What makes CALR-mutated myeloproliferative neoplasms an unusual model is that the mutation simultaneously creates an oncogenic driver and a shared neoantigen. Nearly all pathogenic variants shift the reading frame and generate a positively charged neomorphic C-terminus that is absent from the normal proteome. The two dominant mutation classes, the 52-base-pair deletion known as CALRdel52 and the 5-base-pair insertion known as CALRins5, share an identical 36-amino-acid mutant tail, and type 1 and type 2 proteins share an even longer 44-amino-acid sequence. In principle, this shared sequence could serve as a common immunologic target for vaccines or engineered immune cells across thousands of patients, a situation rarely encountered in cancer, where neoantigens are usually unique to individual tumors.

The immune system does indeed see the mutant protein. Early studies detected spontaneous HLA-DR-restricted CD4-positive T-cell responses to mutant CALR peptides, followed by reports of CD8-positive responses, memory responses in healthy donors, and recognition of autologous mutant cells. Some studies even found stronger responses in healthy individuals than in patients, a pattern compatible with the idea that effective immune surveillance eliminates emerging mutant clones before they become detectable. However, the distribution of responses across studies has been inconsistent, reflecting differences in peptide length, assay sensitivity, HLA restriction, and whether target cells naturally process and present the tested epitopes. Immunogenicity, the review stresses, is not equivalent to effective immune control.

Clinical experience with vaccination has made that distinction painfully concrete. The first long-peptide vaccination studies showed that repeated administration of a mutant CALR peptide was feasible and immunogenic, but most responding cells were CD4-positive T cells and hematologic responses were limited. T-cell receptor sequencing revealed that vaccine-expanded mutant-specific T cells did not substantially enrich in the bone marrow, where the mutant stem cells reside, and the ratio of effector cells to mutant myeloid cells was markedly unfavorable. More recently, a phase 1 study evaluated VAC85135, a heterologous adenoviral and vaccinia-vector vaccine encoding mutant CALR and JAK2 V617F neoantigens, alone or combined with the CTLA-4 inhibitor ipilimumab. Of fourteen enrolled patients, twelve had CALR-mutated disease; post-vaccination mutant-CALR-specific T-cell responses appeared in only five of fourteen patients, no patient experienced a reduction in mutant allele burden, stable disease was the best response in most myelofibrosis patients, and no responses were seen in essential thrombocythemia.

This gap between measurable antigenicity and failed disease control is where immunogenetics enters the picture. HLA class I and class II genotypes determine which mutant CALR-derived peptides can be stably presented to T cells, but peptide binding algorithms capture only one step in a multistage process that also includes proteasomal cleavage, TAP transport, peptide loading, complex stability, source-protein abundance, and the availability of a matching T-cell repertoire. A high-resolution Bulgarian study of 42 CALR-mutated cases, 158 JAK2 V617F-positive cases, and 1,083 controls reported depletion of the HLA-C*06:02 allele in CALR-mutated disease, with a nominal p-value of 0.0464, and proposed that genotype-level capacity to present CALR-mutant versus JAK2-derived peptides may contribute to driver-specific selection. The authors of the review categorize this finding as hypothesis-generating: the allele-level signal was not formally corrected for multiplicity, has not been independently replicated, and a cross-sectional association cannot reveal at which stage of clonal evolution such selection might act.

A large counterpoint complicates the broader immunosurveillance hypothesis. Walkowiak and colleagues predicted MHC class I and class II binding to 40 recurrent clonal hematopoiesis variants in approximately 380,000 UK Biobank participants and found no consistent association between predicted binding and either variant prevalence or clone size. This argues against a strong generalized effect of MHC-dependent immunosurveillance on early expansion of common hematopoietic clones. The review notes, however, that CALR exon 9 frameshift mutations were not directly evaluated in that analysis and generate an extended, shared neomorphic sequence rather than a single amino-acid substitution, so the UK Biobank findings narrow the hypothesis without excluding mutation-specific immune selection in CALR-mutated disease.

