The patient was diagnosed with MPN according to the 2016 World Health Organization criteria in the Second Hospital of Shanxi Medical University. Mutational investigation in this patient was performed using a DNA sample from fresh bone marrow or peripheral blood samples. Patient history and clinical data were extracted from the medical records. This patient provided written informed consent, and this study was conducted in accordance with the tenets stipulated in the Declaration of Helsinki.

A peripheral blood sample of this patient was collected, and peripheral blood mononuclear cells were isolated. The cells were plated at a density of 1×10⁵cells/mL, which provided an optimal density for colony selection without the risk of contamination by neighboring colonies. Erythroid colony-forming units were cultured in methylcellulose-based medium (STEMCELL Technologies,VAN,CAN) containing additional 3 U/mL erythropoietin for 15 days. Granulocyte colony-forming units were incubated in methylcellulose-based medium containing an additional 100 ng/mL granulocyte colony-stimulating factor for 15 days. Erythroid and granulocyte colonies were selected, and DNA was extracted, amplified using a 2×STA Master Mix Kit (BBI Life Science,HK,China), and sequenced for JAK2 exon12 (sense 5’TGGGCCGAAGTCTGA CCCTTT 3’ and antisense 5’ ACAGAGCGAACCAATGC 3’) and CALR exon9 (sense 5’ TGGGGCGTAACAAAGGTG AG 3’ and antisense 5’ TGAAAGTTC TCGAGTCTCA CAGA 3’). After purifying the polymerase chain reaction product, Sanger sequencing (The Beijing Genomics Institute,SZX,China) was used to identify mutation sites.

We performed pathogenicity analyses for the novel missense variant using Mutation Taster score (https://www.mutationtster.org/ ), E-SNPs&GO (https://esnpsandgo.biocomp.unibo.it/), Polymorphism Phenotyping version 2 (PolyPhen2; http://genetics.bwh.harvard.edu/pph2/), and Sorting Intolerant From Tolerant (SIFT; https://sift.bii.a-star.edu.sg/). Alignment of homologous sequences of JAK2 protein from various species was performed using the Unipro UGENE software (49.1version, http://ugene.net/) to assess the conservation of amino acid residues at different sites. Protein three-dimensional (3D) model of wild-type JAK2 (JAK2 ^wt) was used the homodimeric JAK2 pseudokinase-protein tyrosine kinase (PK-PTK) model by JAK2 PK-PTK model (ma-evjj8). The five models of JAK2S533 was constructed based on the above ma-evjj8 by AlphaFold (https://alphafold.com/) and ColabFold (https://github.com/sokrypton/ColabFold) (14, 15). For further analysis, we used the JAK2S533 model with the highest per-residue confidence score (predicted local distance difference test) compared with the other models. The PyMOL system (4.5.0 vision) was used to visualize the results of the protein model. Detailed data of the model and score are provided in Supplementary Material 1 .

Complementary DNAs (cDNAs) for human JAK2N533S and control genes (JAK2 ^wt, JAK2 K539L, and JAK2N542-E543del) were synthesized by GenScript (NKG,CN). All cDNAs were cloned into a pCDH-CMV- MCS-EF1-CopGFP-Puro lentiviral-vector. Lentiviral particles were produced in 293T cells to infect Ba/F3 cells. The efficiency of gene transfer into Ba/F3 cells was assessed by green fluorescent protein (GFP) using laser confocal microscopy (Olympus, Tokyo, Japan) and flow cytometry (Beckman, CA,USA).

Ba/F3 cells were cultured in RPMI 1640 medium (Gibco, USA) containing 15% fetal calf serum (Thermo Fisher Scientific,NK,USA) and 10 ng/mL interleukin-3 (PeproTech,NJ, USA). Cell viability assay was performed using a CCK8 Assay Kit (DOJINDO,kumamato, Japan). For the CCK8 assay, 3,000 cells were seeded in each well of a 96-well plate. Each cell line was cultured for 5 days in the absence of IL-3.

