation develops associated with tumor-necrosis factor-alpha (TNF-α), interleukin 1-beta (IL-1β) and IL-6 upregulation. Cytokine abnormalities in MDS leads to increased apoptosis and marrow hypercellularity. However, decreased marrow cellularity can develop in fraction of MDS patients harboring trisomy 8 karyotypic abnormality. Interestingly, our patient and two previously published cases harbored + 8 in their metaphase analysis at MDS diagnosis while HLH had developed [3, 5]. + 8 cells induce autologous T-cell oligoclonal expansion capable to target MDS precursor/progenitors resulting in characteristic marrow hypoplasia [8, 9]. In most of MDS cases, disease propagation depends on somatic mutations acquisition allowing dysplastic transformation. Mutations in perforin-dependent cytotoxicity are classically described in primary HLH. In our case, it is conceivable that myeloid mutations induced apoptotic and differentiation defects, and “facilitated” acquisition of HLH-like phenotype.

The preferred HLH treatment is HLH-94 protocol, but its administration is limited by MDS-induced cytopenias, as was the case with our patient. This is largely due to hemopoietic progenitor depletion initiated MDS and likely aggravated by HLH. Daitoku et al. reports superior outcome in a HLH-MDS patient treated with Methylprednisolone, CSA and decitabine[4]. Indeed, in vivo administration of hypomethylating agents (HMA) induce Foxp3+ Tregs expansion leading to immune suppression, and attenuates graft-versus-host disease [10]. The combination of HMA and immunosuppressive therapy may be promising treatment in HLH-MDS cases as suggested by the interesting outcomes in the case described by Daitoku et al.

We acknowledge limitations to our interpretation. It is possible that hyperinflammation initiated by HLH led to marrow dysplasia. However, the evidence of +8 strongly suggests a clonal etiology supporting MDS induced HLH. Secondly, sequencing did not include HLH like mutations such as PRF1, STX11, UNC13D, STXBP2, RAB27A, SH2D1A, BIRC4, LYST, ITK, SLC7A7, XMEN, HPS, among others. This may affect our ability to assign HLH as culprit for hyperinflammation in our case. However, Rui et al. recently demonstrated that epigenetic perturbations induced by DNMT3A mutations results in aberrant stem cell gene-expression associated with immune deregulation, which may have contributed to inflammatory manifestations in our case [11]. Additionally, recent data demonstrated that myeloid specific mutations including TET2, ASXL1, and DNMT3A can induce inflammasome activation in myelodysplasia and exacerbate inflammation [12-15]. In summary, HLH associated with MDS is an aggressive entity and should prompt careful evaluation hyperinflammatory signs. It is possible that targeting MDS hematopoiesis with hypomethylating agents in combination with immunosuppressive therapy to minimize hyperinflammation could improve life-threatening HLH in MDS patients.