In-Frame Exon 9 CALR Deletions Co-Occur with Other Alterations in the JAK-STAT Pathway in Myeloproliferative Neoplasms

Introduction: Mutations in the chaperone gene calreticulin (CALR) have been recently identified in essential thrombocythemia (ET) and primary myelofibrosis (PMF), and are essentially mutually exclusive with JAK2 or MPL mutations. Normal and mutant CALR proteins may differentially affect the subcellular trafficking of JAK-STAT signaling components. CALR mutations previously reported in ET and PMF have been +1 frameshift (fs) mutations localized to exon (E) 9 that generate a novel C-terminal protein sequence with a shift from acidic to basic residues. CALR E9 in-frame (IF) deletions have been recently rarely reported as polymorphisms such as TMP_ESP_19_13054686_13054688 and TMP_ESP_19_13054650_13054658 (Ensembl database entries). We sought to determine the frequency and associated clinical features of CALR with E9 IF alterations in samples submitted for suspicion of a myeloproliferative neoplasm (sMPN). We also assessed whether CALR IF alterations are differentially associated with +1fs mutations or with JAK2 V617For other somatic mutations in MPN-associated genes.

Materials and Methods: CALR mutation analysis of E9 was performed on genomic DNA extracted from blood, bone marrow (BM) aspirate or fixed BM biopsy sections using a Sanger sequencing assay with an analytic sensitivity of at least 15%. E9 IF cases were further assessed and mutations quantified by an Ion torrent sequencing panel assessing CALR, CSF3R, JAK2 and MPL, a second panel containing ASXL1, EZH2, IDH1, IDH2, KRAS, NRAS and TET2 and an Illumina MiSeq extended panel with 20 additional MPN-associated genes. These assays had a sensitivity of approximately 5%. JAK2 V617Fmutations were quantitated using a pyrosequencing assay with an analytic sensitivity of 1%.

Results: We assessed CALR E9 mutation status in 733 sMPN samples that were negative for JAK2 V617F mutation. 148 (20.1%) had typical +1fs mutations (95 type 1 and variants, 53 type 2 and variants); 2 (0.3%) had point mutations (E381A and D7373M); 7 (1.0%) had IF deletions including E381_A382>A, D397_D400>D (n =4), D400_K401>D and E405_V409>V. All E9 IF deletions were present at ~50% of reads. Clinical diagnoses were cytopenia/BM fibrosis, ET, thrombocytosis/anemia, and sMPN unspecified. Mutation analysis for 27 additional MPN-associated genes revealed mutations in 5/7 (71.4%) IF deletion cases including in MPL (W515L,40%; D163Y,12%), CSF3R (A470T 46%), ASXL1 (D954fs*26, 45%) and ZRSR2 (S449_R450dup, 27%). No additional mutations were found in the 2 cases with non-synonymous CALR point mutations/SNPs. In a parallel set of 76 MPN samples that had JAK2 V617F at varying levels, we noted 1 E9 IF deletion (D397_D400>D) in a sMPN case with 21.6% JAK2 V617F, and a typical +1fs mutation (K385fs*47) in a case with low (4.2%) JAK2 V617F. All other JAK2 V617F cases had no E9 CALR alterations.

Conclusions: CALR E9 in-frame deletions occur in up to 1% of sMPN samples and involve a variety of codons in the acidic domain. Therefore, sizing assays without DNA sequencing are not sufficient to unequivocally distinguish IF deletions from the characteristic +1 frameshift somatic mutations associated with ET and PMF. Given their level, these CALR IF deletions are likely germline sequence variants but are associated with a high frequency of somatic mutations in other MPN-associated genes but not with CALR +1fs mutations. Their co-occurrence with pathogenic somatic mutations in JAK2, MPL and CSF3R affecting the JAK-STAT pathway raises the possibility for a contributory role of altered CALR proteins produced by these E9 deletions in the pathogenesis of MPN.