Ba/F3 cells were deprived of IL-3 for 2 days. Total protein was extracted, processed using radioimmunoprecipitation assay (RIPA) buffer (Thermo Fisher Scientific, NK,USA), and supplemented with protease and phosphatase inhibitor mixture tablets (Thermo Fisher Scientific,NK,USA). Western blotting was performed by Simple Western (Protein Simple Technology,SV, USA). Antibodies against signal transducer and activator of transcription 3 (STAT3), STAT5, phospho-STAT3 (pSTAT3), and pSTAT5 were obtained from Cell Signaling Technologies (MA,USA).

A 75-year-old female patient presented with a gradual increase in platelet counts over 3 years. In December 2016, she was referred to the Second Hospital of Shanxi Medical University for evaluation of this marked thrombocytosis ( Figure 1A ). Despite the elevated platelet count, the patient was in good physical condition and did not experience any clinical symptoms, such as fever, itchy skin, facial flushing, bone pain, splenomegaly, or others. Furthermore, she did not have any complications, such as bleeding, thrombosis, or cerebrovascular disease.

A bone marrow biopsy was conducted to investigate the underlying reason for the thrombocytosis. The findings indicated normal cellularity with a normal ratio of myeloid to erythroid cells without fibrosis. However, there was a significant increase in megakaryopoiesis. Chromosomal analysis of 20 karyotypes revealed that the patient had a normal karyotype of 46, XX. Mutational analysis using peripheral blood revealed that the patient carried a JAK2 exon12 mutation (JAK2N533S) and a CALR mutation (CALRtype1). The patient was diagnosed with ET based on the platelet count (707 per mm³), proliferation mainly of the megakaryocytic lineage, and the presence of CALRtype1. However, there was no evidence of PV apart from the presence of JAK2N533S. The patient’s blood routine returned to normal and remained stable after treatment with hydroxyurea. Currently, she remains in good health despite carrying JAK2N533S and CALRtype1 ( Figures 1A, B ).

The clonal relationship between JAK2N533S and CALRtype1 was further investigated in this study. Peripheral blood mononuclear cells were isolated, and colony-forming assays were performed for erythroid cells and granulocytes. All 17 selected erythroid colonies exhibited a heterozygous JAK2N533S mutation (JAK2N533S^{het +} ). Of those, 14 colonies (82%) also carried a heterozygous CALRtype1 mutation (CALRtype1^het+), one colony (6%) harbored a homozygous CALRtype1 mutation (CALRtype1^hom+), and two colonies (12%) exhibited a wild-type CALR (CALR ^wt) genotype. In addition, all 21 selected granulocyte colonies exhibited a JAK2N533S^het+. Among those, 20 colonies (95%) had a CALRtype1^het+, and one colony (5%) exhibited a CALR ^wt genotype ( Figures 1C, D ).

Because JAK2N533S^het+ was present in all colonies, JAK2N533S was further tested using DNA extracted from hair follicles ( Figure 1E ). This analysis yielded positive results for the presence of JAK2N533S^het+, indicating that this is a germline mutation. In contrast, among these colonies, 20 (95%) exhibited a CALRtype1^het+, while one (5%) had a CALR ^wt genotype. This evidence further indicated that CALRtype1 was a late somatic event and occurred in a multipotent hematopoietic stem cell compartment capable of generating both myeloid and erythroid progeny.

Individual colony analysis revealed three distinct tumor subclones, namely JAK2 N533S^het+/CALRtype1^het+, JAK2N533S^het+/CALR ^wt, and JAK2N533S^het+/CALRtype1^hom+. Notably, significant expansion was observed in the JAK2N533S^het+/CALRtype1^het+ clone compared with the JAK2N533S^het+/CALR ^wt clone. This finding suggested that the acquisition of CALR mutation provides a growth advantage over the presence of JAK2N533S alone ( Figures 1D, E ).

