Study design

This prospective cohort study enrolled patients treated at Hospital Universiti Sains Malaysia (USM) over a three-year period (2015–2018). Eligibility criteria comprised individuals with BCR-ABL-negative myeloproliferative neoplasms (MPNs)—PV, ET, and PMF—patients who subsequently progressed to secondary myelofibrosis, and those with MDS/MPNs. Diagnostic criteria followed the World Health Organization (WHO) Classification of Tumours of Haematopoietic and Lymphoid Tissues for myeloid neoplasms and acute leukaemia, 2008 edition and its 2016 revision.

Patient samples

The study enrolled patients diagnosed with either BCR-ABL-negative MPNs or MDS/MPNs who underwent JAK2 V617F mutation analysis at Hospital USM. Hematological parameters, including hemoglobin (Hb), platelet, red blood cell (RBC) and white blood cell (WBC) counts, as well as JAK2 V617F mutation status, were retrieved from the Hematology Laboratory Information System (LIS) of Hospital USM. Peripheral blood specimens from healthy controls (n = 12) were also collected. All procedures were conducted in accordance with the USM Human Research Ethics Committee guidelines (USM/JEPeM/20020079).

Genetic Analyses

Genomic DNA was isolated from peripheral whole blood using the QIAamp DNA Blood Mini Kit (Qiagen, Germany). Mutational analysis was performed on JAK2 exon 12, TET2 exons 1 and 7, and MYD88 exons 3, 4, and 5 using Sanger sequencing. Primers for JAK2 exon 12 were adopted from dos Santos et al. 2014 22, whereas primers targeting TET2 exons 1 and 7 and MYD88 exons 3–5 were designed de novo with Primer3Plus (primer3plus.com) and validated with NCBI BLAST. Target fragments were amplified by polymerase chain reaction (PCR) with exTEN II PCR Master Mix (1st Base, Singapore) containing 100 ng of template DNA per reaction. Amplicons were verified on 1 % agarose gels and then sent to 1st Base (Singapore) for capillary Sanger sequencing. Primer sequences and PCR conditions are provided in Additional File—Table 1.

Data and statistical analysis

Sequencing data analysis was conducted using Sequence Scanner software 2.0 (Thermo Fisher Scientific, USA). The identified sequence variations were cross-referenced with the single nucleotide polymorphism database (dbSNP) online tool to identify any existing polymorphisms. Statistical analysis was conducted utilizing the statistical software package GraphPad Prism version 9 (GraphPad Software, USA). Continuous variables are expressed as medians and ranges, whereas categorical variables are expressed as counts and percentages. The comparison of continuous variables was performed using an independent t test, whereas Fisher’s exact test was applied for categorical variables. A p value <0.05 was considered to indicate statistical significance.

Demographics and Hematological Profiles

In this prospective study, a total of 54 patients were enrolled, comprising 48 with BCR-ABL-negative MPNs—20 with PV, 13 with ET, 11 with PMF, and 4 with post-PV MF—and 6 with MDS/MPNs. The median age was 63 years (range, 30–90 years), with 56.3 % male predominance. Hematologic parameters varied according to disease subtype. PV patients exhibited significantly elevated RBC counts (median, 6.9 × 10/L), WBC counts (median, 15.9 × 10/L), and platelet counts (median, 520.5 × 10/L). Patients with ET had a median platelet count of 789.0 × 10/L, whereas those with PMF presented with anemia (median hemoglobin, 10.0 g/dL), lower RBC counts (median, 4.1 × 10/L), and variable leukocytosis and thrombocytosis. Compared with BCR-ABL-negative MPNs, MDS/MPN cases demonstrated higher WBC and platelet counts (p = 0.0054), consistent with the hyperproliferative phenotype of this entity. The demographic and hematologic characteristics of the patients are summarized in Table 1a and Table 1b.

Mutation Profiling

MYD88 Mutation

We detected an intronic SNP, rs4988457 (1944C>G), in three MPN patients (PV = two patients; ET = one patient) (Figure 1A, 1B, 1C), all co-mutated with JAK2 V617F and/or exon 12. No MYD88 variants were found in PMF or MDS/MPNs. Correlation study with hematological parameters observed no significant differences between MYD88-positive and MYD88-negative patients within the PV and ET groups (Table 4).

