Introduction

The PICALM::MLLT10 (also known as CALM::AF10) fusion transcript has first been discovered in the U-937 histiocytic lymphoma cell line [1, 2]. It is the product of a translocation of chromosomes 10 and 11 (t(10;11)(p12.3;q14.2)). MLLT10-rearranged acute myeloid leukemia (AML) is uncommon. It most frequently occurs as KMT2A::MLLT10-fusion, which one study estimates as the third most prevalent KMT2A-rearrangement, being present in 19% of KMT2A-rearranged AML [3]. In comparison, PICALM::MLLT10 rearranged AML is extremely rare, with the largest published cohort in the adult setting only including 18 cases [4, 5], and the largest pediatric cohort including 39 patients [6]. This is in contrast to MPAL, where the fusion seems more frequent [7], and to gamma-delta-T-cell acute lymphoblastic leukemia (ALL) where the fusion also occurs more frequently [8]. Given its low incidence in AML, the diagnosis can be challenging, as is evident in the case presented here.

It is not entirely clear how PICALM::MLLT10 leads to leukemogenesis. It has been shown that the interaction of the clathrin-binding-domain of PICALM and the OM-LZ region of MLLT10 is the relevant outcome of the fusion [9]. It has also been shown that several genes, including BMI1 posterior HOXA genes, and MEIS1 are overexpressed in PICALM::MLLT10-positive acute leukemias [10, 11]. However, there are two proposed downstream mechanisms. One focuses on the interaction of a nuclear export signal contained within PICALM and the DOT1L [9, 12, 13]. This prevents nuclear export of the fusion transcript, and facilitates PICALM-dependent upregulation of HOXA-family genes (HOXA 5, 9, 10, MEIS1), which in turn blocks differentiation of hematopoietic stem cells (HSCs) [14]. The other hypothesis focuses on BMI1, which encodes a ring finger protein that is part of the Polycomb Repressive Complex 1 (PRC1). Interestingly, it is not upregulated in KMT2A-rearranged AML. The gene is located downstream of AF10. PRC1 stabilizes long-term silencing of regions targeted by Polycomb Repressive Complex 2 (PRC2), which represses multiple tumor suppressor genes (e.g., CDKN2A and CDKN2D). Thus, its upregulation is oncogenic. It has been shown that inhibition of BMI1 is capable of both inhibiting PICALM::MLLT10-driven leukemic transformation and impairing the growth of PICALM::MLLT10 rearranged AML in several models [15].

We conclude that the case report and literature review presented here will aid clinicians in recognizing this rare entity and understanding its disease biology.

Case report

A 30-year-old man presented to the Emergency Department of a local hospital 4 times in 11 days with increasing thoracic pain. Due to unspecific ST-elevations on ECG, pericarditis was diagnosed. He was repeatedly discharged with anti-inflammatory and analgesic drugs, until he returned with nausea, myalgias, fever, chills, and sweating. A computed tomography scan was performed (Fig. 1), which revealed a mediastinal mass of 6.5 × 4 × 7 cm, and putative pyelonephritis. The patient was admitted under the suspicion of sepsis. Laboratory studies were compatible with disseminated intravascular coagulation (DIC;Tc: 9 G/l, D-Dimer: >128.000 µg/l, PT: 21 s, Fibrinogen: 2.1 g/l). immune-thrombopenia (ITP) and paraneoplastic DIC were considered, and the patient was transferred to our academic hospital.

Fig. 1
figure 1

Axial image from the CT scan taken at diagnosis showing the largest diameter of the mediastinal myelosarcoma.

Full size image

At our center, petechiae on abdomen and legs were noted, and the patient reported bleeding easily when brushing his teeth. Neither lymphadenopathy nor organomegalies were present. There was no palpable scrotal mass. An ultrasound of the urogenital tract did not reveal a germ cell tumor. Laboratory studies were notable for massive elevation of acute-phase markers (Table 1).

