Introduction

The protein of the upstream binding transcription factor (UBTF) plays a key role in ribosomal DNA and RNA transcription. Recent molecular studies on paediatric acute myeloid leukaemia (AML) led to the identification of tandem duplications (TDs) in exon 13 of the UBTF gene in 9% of relapsed AML¹ and 1.2%–4% of primary paediatric AML.^{2, 3} The presence of UBTF-TD was mutually exclusive with other genetically defined AML subgroups and associated with a particular genetic pattern with frequent accompanying FLT3-ITD or WT1 mutations, normal karyotype or trisomy 8. Subsequent studies identified UBTF-TD in 3% of adult AML patients below the age of 65 years, and a strong positive correlation with younger age.^{4, 5} Like in paediatric AML, the presence of UBTF-TD in young adults was a prognostic factor for inferior outcome, with lower complete remission rates and inferior survival independent of the presence of associated mutations.

Interestingly, the bone marrow morphology of UBTF-mutated AML frequently displayed myelodysplastic changes, and in some cases was compatible with the diagnosis of AML-M6 according to the FAB classification.⁴ A most recent observation described UBTF-TD as a driver mutation in a substantial portion of children with myelodysplastic syndrome (MDS).⁶ The recognition of the histopathological pattern of these cases will support the identification of patients in need for UBTF analysis, and thus greatly improve therapeutic decision-making. Therefore, we investigated the histopathology of 14 paediatric patients with de novo MDS with excess blasts (MDS-EB)^{7, 8} and UBTF-TD.

Materials and Methods

Results

The median age of the 14 patients with de novo MDS-EB and UBTF-TD was 10.9 years (range 4.7–17.1). There were 11 males and three females (Table S1). Absolute neutrophil count (ANC) was 5.9 g/L (range 2.0–14.5), and all patients presented with anaemia (median haemoglobin concentration 8.2 g/dL, range 6.5–9.4) and thrombocytopenia (median platelet count 60 g/L, range 20–129). The mean corpuscular volume of red cells was elevated for age in five patients (38%). All patients had normoblasts on blood smear (median 3.4/100 WBC, range 1–4), and myeloid precursors were noted in 12 cases (86%).

Histomorphology of bone marrow biopsies

Bone marrow biopsies obtained from UBTF-TD patients exhibited a consistently hypercellular marrow for age in the majority of cases (N = 13; 93%), with a median cellularity of 88% on core biopsies. The distribution pattern of haematopoiesis was found to be diffuse rather than patchy. The architecture of the bone marrow was visibly disturbed, showing atypically located erythropoiesis and a prominent hyperplastic, atypical megakaryopoiesis in all cases. A summary of the morphological features is provided in Table 1 and displayed in Figure 1.

Table 1. Morphological features of paediatric MDS with excess blasts and UBTF-TD

Cytology (bone marrow aspirate)
Blast count (%)	Median (range)	11.2% (5%–19%)
EP to MP ratio	Median (range)	1.8 (0.3–12)
Dyserythropoiesis (>10%)	Number (%)	14 (100)
Severe immaturity in erythropoiesis	Number (%)	14 (100)
Dysgranulopoiesis (>10%)	Number (%)	14 (100)
Dysmegakaryopoiesis (>10%)	Number (%)	14 (100)
Histology (core biopsy)
Hypercellular in BM histology	Number (%)	13 (97)
Severe immaturity in erythropoiesis	Number (%)	14 (100)
Abnormal localisation of erythropoiesis	Number (%)	14 (100)
Megakaryocytes (n/HPF)	Median (range)	43 (34–51)
Micromegakaryocytes (n/HPF)	Median (range)	17 (12–19)
Monolobulated megakaryocytes in % of total megakaryocyte number	Median (range)	65 (57–69)
Bilobulated megakaryocytes in % of total number of megakaryocytes	Median (range)	12 (11–15)
Megakaryocyte with separated nuclei in % of total number megakaryocytes	Median (range)	2 (1–4)

• BM, bone marrow; EP, erythropoiesis; MP, myelopoiesis; N, number; HPF, high-power field.