Innate immune genetics adds further candidate modifiers. NKG2D, an activating receptor encoded by KLRK1 and expressed by natural killer cells and cytotoxic lymphocytes, recognizes stress-inducible ligands including MICA and MICB. High-resolution genotyping in 43 CALR-mutated patients and 156 Bulgarian controls identified the MICA*004:001 allele as associated with CALR-mutated disease, with p equal to 0.004 and a Bonferroni-adjusted p of 0.047, while a KLRK1 haplotype association in a smaller subgroup showed an odds ratio of 3.61 but did not undergo independent replication. Separately, a study of 167 myeloproliferative neoplasm patients, including 30 CALR-mutated cases, found that the combination of inhibitory KIR2DL2 with HLA-C2 ligands was most strongly associated with CALR-mutated disease, with an odds ratio of 7.91 and a 95 percent confidence interval of 2.32 to 29.66. The review emphasizes that both signals derive from a single Bulgarian research network with partly overlapping cohorts and therefore require independent, ancestry-matched replication before any mechanistic role can be assigned.

Acquired immune escape mechanisms likely compound inherited vulnerabilities. Experimental work has shown that extracellular mutant CALR can suppress phagocytosis and antitumor immunity, and recent mechanistic studies demonstrated that in CALRdel52-driven systems and patient samples, mutant CALR signaling increased TGF-beta expression and regulatory T-cell expansion through an ERK/Sp1 axis; pathway interference restored antitumor T-cell activity in preclinical models. Patients with myeloproliferative neoplasms also harbor frequent CD8 T-cell-platelet aggregates associated with impaired T-cell function, suggesting that the disease’s marked megakaryocytic phenotype may itself contribute to immune suppression. Chronic inflammation, checkpoint-ligand expression, altered antigen-processing gene expression, and the suppressive marrow niche further weaken immune control, helping to reconcile demonstrable CALR antigenicity with the persistence of established disease.

Therapeutically, the field remains at an early stage. Clinically tested CALR-directed approaches are limited to vaccination and checkpoint modulation, neither of which has produced consistent molecular responses. Preclinical strategies have advanced further: mutant-CALR-directed CAR-T cells showed selective cytotoxicity against mutant-positive cell lines and patient-derived cells, controlled disease in xenograft models, and revealed a BCL-2-associated resistance mechanism that could be pharmacologically reversed with venetoclax, while the neoepitope-directed monoclonal antibody 4D7 selectively disrupted mutant CALR-MPL signaling in preclinical models. NKG2D-based approaches, MICA/MICB shedding inhibitors, and KIR-directed interventions remain extrapolated from other cancers and have not demonstrated CALR-specific efficacy. The review concludes that no immunogenetic biomarker is ready for clinical use and that progress will require multicenter, ancestry-matched cohorts with harmonized high-resolution genotyping, immunopeptidomic confirmation of naturally presented peptides, functional validation with autologous mutant targets, and prospective trials embedding immune correlates alongside molecular-response endpoints.

Subject of Research: Immune surveillance and immunogenetic modifiers of CALR-mutated myeloproliferative neoplasms

Article Title: Immune surveillance and emerging immunogenetic modifiers in CALR-mutated myeloproliferative neoplasms

Article References: Shivarov, V., & Ivanova, M. (2026). Immune surveillance and emerging immunogenetic modifiers in CALR-mutated myeloproliferative neoplasms. Frontiers in Hematology, 5, Article 1969104. https://doi.org/10.3389/frhem.2026.1969104

Image Credits: AI Generated

DOI: 10.3389/frhem.2026.1969104

Keywords: CALR mutations, myeloproliferative neoplasms, immune surveillance, HLA, neoantigen, NKG2D, MICA/MICB, KIR, natural killer cells, cancer vaccination, immunogenetics, clonal hematopoiesis