Table 1 Laboratory studies of the clinical case.
Full size table

A blood smear revealed blasts which were morphologically judged to be of primarily lymphatic aspect with immature nucleus and multiple nucleoli. A left-shift of the myeloid lineage was noted. No fragmentocytes were observed. In contrast to the primary cytomorphologic aspect, immunophenotyping of the peripheral blood revealed a large population of myeloblasts (strongly positive for MPO, positive for CD15, CD33, CD34, CD38, CD71, CD105, and CD123), partially positive for CD117, HLA-DR, negative for CD13). A bone marrow biopsy confirmed the diagnosis of AML (FAB M1). The patient received all-trans retinoic acid and arsenic trioxide until PML::RARα-rearrangement was ruled out. Cytogenetics revealed t(10;11)(p12.3;q14.2). Interphase fluorescence in-situ hybridization (FISH) and myeloid gene panel (Illumina TSO500 panel) sequencing confirmed a PICALM::MLLT10 fusion transcript (variant allele frequency (VAF): 38%) and found two mutations in EZH2 (c.2069 G > A, p.R690H, VAF 36%, ClinVar: likely pathogenic; c.2084 C > T, p.S695L, VAF 36%, ClinVar: pathogenic/likely pathogenic), but no other mutations or rearrangements (65 genes and 688 rearrangements analyzed). Though not definitively deductible from our sequencing data, the presence of two separate EZH2mut could hint at a biallelic loss of function. Both EZH2 mutations we have detected are located in the SET-domain of EZH2. According to the Catalog of Somatic Mutations in Cancer (COSMIC) database, one of these mutations (p.S695L) has previously been described in AML, myelodysplastic neoplasm (MDS), ALL, all well as solid tumors, while the other (p.R690H) has been described in a number of hematologic tumors, including AML, MDS, myelofibrosis, Chronic myeloid leukemia, and ALL. Missense mutations in this region abrogate catalytic activity in vitro [16]. Both mutations in the SET-domain as well as homozygous aberrations of EZH2 show a trend towards worse survival in one large study [17]. Conventional cytogenetics revealed no aberration besides the reciprocal fusion, with t(10;11)(p12.3;q14.2) being detectable in 18/20 metaphases.

The patient then underwent intensive induction chemotherapy with cytarabine and idarubicin (“7 + 3” regimen). Bone marrow biopsy after two cycles showed complete morphological remission, as well as flow-cytometric and molecular measurable residual disease (MRD) negativity, and the remnant of the strongly regressive mediastinal mass was not 18-fluor-desoxy-glucose-avid on positron emission tomography (PET)-CT. We thus considered it most likely that the mass was a myelosarcoma of the present AML and refrained from a biopsy.

The patient then underwent allogeneic stem cell transplant from his matching sister after myeloablative conditioning with cyclophosphamide and busulfan. He received cyclosporine A (CyA) and methotrexate (MTX) as graft versus host disease (GVHD)-prophylaxis. Eleven months after allogeneic stem cell allogeneic transplant, molecular relapse became apparent with an increasing PICALM::MLLT10 fusion transcript. However, under salvage therapy with azacitidine and venetoclax (aza/ven), the fusion transcript disappeared below the limit of detection (VAF of 2%) again. Unfortunately, seven months after initiation of aza/ven an extramedullary relapse in the form of a progression of the previously described mediastinal mass occurred, while the bone marrow remained MRD negative. Video-assisted thoracoscopic surgery was performed to take a biopsy of the mass, which confirmed the presence of a myelosarcoma, the known PICALM::MLLT10 fusion and the EZH2-mutations. However, in addition mutations in BRAF (G469A; VAF 43%), NRAS (G12S; VAF 44%) and TP53 (L145R; VAF 57%) were detected in the mediastinal mass. Activating mutations in BRAF are rare in AML and are clearly associated with adverse outcomes [18, 19]. After subsequent radiation of the mediastinal mass (12 × 2 Gray), further extramedullary manifestations of the AML appeared in the patient’s pancreas, ribs, and skull. He underwent a second allogeneic stem cell transplant from the same donor (conditioning: Fludarabine and total body irradiation (8 Gray); GVHD-prophylaxis: CyA, MTX) five months after discontinuation of aza/ven. Subsequently, a maintenance therapy with the BRAF-Inhibitor darafenib and the MEK-Inhibitor trametinib was initiated. This therapy is the standard of care in malignant melanoma with activating BRAF-mutations [20]. In AML, BRAF-inhibition is not established, but case reports have also shown responses [21]. Unfortunately, the patient did not achieve a long-term remission and died four months after the second transplant.

