PAX5 P80R-mutated B-cell acute lymphoblastic leukemia with transformation to histiocytic sarcoma: clonal evolution assessment using NGS-based immunoglobulin clonality and mutation analysis

A 60-year-old male patient presented with fatigue. Analysis of the peripheral blood showed pancytopenia and 40% of blasts. Subsequent analysis of a bone marrow biopsy and aspirate resulted in a diagnosis of BCR-ABL-negative B-cell acute lymphoblastic leukemia (B-ALL, for immunophenotype and karyotype, see supplementary Tables 1 and 2), with absence of extramedullary and central nervous system localization. The patient started intensive chemotherapy and rapidly reached complete remission with no detectable minimal residual disease (MRD) by cytomorphology and immunophenotyping. However, 10 months after the initial diagnosis of B-ALL and during consolidation courses, the patient presented with recurrent thrombocytopenia, progressive fatigue and a skin lesion on the forehead. A positron emission tomography (PET) scan showed generalized disease with foci of increased FDG uptake in bone, lymph nodes, the liver, the skin, and the lungs. Morphology and immunohistochemistry of biopsies obtained from the bone marrow, skin, and a supraclavicular lymph node showed an infiltration of rather large, pleomorphic malignant cells (Fig. 1A) with a high proliferation (Ki-67: 80%) and histiocytic differentiation (CD14⁺, CD68⁺, and CD163⁺, supplementary Table 3), consistent with a diagnosis of histiocytic sarcoma (HS). The biopsies, peripheral blood, and bone marrow aspirate did not demonstrate any evidence of B-ALL relapse. To determine whether the two malignancies in this patient were clonally related, we performed clonality assessment by using GeneScan and NGS-based immunoglobulin (IG) clonality analysis and mutation analysis.

To determine whether the B-ALL and HS had a common clonal origin, we performed immunoglobulin (IG) clonality analysis on a bone marrow sample at the time of B-ALL diagnosis and a skin biopsy with HS infiltration. With conventional clonality analysis using GeneScan fragment length analysis according to the EuroClonality/BIOMED-2 assay[1], IGHV-IGHD-IGHJ rearrangement analysis showed similarly sized products for framework (FR)1, FR2, and FR3 targets in both the B-ALL and HS sample, but also products that were unique for the B-ALL or HS sample were observed (Table 1, Supplementary Fig. 1). For IGHD-IGHJ, IGKV-IGKJ, and IGKV/Intron-KDE targets, only clonal products unique for one of the two samples could be detected. This analysis was therefore not firmly conclusive to establish whether the B-ALL and HS were clonally related, since only one IGHV-IGHD-IGHJ gene rearrangement (as convincingly detected in the FR1 and FR2 PCRs) showed an identically sized fragment, which could be a coincidence. In addition, the results of the B-ALL sample suggested the presence of more than one clone based on the number of IGH gene rearrangements.

To more reliably determine the clonal relationship between the B-ALL and HS in this patient, we performed NGS-based clonality analysis, according to the protocol that was recently published by the EuroClonality-NGS Working Group [2, 4]. With NGS-based IG clonality analysis, two clonal IGHV-IGHD-IGHJ rearrangements (clonotypes V6-1 -0/39/-5 J6 and V3-19 -0/44/-2 J6) and one clonal IGKV-IGKJ rearrangement (V1(D)-33 -11/2/-7 J3) were detected in both the B-ALL and HS with 100% identical sequences, providing evidence for a direct clonal relationship (Table 1, Supplementary Fig. 2). Nonetheless, additional IG rearrangements unique for each sample were also detected. Detailed analysis of the sequences of the five IGH rearrangements in the B-ALL sample revealed that four of the rearrangements, i.e., one incomplete IGHD-IGHJ and three complete IGHV-IGHD-IGHJ rearrangements, contained the exact same D and J genes, including an identical N2 junction (D-J stem: D2-2 -2/7/-5 J6). Two different V genes were combined with this IGHD-IGHJ stem, resulting in three different clonotypes (V6-1 -1/37/-5 J6, V6-1 -0/39/-5 J6 and V3-7 -1/34/-5 J6) (Fig. 1B). This suggests that (at least) three subclones in the B-ALL arose from the same IGHD-IGHJ stem. Based on the rearrangements detected in the HS sample, only one of the subclones evolved into the HS (Fig. 1C).

