Two cases of an imatinib-sensitive PRKG2::PDGFRA-positive chronic myeloid neoplasm with basophilia

The clinical and morphological phenotype of myeloid/lymphoid neoplasms with tyrosine kinase (TK) gene fusions (MLN-TK) is influenced by both the TK and the partner gene [ 1 ]. Here, we report on two patients with an initial diagnosis of a chronic myeloid neoplasm with prominent basophilia, absence of eosinophilia and elevated serum tryptase levels in association with myelofibrosis. Patient #1 : A 59-year old female presented with thrombocytosis of up to 3500 × 10 9 /L, basophilia (1.1 × 10 9 /L) in the absence of eosinophilia or monocytosis, and marked splenomegaly (length 25 cm by ultrasound). The serum tryptase was increased at 30 µg/L (normal value < 11.4). The bone marrow (BM) was normocellular with increased and dysplastic megakaryocytes without myelofibrosis and no increase in basophils, mast cells (MC) or blasts. Genetic analyses were negative for BCR :: ABL1 and mutations in JAK2 , CALR , or MPL . The patient was diagnosed as triple-negative essential thrombocythemia (ET) with basophilia. Cytoreductive treatment consisted of hydroxyurea and anagrelide. Three years later, the patient presented with weight loss (−6 kg), night sweats, fatigue, and worsening splenomegaly. At referral, leukocytes were at 22 × 10 9 /L with basophilia (4.6 × 10 9 /L, 21%), and 8% blasts. Neutrophils, monocytes and eosinophils were normal. Hemoglobin was 9.8 g/dL and platelets 1,200 × 10 9 /L. Lactate dehydrogenase (LDH) and serum tryptase levels were elevated at 930 U/L and 45 µg/L, respectively. BM histology revealed a myeloproliferative neoplasm (MPN)-like picture, with increased basophils, interstitial MC, and myelofibrosis grade 2-3 (Fig. S1 ). A diagnosis of post-ET-myelofibrosis was made. The karyotype was normal; KIT D816V was not detected. A 63-gene myeloid panel did not identify any pathogenic or likely pathogenic variants. (Table S2 ). Patient #2 : A 44-year old male presented with early satiety, abdominal pain, and unintentional weight loss (−4 kg) over six months. Blood counts revealed leukocytosis (23.9 × 10 9 /L), mild anemia (hemoglobin 12.3 g/dL), and thrombocytopenia (114 × 10 9 /L). The differential showed basophilia (absolute 15.5 × 10 9 /L, 65%), while neutrophil, monocyte, and eosinophil counts were normal. Serum tryptase was elevated (125 µg/L), LDH was normal (210 U/L). The spleen (25 cm) was enlarged. Treatment with hydroxyurea was initiated and the patient was referred for further evaluation. A BM biopsy showed markedly increased basophils, interstitial MC and myelofibrosis grade 3 (Fig. S1 ). The karyotype was normal. BCR :: ABL1 was negative and no mutation in JAK2 or KIT was identified (Table S1 ). A 63-gene myeloid panel identified a pathogenic nonsense mutation in TET2 (c.2368 C > T; p.Gln790*; VAF 41%) and a pathogenic truncating mutation in IKZF1 (c.1004dup; p.Gly337Argfs152; VAF 41%). Both patients were included in the SIRIUS project to comprehensively analyze genomic events that may underlie the clinical conditions by whole genome sequencing (WGS) and whole transcriptome sequencing (WTS) [ 2 ]. WGS-based structural variant analysis, together with adaptive nanopore sequencing [ 3 ], identified an inversion on chromosome 4 with genomic breakpoints in exon 12 (patient #1) and intron 11 (patient #2) of PDGFRA (4q12) and in intron 3 (patient #1) and intron 10 (patient #2) of PRKG2 (4q21), both consistent with an in-frame PRKG2 :: PDGFRA fusion gene (Tables S2 – S5 ). Subsequent FISH analyses on interphase nuclei and metaphases confirmed rearrangement of PDGFRA and the inversion in the long arm of chromosome 4 [inv(4)(q12q21)], which is cytogenetically cryptic by conventional karyotyping due to the small size and similar G-banding pattern of the inverted chromosomal fragment (Fig. 1 ). Fig. 1: Detection and genomic characterization of the fusion genes. Full size image A Chromosome 4 (ideogram with G-banding pattern) with the genes PDGFRA (4q12) and PRKG2 (4q21) that are involved in an inversion. B FISH on interphase nuclei (patient #1) and on metaphases (patient #2) with a PDGFRA break-apart probe (green: 5’region of PDGFRA , red: 3’region of PDGFRA ) showing a PDGFRA rearrangement. C Illustration of genomic breakpoints (red arrows) in exon 12 (patient #1) and intron 11 (patient #2) of PDGFRA and in intron 3 (patient #1) and intron 10 (patient #2) of PRKG2 involved in an inversion on chromosome 4 identified by WGS ( PDGFRA : ENST00000257290.10; PRKG2 : ENST00000264399.6; reference genome: GRCh37/hg19), and resulting PRKG2::PDGFRA fusion transcripts detected by WTS. Abbreviations: FISH, fluorescence in situ hybridization; WGS, whole-genome sequencing; WTS, whole-transcriptome sequencing. Initial RNA fusion calling (Arriba, STAR-Fusion, FusionCatcher and Manta) did not detect a PDGFRA fusion transcript, probably because of low-level expression, but re-analysis of the short-read WTS data using the DRAGEN fusion pipeline (Illumina, San Diego, CA, USA) identified PRKG2 :: PDGFRA in both cases, albeit with low read support. The fusion transcripts consisted of exons 1 to 3 (patient #1) and 1 to 10 (patient #2) of PRKG2 and truncated exons 12 to 23 of PDGFRA encompassing the tyrosine kinase domain (Fig. 1 ). Predicted mRNA fusions from the DNA junction sequences were used to design patient-specific RT-PCR assays that enabled detection of distinct in-frame PRKG2 :: PDGFRA mRNA fusions in both patients. Corresponding quantitative RT-PCR assays were established to