Literature review

We searched PubMed for possible spellings of the translocation of interest (“PICALM-MLLT10”, “CALM-AF10”, “PICALM::MLLT10”, “CALM::AF10”) alone and in combination with “AML” or “Acute myeloid leukemia”. Papers including clinical characterizations of patients diagnosed as AML by the respective authors were included. Papers for which no English-language full-text was available were excluded. We identified 118 cases of PICALM::MLLT10 rearranged AML published in case reports or case series [4, 6, 22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44]. Relevant clinical, immunophenotypic and genetic characteristics of the reported cases are summarized in Table 2. The reported patients are typically younger than the average AML patient with a median age of 14 years at diagnosis (range 6-66). There is a trend towards male sex (71/118). Extramedullary disease is common, with 35 cases reporting some extramedullary involvement. Of these, at least 25 were myelosarcomas (10 mediastinal, 2 abdominal, 1 mammary, 1 uterine, 11 unknown location). However, the type of extramedullary disease was not always specified. Thus, the young, male patient with extensive extramedullary involvement we present here has quite typical clinical features of PICALM::MLLT10-rearranged AML.

Table 2 Clinical characteristics, karyotypes and immunophenotypes of all adult/adolescent cases with PICALM::MLLT10 and AML published in the literature.
Full size table

Karyotypes were available for 63 cases. There was a median of 1 cytogenetic abnormality (range: 0–15) besides PICALM::MLLT10. 29 cases had cytogenetic aberrations associated with a poor prognosis, while only one each had an aberration that is associated with favorable/intermediate outcome according to the European Leukemia Net risk stratification from 2022. Eight had trisomy 4 and eight had trisomy 19. In seven patients i17(q10) co-occurred, while no other cytogenetic aberration was co-occurring in more than three cases. Immunophenotypic markers are available for 69 cases. Markers commonly associated with AML were widely reported, while CD7 was the most frequent aberrant marker (27/34 cases). Unfortunately, molecular genetic analyses are not available for the majority of previously reported cases. However, following the introduction of next generation sequencing analysis in increasing numbers of AL patients, newer series make it apparent that PICALM::MLLT10 has a characteristic comutational profile. The most frequent co-mutations occurred in PHF6 (22), TP53 (12), NF1, NRAS, SUZ12 (8 each), WT1 (6) and EZH2 (5). As our patient also had two EZH2 mutations, we investigated these co-occurrences further. In a series of 15 AL (12 T-ALL, 1 MPAL, 2 AML) patients with PICALM::MLLT10 rearrangement 4 (including one AML) cases carried mutations in EZH2, which usually is very rare in AL, with one report finding one mutation amongst 113 AML cases and another finding no EZH2mut in 54 AML patients [16, 29, 45]. Notably, four more patients in the series (including the other reported AML) carried a polymorphism of EZH2. Another analysis of PICALM::MLLT10-rearranged AL found EZH2mut in 6/20 cases (ALAL, T-ALL and AML) [35]. Furthermore, one analysis of MLLT10-fusions in pediatric AML found another three patients with the co-occurrence of PICALM::MLLT10 and EZH2 mutations [24].