In 12 cases (86%), the ratio of myeloid to erythroid precursors was decreased, with a median ratio of 1.8 (range 0.3–12). Granulopoiesis was left-shifted with immature precursors in the centre of the core biopsy. Notably, 10 patients (71%) displayed myeloid precursors with few but giant nucleoli. Erythropoiesis presented significant dysplastic features, with peritrabecular, atypical localization. Additionally, erythropoiesis was immature, with numerous proerythroblast clusters present in all cases. Megaloblastoid changes were evident in five patients (35%).

Megakaryocytes were markedly increased in number and were observed in small occasional peritrabecular clusters. Prominent atypical features were evident in all cases, with an increase in nuclear bi- or monolobulation and the presence of micromegakaryocytes. Furthermore, blastoid immature megakaryocytes were increased in 10 patients (71%). Reticulin fibrosis was detected in two patients, characterized by the formation of a loose network of reticulin fibres with rare, perivascular intersections, corresponding to a myelofibrosis grade 1. Sinusoids were enlarged in six patients (43%).

Immunohistochemistry and immunofluorescence of bone marrow biopsies

Immunohistochemical analysis revealed scattered CD20-positive B cells, intermingled with CD3-positive T cells, without dense infiltrates or a significant increase in number. The TdT-positive hematogones exhibited a normal distribution and number. Notably, a clear increase in Naphtol–Chloracetate–Esterase-negative blasts was observed in all cases. In all but one case, the blasts were found to be negative for CD34, but positive for CD33. In 36%, blasts were also positive for CD123. CD42b immunohistochemistry highlighted micromegakaryocytes in all cases. Only isolated round-oval mast cells were visible. Additionally, a high number of isolated and small aggregates of CD123 plasmacytoid dendritic cells were observed in close to half of the cases (43%). In all instances, myeloid precursors positive for MPO displayed prominent UBTF-positive nucleoli. Immunohistochemistry and Naphtol–Chloracetat–Esterase stain is shown in Figure 2.

Cytogenetics and molecular findings

Cytogenetic analysis revealed a normal karyotype in 10 patients (71%), with no detection of monosomy 7. Four patients (29%) exhibited a trisomy 8.

Genetic analysis of UBTF-TD patients did not reveal germline predisposition mutations in GATA2, RUNX1, and SAMD9/9L. Somatic mutations were observed and are summarized in Figure 3. Among the detected mutations, those in WT1 were the most common, being present in six patients (43%). Morphologically, no significant differences in WT1-mutated versus WT1-wildtype could be detected.

Discussion

Since UBTF-TD has emerged as a novel subgroup-defining genetic lesion of AML and MDS in young patients, it is crucial that pathologists recognize the morphological features of these cases. In this study, we aimed to describe the characteristics of cytology and histopathology, which may indicate the presence of UBTF-TD to prompt further molecular profiling. We studied 14 UBTF-TD-positive cases with de novo MDS having excluded disease with prior chemo−/radiation therapy, prior aplastic anaemia, or known inherited bone marrow failure/predisposition syndrome. All cases were classified as MDS-EB according to the International Consensus Classification,^{7, 8} and qualify as childhood MDS with increased blasts cMDS-IB in the current WHO classification.¹¹

Except for one case with normocellularity, bone marrow biopsies were hypercellular for age. This contrasts with most cases of MDS-EB in GATA2 deficiency,¹² SAMD9/9L syndrome,¹³ or the classical inherited bone marrow failure disorders. Furthermore, it is distinct from MDS presenting as refractory cytopenia of childhood (RCC).^{8, 14, 15} Next to a hypocellular bone marrow in most instances, RCC displays a typical morphological pattern featuring patchy erythropoiesis and reduced megakaryopoiesis with occasional micromegakaryocytes, which is absent in UBTF-mutated myeloid neoplasia.

Notably, we observed marked dysplasia in all three lineages, particularly in the megakaryopoiesis. While megakaryocytes with dysplastic features are increased in de novo MDS and MDS with germline predisposition, such as GATA2 deficiency,^16-18 the predominance of monolobated cells in small clusters and an extremely high rate of atypia with only few normal megakaryocytes (<10%) can serve as a morphologic clue to a diagnosis of UBTF-TD. Another striking feature of UBTF-TD is morphologic dysplasia in the erythropoietic lineage. Erythropoiesis was severely immature, with numerous proerythroblasts and circulating normoblasts. This is in line with the observation that UBTF-TD in adult AML is associated with erythroid leukaemia (M6) and may be attributed to the overexpression of regulators of erythroid differentiation, such as HBB, HBA2, and ABO.^{4, 5}