monitor molecular response to treatment (Supplement). Treatment with imatinib was initiated at variable doses (100–400 mg daily) with a long-term maintenance dose of 100 mg daily in both patients. Within weeks, complete blood counts, differentials, and serum tryptase levels returned to normal values. In patient #1, a BM biopsy at month 6 demonstrated complete resolution of basophils, MC, and reticulin fibrosis. In patient #2, no follow-up BM biopsies were performed. During imatinib treatment, FISH for PDGFRA rearrangement and RT-PCR for the PRKG2 :: PDGFRA fusion became negative. In patient #2, variant allele fractions of the TET2 and IKZF1 mutations declined in parallel with PRKG2 :: PDGFRA transcript levels, consistent with clearance of the pathogenic clone below the assay’s limit of detection (Fig. 2 ). Fig. 2: Clinical and genetic response during treatment. Full size image Treatment periods are indicated above each case. Abbreviations: FISH fluorescence in situ hybridization, qPCR quantitative polymerase chain reaction, VAF variant allele frequency. The two cases highlight the challenging decision-making of when and how to pursue detection of pathogenetically relevant somatic variants or fusion genes in morphologically distinct myeloid neoplasms when first-line screening, such as conventional karyotyping, targeted FISH/PCR assays, or routine myeloid DNA panels, remains non-informative. In routine practice, this scenario is often encountered in the workup of chronic myeloid neoplasms that test negative for BCR :: ABL1 or MPN driver mutations, or those with eosinophilia that test negative for FIP1L1 :: PDGFRA . Conventional karyotyping is a screening tool with limited resolution and may therefore miss cryptic rearrangements. If a fusion gene is suspected, the underlying rearrangement should be confirmed by locus-specific FISH and, ideally, by RT-PCR. However, RT-PCR is largely restricted to known fusion transcripts, whereas routinely used break-apart FISH probes can identify the rearranged tyrosine kinase but generally fail to identify the fusion partner. Importantly, the clinical and morphologic phenotype is determined not only by the tyrosine kinase involved but also by the fusion partner, as well as by the frequency and pattern of additional somatic mutations [ 1 ]. Distinct morphological patterns, such as eosinophilia, monocytosis, or basophilia, can guide the selection of more comprehensive genetic testing, including specialized PCR-based approaches (historically, RACE- or Bubble-PCR), WTS, long-read nanopore DNA sequencing, or WGS [ 3 ]. In MLN-TK, a significant minority of the >100 currently described fusion genes, most commonly involving PDGFRA , PDGFRB , FGFR1 , JAK2 , ABL1 , or FLT3 , are cytogenetically cryptic due to small deletions, inversions, or insertions [ 1 ]. Overreliance on single morphologic parameters can be misleading: eosinophilia is present in >90% of patients primarily in MLN- FIP1L1 :: PDGFRA and MLN- ETV6 :: ABL1 , whereas it is absent in >40–50% of MLN- FGFR1 cases [ 1 ]. In patients with a clinical picture of chronic myelomonocytic leukemia, who test negative for typical mutations by myeloid panel, the concurrent presence of eosinophilia >0.5 × 10 9 /L may point towards underlying fusion genes such as FIP1L1 :: PDGFRA or ETV6 :: ABL1 [ 1 ]. The combination of monocytosis and eosinophilia is also recurrently observed in KIT D816V-positive systemic mastocytosis, usually of the advanced subtype. Irrespective of the involved tyrosine kinase, BCR as a fusion partner of PDGFRA, PDGFRB, FGFR1, JAK2, or ABL1 appears to confer a chronic myeloid leukemia (CML)-like phenotype without or only mild eosinophilia. Basophilia is a recurrent finding in BCR :: ABL1- positive CML and is incorporated as a prognostic variable in established risk models. Careful attention should be paid to the differentiation between basophils and circulating MC, which may occur in patients with mast cell leukemia. MYB :: GATA1 has been reported in myeloid neoplasms with basophilia [ 4 ], whereas in contrast to previous reports, more recent studies indicate that basophilia is not a consistent feature of patients with DEK :: NUP214 fusions [ 5 ]. To date, only two patients (1 male and 1 with undisclosed sex) with PRKG2 :: PDGFRA fusions have been reported in the literature, with detailed clinical annotation available for one patient, who presented with MPN-NOS with marked basophilia, elevated serum tryptase, and a rapid response to imatinib [ 6 , 7 ]. PRKG2 has also been described as a fusion partner of PDGFRB , and the eight reported patients (6 males, 2 females) shared a phenotype characterized by basophilia and increased MC with absent or only mild eosinophilia [ 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 ]. PRKG2 encodes the membrane-associated cGMP-dependent serine/threonine kinase PKGII, a key regulator of intestinal secretion (including CFTR activation/trafficking) and bone growth. The coiled-coil domain of PKGII fused to the tyrosine kinase domain of PDGFRA could lead to dimerization and constitutive activation of the fusion protein, as has been assumed for PKGII::PDGFRβ [ 8 , 10 ]. Collectively, these observations suggest that PRKG2, as a fusion partner, shapes the phenotype toward a chronic myeloid neoplasm characterized by prominent basophilia. These cases highlight the value of genomic analysis to uncover cryptic, targetable kinase fusions when routine testing is non-informative. Imatinib led to rapid and durable remission of MLN- PRKG2 :: PDGFRA .