Discussion

The literature to date suggests, that PICALM::MLLT10 rearranged AML is associated with unusually young age, male sex, and (mediastinal) myelosarcoma, which our case confirms. Unfortunately, our patient seems to confirm the previously reported trend towards adverse outcome. A recent large retrospective analysis also confirms the negative prognosis in the pediatric population [6]. A large analysis suggests that MLLT10, regardless of the fusion partner, always seems to convey adverse outcomes [5].

A significant percentage of published cases (23/63) had a complex karyotype. However, it is unclear whether the genomic instability required to produce these karyotypes is cause or effect of the PICALM::MLLT10-rearrangement. While the high incidence of trisomies 4 and 19 that was first noted by Borel et al. [4] is confirmed by other published cases, our patient did not have such aberrations.

As described in the literature, the occurrence of this translocation not only in AML, but as well in both ALL and MPAL points towards its occurrence in a pluripotent progenitor cell compartment [32, 46]. The descendants of this progenitor might then develop into different ALs depending on the differentiation they undergo and secondary alterations accumulated. This could also explain the high latency of AL development in a transgenic murine model [47]. Research in this area is ongoing, with the current WHO classification of myeloid disease hinting at a potential re-definition of PICALM:: MLLT10 as a MPAL-defining genetic aberration in the next update of the classification [7].

While better genetic characterization of this entity is needed, a pattern of typical co-mutations seems to emerge from our literature analysis. Our case-report matches the published co-mutations well, as the mutations in EZH2, NRAS and TP53 all seem frequent in this population. However, NRAS and TP53 are also frequently mutated in other AML [48]. EZH2 on the other hand seems more closely associated with PICALM::MLLT10. EZH2 is the catalytic subunit of the Polycomb Repressor Complex 2 (PRC2), which recruits PRC1 via methylation markers. Together, they mediate long term epigenetic silencing of chromatin. Studies in hepatocellular carcinoma, esophageal squamous cell carcinoma and prostate cancer have shown a synergistic effect of BMI1- and EZH2-upregulation, while the interaction in breast cancer seems more complex [49,50,51,52,53,54,55]. Therefore, we hypothesize that, if BMI1-upregulation is the relevant mechanism of leukemogenesis in PICALM::MLLT10 positive AL, acquiring an EZH2 mutation constitutes a crucial survival benefit, leading to selection of EZH2 mutated clones. This hypothesis would explain the strikingly elevated frequency of EZH2 mutations in PICALM::MLLT10 positive AML. Multiple large cohorts of AML patients (Basheer et al.: n = 2434; Papapemmanuil et al.: n = 1540; Rausch et al.: n = 1138; Stasik et al.: n = 1604; Wang et al.: n = 714), have found EZH2mut in 5% (Basheer), 4% (Papaemmanuil, Rausch, Stasik), and 2% of AML cases, respectively, and some smaller studies show barely any cases (1/113, 0/54) [16, 17, 45, 56,57,58,59]. EZH2mut are associated with worse OS, with one large study observing a trend towards even worse outcomes for homozygous mutations and mutations of the SET-domain [17]. In contrast to the low incidence of EZH2mut observed in large unselected cohorts of AML, we and others have shown frequent occurrence of EZH2mut in PICALM::MLLT10 positive AML [24, 29, 35]. However, the potential association between PICALM::MLLT10 rearrangements and EZH2mut is based on small samples and further studies are needed.

In addition, our hypothesis strengthens the case for BMI1-mediated leukemogenesis. Clinically, the possibility of targeting both BMI1 and EZH2 therapeutically makes this finding especially relevant [15]. In follicular lymphoma, the EZH2-Inhibitor tazemetostat is already approved by the FDA, proving the safety of targeting EZH2 in humans [60]. In AML EZH2-inhibitors have already shown promise in PDX-mouse-models [61].

In conclusion, this report not only highlights and confirms the clinical characteristics of PICAL::MLLT10mut AML, but also elucidates key molecular mechanisms as well as potential targets for eventual therapeutic intervention.