AUTHORS
Danielle Rachel Isakov, PhD, Breana Sinkfield-Tyler2, Tsega-Ab Abera3, Olivia Sakaguchi4, Ronald Hoffman, MD5, Anna Nam, MD6
AFFILIATIONS
1 Tri-Institutional MD-PhD Program, New York, New York, United States, Weill Cornell Medicine, New York, New York, United States
2 Weill Cornell Medicine, New York, New York, United States
3 Weill Cornell Medicine, New York, New York, United States
4 Weill Cornell Medicine, New York, New York, United States
5 Icahn School of Medicine at Mount Sinai, New York, New York, United States, Myeloproliferative Neoplasms Research Consortium, New York, New York, United States
6 Sandra and Edward Meyer Cancer Center, New York, United States, Weill Cornell Medicine, New York, New York, United States
ABSTRACT
Calreticulin (CALR) mutations drive 20% of myeloproliferative neoplasms (MPN). The two most common variants, a 52-bp deletion (Type 1) and a 5-bp insertion (Type 2), cause frameshifts in the last exon (exon 9), generating novel C-termini that aberrantly activate the thrombopoietin receptor MPL. Other pathogenic frameshift variants cause similar neo-morphic variants and downstream events. To explore the possibility of eliminating the mutated exon 9 as a therapeutic approach, we studied an unusual CALR frameshift variant associated with a novel isoform that skipped the mutated exon 9.
We recently studied blood stem and progenitor cells from patients with CALR-mutated MPN, who were treated with interferon-alpha (IFNa, Lama*, Isakov*, et al, Nat Genet, in press). We performed single-cell multi-omics that co-detects somatic mutations and single-cell RNA-seq (Genotyping of Transcriptomes, GoT) of bone marrow CD34+ stem and progenitor cells obtained at baseline and following IFNa therapy for at least one year. Within this cohort, we identified a patient with essential thrombocythemia caused by a 31-base pair deletion in CALR exon 9, who experienced a remarkable molecular response with IFNa therapy (variant allele frequency (VAF) 0.49 to VAF 0.04 at year 1 of therapy). GoT confirmed the depletion of the CALR mutant clone in hematopoietic stem cells (HSC). Examination of the genotyping reads revealed the presence of a novel isoform that skipped exon 9 and a portion of 3’UTR, suggesting that the loss of the pathogenic exon 9 resulted in clonal depletion with IFN treatment.
To determine whether this rare 31-bp deletion variant in CALR may induce a novel exon 9 skipping event, we modeled the 31-bp deletion in a splicing reporter construct spanning CALR exons 8 and 9 with their respective introns, as well as the 3’ UTR. We then transfected HEK293T cells with CALRWT, CALR31del, CALRType1, CALRType2, or control vectors. The CALR31del alone was sufficient to produce the expected splice product that matched the isoform identified in the original patient, whereas CALRWT, CALRType1, CALRType2, and control constructs did not.
To further characterize the alternative splicing activity, we designed 12 antisense-oligonucleotides (ASO) by tiling the region spanning the splice sites. We then transfected HEK293T cells with individual ASOs and a CALRWT, CALR31del, or Type 1 CALR or Type 2 CALR pathogenic variant vectors. Although the ASOs decreased CALR transcripts overall, they induced the most significant reduction of the exon 9 splice transcript in the CALR31del condition (p < 0.01). This result led us to hypothesize that the splice event may be contingent upon the RNA secondary structure of the transcript, which is disturbed upon ASO binding. We therefore performed predictive RNA fold modeling, which suggested that the 31-bp deletion disrupted a stable hairpin present in the 3’UTR by creating a stem-loop that exposed a cryptic acceptor site with higher accessibility compared to the wildtype, Type 1, or Type 2 CALR transcripts.
To determine how IFNa may have elicited such a dramatic molecular response via exon 9 skipping, we re-purposed GoT to enrich for the alternative splice site in the original patient samples. Consistent with the splicing reporter assay, exon 9 skipping event occurred within CALR31del cells (and not in CALR wildtype). At baseline, the proportion of mutated cells with exon 9 skipping was enriched in multipotent lymphoid progenitors compared to myeloid progenitors. Following therapy, IFNa enhanced lymphoid differentiation overall in both CALR31del and wildtype, expanding the B-lymphoid progenitors within the CD34+ compartment (a finding consistently observed in the CALR-mutated MPN patient cohort in our study). These B-lymphoid progenitors displayed enhanced production of the exon 9 skipped isoform upon IFNa therapy, suggesting that IFNa-induced B-lymphoid differentiation of HSCs may enhance exon 9 skipping and thereby induce the loss of fitness advantage afforded by the CALR31del mutation.
Together, these results demonstrate that the 31-bp CALR deletion induces an exon 9 exclusion event that was enhanced by IFNa therapy and potentially linked to the remarkable molecular response. These findings motivate the development of novel therapeutic strategies that enhance exon 9 exclusion events by leveraging the endogenous splicing machinery.
DISCLOSURES
Danielle Rachel Isakov, PhD: No relevant conflicts of interest to declare.