We sought to better understand the pathogenicity of JAK2N533S. Hence, a bioinformatics analysis was conducted to determine the function of JAK2N533S. In this analysis, various mutation impairment prediction tools were utilized to predict the impact of Asn533Ser substitution on the JAK2 protein. The Mutation Taster score indicated that Asn533Ser substitution was pathogenic, suggesting a detrimental effect on protein function. However, other tools (e.g., E-SNPs&GO, PolyPhen-2, and SIFT) classified the N533S variant as benign, indicating that it is unlikely to significantly impact the protein function. The contradictory results obtained from the pathogenic prediction analysis of JAK2N533S are perplexing ( Figure 2A ). It is important to consider that different prediction methods may be characterized by inherent biases and limitations, which could contribute to discordant conclusions.

Additionally, a conservation analysis was performed to assess the level of conservation of Asn533 across different species. It was found that Asn533 is conserved in the JAK2 gene among species, such as humans, Pan troglodytes, Macaca mulatta, etc. However, of note, in Danio rerio, Asn533 can naturally be replaced with Ser ( Figure 2B ).

Furthermore, a comparison of the three-dimensional model of the homodimeric JAK2 PK-PTK model containing N533 and S533 was conducted ( Figures 2C, D ). The S533 mutation does not appear to significantly alter the structure of the JAK2 protein based on the merge of the JAK2 PK-PTK model of N533 and S533 ( Figure 2C ). The substitution of Asn with Ser could not trigger significant structural changes due to the slightly shorter length of Ser compared with Asn and their similar polar nature. Hence, it may cause subtle and confined alterations in JAK2 structure.

Previous research has determined that the JAK2K539L mutation in JAK2 exon12 is a pathogenic mutation that leads to JAK2 constitutive activation. This activation occurs by disrupting the highly charged region of D620 and altering the salt bridge interaction between residues D620 and E621. Moreover, the change of salt bridging among Lys539, Glu592, and Glu596 in the periphery of the hydrophobic interface could potentially lead to the stabilization of a constitutive active dimer (15–18). Could the N533S mutation influence them by establishing or disrupting the surrounding salt bridging and hydrogen bonds to promote JAK2 activation? There were no changes observed in the surrounding hydrogen bonds, and the N533S mutation did not establish a salt bridge interaction with residues D620, E621, and K539, as the distance between them exceeded 5 Å.

Collectively, based on the results of the bioinformatics analysis, JAK2N533S may exert only a minor effect on protein function. Further investigation and experimental studies are required to fully understand the functional consequences of this substitution.

To investigate the functional effects of JAK2N533S, we cloned the cDNAs of JAK2N533S and control genes (JAK2 ^wt, JAK2K539L, and JAK2N542-E543del) into a lentiviral expression vector with an GFP and puromycin; of note, JAK2K539L and JAK2N542-E543del are the most common gain-of-function mutations in JAK2 exon12 ( Figure 3A ). We used a lentiviral transduction to transfect the JAK2 gene with different mutation sites into Ba/F3 cells. We sorted the transgene-positive cells utilizing puromycin (2 ng/mL) and detected the positive rate of GFP by flow cytometry ( Figure 3B ). Next, we measured the IL-3-independent proliferation of cells. Each cell line was cultured for 5 days in the absence of IL-3. The experiment was conducted in triplicate. Cell viability assay was performed using CCK8. The control cells expressing JAK2K539L and JAK2N542-E543del showed significant accumulation, whereas there was no difference observed between cells carrying JAK2N533S and control cells expressing JAK2 ^wt, empty-vector, and untransfected Ba/F3 cells ( Figure 3C ). We further examined the levels of phosphorylated JAK2-downstream signal proteins, including STAT5 and STAT3. Cells were deprived of IL-3 cytokine for 2 days and analyzed by western blotting (ProteinSimple Technology) for phosphorylation of STAT5 and STAT3. Control cells carrying JAK2K539L and JAK2N542-E543del presented increased phosphorylation of STAT5(Tyr694) and STAT3(Tyr705), whereas those harboring JAK2N533S, JAK2 ^wt, empty-vector, and untransfected Ba/F3 cells did not show obvious activation of STAT5 and STAT3 ( Figures 3D, E ). The data obtained from the functional analysis indicated that JAK2N533S did not alter the function of JAK2 protein.