To further study the pattern of clonal evolution from the B-ALL (sub)clone to HS, molecular analysis was performed with the TruSight Oncology 500 assay (Illumina), which identified multiple mutations and several copy number variations in both the B-ALL and HS sample (Supplementary Tables 4 and 5). A pathogenic mutation in the PAX5 gene (resulting in the p.P80R change) was detected indicating that the B-ALL belongs to the newly recognized subtype of B lymphoblastic leukemia with PAX5 P80R [5, 6]. In addition, mono-allelic loss of PAX5 was observed, leading to bi-allelic inactivation of the PAX5 gene. These PAX5 alterations, together with a mutation in KRAS (p.G12D) and loss of CDKN2A, were identified in both the B-ALL and HS tissue, which supported the clonal relationship. An additional, likely oncogenic RAF1 (p.R391W) mutation, was detected in the HS sample only [7].

Lineage switch of a lymphoid malignancy to HS is rare, but has been well documented in previous case reports and in small case series [8‐11]. A clonal relationship between the lymphoid malignancy and the HS was established in a subset of cases by the detection of identical cytogenetic abnormalities or rearrangements of the immunoglobulin or T-cell receptor genes. In our case, conventional clonality analysis yielded ambiguous results, but NGS-based IG clonality analysis was able to confirm the clonal relationship between the B-ALL and the HS. In addition, using the sequence information of the NGS-based IG clonality analysis, multiple related subclones could be distinguished in the B-ALL, and the clonal evolution to HS could be unraveled (Fig. 1C). The clonal relationship was further supported by the presence of identical somatic mutations in PAX5 and KRAS.

The reason why some B-ALL patients develop a histiocytic malignancy is not clear, yet. Several studies show that loss of PAX5 expression in B-cells leads to dedifferentiation to uncommitted precursor cells [12, 13]. Interestingly, the present B-ALL case displayed inactivated PAX5 due to a mutation combined with loss of the other PAX5 allele, which will result in a lack of PAX5 expression. This likely makes these cells prone to loss of their B-cell phenotype and differentiate into another lineage after acquiring one or more additional genetic aberrations. Furthermore, specific aberrations in B-ALL, including PAX5-P80R, were recently found to predispose to an early monocytic phenotype switch in pediatric patients during initial chemotherapy [14]. The monocytic switch was accompanied by a gradual loss of CD19 expression and increase in expression of at least one monocytic marker. This observation also shows that PAX5-mutated cells are able to rather easily switch their phenotype. In this pediatric PAX5-P80R-mutated cohort, no subsequent malignancies, like histiocytic sarcoma, were described during follow-up.

In our adult case, the B-ALL cells harbored a PAX5 mutation and subclonal KRAS, NRAS, PTPN11, and CDKN2A mutations. In addition, mono-allelic loss of PAX5 and CDKN2A was detected. All these aberrations are frequently observed in PAX5-mutated B-ALL patients. Passet et al. [5] described a cohort of 30 PAX5-P80R-mutated adult B-ALL patients which all showed inactivation of the second PAX5 allele (mutation or loss) and the majority harbored CDKN2A loss (74%) and an NRAS and/or KRAS mutation (73%). All patients achieved complete remission after treatment. Monocytic shift or subsequent histiocytic malignancies were not reported in this paper.

Interestingly, the HS cells of our patient harbored the same PAX5 and KRAS mutation as the B-ALL cells, with an additional RAF1 mutation. The combinaton of multiple mutations effecting the RAS pathway has been reported before in HS, including the combination of a KRAS with a RAF1 mutation [9, 15]. Also in one of these previously described cases, the RAF1 mutation was present only in the HS sample and not in the concomitant chronic myelomonocytic leukemia (CMML), whereas both harbored the same KRAS mutation [9]. These observations suggest that RAF1 activation might play a role in the final transformation to HS in B-cells that are already prone to loss of their B-cell phenotype.

In conclusion, we report a case of PAX5 P80R-mutated B-ALL followed by histiocytic sarcoma in which combined NGS-based IG clonality and mutational analysis elucidated the clonal relationship and evolution. A combination of genetic events, including PAX5 inactivation in combination with an acquired RAF1 mutation, may have played a role in the lineage switch of the malignant cells and subsequent development of HS. Our study demonstrates that NGS-based mutation and IG clonality analysis provide clear added value for clonal comparison, which helps to unravel the underlying clinicopathological mechanisms of disease evolution.