Molecular analysis of UBTF-TD patients revealed that WT1 mutations were the most common co-occurring mutations, while germline predisposition mutations such as GATA-2 or SAMD9/9L were absent. Additionally, no monosomy 7 was detected. A similar genomic landscape was observed in UBTF-TD-positive adult AML.¹⁹ This molecular profile is remarkable, as previously, germline mutations such as GATA-2 deficiency as well as monosomy 7 were the most frequent genetic aberrations in advanced primary MDS of childhood.¹²

On H&E sections, we noticed prominent nucleoli in UBTF-TD bone marrow biopsies. Double immunofluorescence images confirmed the presence of prominent UBTF-positive nucleoli in myeloid precursors positive for MPO. Interestingly, this finding is consistent with previous reports of gain-of-function mutations in UBTF, such as UBTF E210K, observed in developmental neuroregression and associated with nucleolar dysregulation characterized by few large nucleoli in cultured fibroblasts.¹⁸

Nearly half of the patients in our study showed an increase in CD123-positive plasmacytoid dendritic cells. While in adult MDS/myeloproliferative neoplasia, bone marrow dendritic cell aggregates are associated with disease progression and immune dysregulation,¹⁷ lymphocyte counts assessed by CD20 and CD3 in UBTF-TD patients were unremarkable.

Immunohistochemical stainings revealed a high number of CD34-negative blasts. The enumeration of myeloid blasts is critical to the classification of MDS, and in this study was completed by quantification of blasts in aspirate smears. It can be assumed that histopathological features in UBTF-mutated disease classified as AML are similar to those outlined in this study for MDS-EB. Biological features rather than an arbitrary cutoff in blast count (of 20% or 30%) are important to differentiate the MDS from AML in young individuals. AML in the young is generally a chemotherapy-sensitive disorder, and most patients are cured by chemotherapy only. In contrast, survival with MDS-EB is extremely poor in the absence of allogeneic stem cell transplantation. Given the severe dysplastic features described above, the poor response to chemotherapy,^1-5 and a strikingly lower blast count in UBTF-mutated AML compared to AML with wildtype,^{4, 5} we propose that myeloid neoplasia with UBTF-TD either classified as MDS or AML may be considered one disease entity requiring a unified MDS-like therapy approach. Current next-generation sequencing (NGS) pipelines for AML and MDS may not include or detect UBTF-TDs, rendering the interpretation of the histopathology by experienced pathologists crucial for expanding standard diagnostic approaches to improve recognition of this distinct subgroup. Lastly, clinical studies providing a therapy algorithm with novel agents, like BCL-2 inhibitors¹⁹ or menin inhibitors,²⁰ to patients with UBTF-mutated myeloid neoplasia my pave the way for more rational and disease-specific therapy.

Conclusion

In this study we provide a comprehensive description of the diagnostic features observed in paediatric MDS with UBTF-TD. Our findings reveal distinct histopathologic characteristics that may serve as valuable clues for identifying UBTF-TD in affected patients. Notably, hyperplastic myelopoiesis displaying severe atypia, accompanied by a high frequency of micromegakaryocytes and monolobated cell forms, emerged as prominent histologic features. Furthermore, we observed an immature erythropoiesis with circulating normoblasts, which adds to the diagnostic profile of UBTF-TD. The identification of these specific histopathologic traits is of utmost importance, as it can prompt clinicians to conduct further molecular studies, leading to early and accurate diagnosis. Early detection of UBTF-TD is critical in guiding appropriate therapeutic strategies tailored to the individual patient's needs. Timely intervention may significantly impact patient outcomes, particularly considering the aggressive nature of UBTF-TD-associated MDS.

Author contributions

All authors read and approved the final article. Stephan Schwarz-Furlan, Carole Gengler, Ayami Yoshimi-Nöllke, and Franziska Schreiber analysed the data. Christian Thiede provided molecular data. Martina Rudelius, Miriam Erlacher, and Charlotte Niemeyer designed the study and edited the article. Guido Piontek, Yuki Schneider-Kimoto, and Markus Schmugge performed experiments and interpreted data. Martina Rudelius wrote the article.

Conflicts of interest

The authors declare no conflicts of interest.