Patient’s demographics, hematological and mutation profiles	MYD88-positive		MYD88-negative		p-value (PV)	p-value (ET)
		PV (n=2)	ET (n=1)	PV (n=18)			ET (n=12)
Gendera
Male	1 (50.0)	0 (0.0)	10 (55.6)	7 (58.3)	>0.9999	0.4615
Female	1 (50.0)	1 (100.0)	8 (44.4)	5 (41.7)
Age (years)b	69.0 (66.0-72.0)	42.0	64.5 (40.0-79.0)	57.0 (30.0-90.0)	0.4006	0.5000
Age (years)a
≤59	0 (0.0)	1 (100.0)	4 (22.2)	7 (58.3)	NA	NA
60-69	1 (50.0)	0 (0.0)	9 (50.0)	2 (16.7)
70-79	1 (50.0)	0 (0.0)	5 (27.8)	1 (8.3)
≥80	0 (0.0)	0 (0.0)	0 (0.0)	2 (16.7)
Complete blood countsb
Hb (g/dL)	17.5 (16.4-18.6)	13.4	14.7 (8.8-23.4)	13.3 (10.2-14.8)	0.4394	0.6588
RBC count (x 1012/L)	6.6 (6.2-7.1)	4.6	6.9 (4.4-9.4)	5.0 (3.6-5.8)	0.7531	0.8508
WBC count (x 109/L)	14.5 (10.4-18.5)	7.2	15.9 (7.1-31.3)	11.7 (7.2-36.5)	0.7586	0.4428
Platelet count (x 109/L)	759.0 (475.0-1043.0)	867.0	520.5 (195.0-1388.0)	779.0 (297.0-867.0)	0.5238	0.9136
Gene mutationsa
JAK2 V617F	2 (100.0)	1 (100.0)	16 (88.9)	10 (83.3)	>0.9999	>0.9999
JAK2 exon 12	2 (100.0)	0 (0.0)	7 (38.9)	3 (25.0)	0.1895	>0.9999
TET2	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	>0.9999	>0.9999

JAK2 V617F mutation

The high prevalence of JAK2 V617F in our cohort (81.5%) is consistent with its well-established diagnostic value in BCR-ABL-negative MPNs according to WHO criteria 23,24. A separate Malaysian series documented the mutation in 73.6% of patients with MPN 25. Relative to neighbouring populations, the frequency was lower in Thailand (32.9%) 26, comparable in Singapore (77.5%) 27, and ranged from 56% to 67% in Chinese cohorts 28,29,30. The markedly elevated platelet counts observed in PV cases harbouring JAK2 V617F imply a higher thrombotic risk, in line with previous evidence of JAK2-driven platelet activation 31,32,33. Among patients with ET, mutation-positive individuals were significantly older, supporting an age-related accumulation of the mutation as reported in earlier studies 34,35. Although earlier publications demonstrated significant differences in haematological indices between MDS/MPN patients stratified by JAK2 status 36, the present work detected no such differences, probably because of the small sample size (n = 6). Overall, the robust association between JAK2 V617F and BCR-ABL-negative MPNs underscores its diagnostic and prognostic importance, although its precise relationship with clinical phenotype and thrombosis remains incompletely defined 17,37,38,39.

JAK2 exon 12 mutation

The analysis revealed that six patients (five with PV and one with ET) harbored a JAK2 exon 12 mutation. Worldwide, JAK2 exon 12 mutations account for approximately 2–5 % of PV cases, occur almost exclusively in JAK2-V617F–negative disease, and are exceedingly rare in ET/PMF and MDS/MPN 40,41.

Among the identified variants, D544N and c.1194+12G>A were consistently detected in repeat sequencing runs, suggesting that these variants are highly reliable. Previous literature has documented more than ten distinct sequence variations within the JAK2 exon 12 region, the majority of which cluster between codons 536 and 544 42, which is a recognized mutation hotspot 43,44,45. Mutations occurring in this region may disrupt the JH2 pseudokinase domain, analogous to V617F, thereby resulting in constitutive activation of JAK–STAT signaling 46,47. Notably, alterations involving the E543–D544 sequence have been reported in multiple studies 48,49, and the specific D544N variant has been detected in the Chinese population 28. The c.1194+12G>A variant, located 12 nucleotides downstream of exon 12 within intron 12, lies in close proximity to a previously described variant, c.1194+6T>C, identified in the same genomic region 22.

Interestingly, all JAK2 exon 12-positive cases also harbored the JAK2 V617F mutation. This co-occurrence is atypical, as exon 12 mutations are generally considered mutually exclusive with V617F in classical MPNs 50. Nonetheless, familial and complex PV cases with such co-mutations have been reported 48,51. A study from Qatar (Western Asia) likewise reported a high frequency of JAK2 exon 12 mutations in PV and ET that co-occurred with JAK2 V617F 52, suggesting possible geographical variation. The co-detection of JAK2 V617F and exon 12 variants may indicate subclonal heterogeneity arising from clonal mosaicism and disease evolution, although such events are rare 53. Because our analysis relied on Sanger sequencing—which is less sensitive than next-generation sequencing (NGS) and susceptible to artefacts—accurate quantification of low-level variant allele frequencies (VAF < 15–20 %) was not feasible. Accordingly, definitive conclusions cannot be drawn, and this observation warrants further investigation.

TET2 mutation

A recent meta-analysis reported TET2 mutation prevalences of 20.8 % in Asia, 17.4 % in North America, 13.0 % in Europe, and 7.0 % in Australia 54. Although TET2 is a common epigenetic modifier in MPNs, we did not detect mutations in exons 1 or 7 within our cohort. This observation aligns with prior studies demonstrating that pathogenic variants cluster predominantly in exons 3 and 11 55. Exon 1 contains promoter sequences that regulate gene expression, whereas exon 7 resides within a conserved domain. The absence of variants in these exons may therefore reflect either a low mutation frequency or epigenetic silencing mechanisms that are not detectable by conventional sequencing 56,57,58. Moreover, previous reports suggest that TET2 mutations alone are insufficient to initiate the MPN phenotype, although they may facilitate genomic instability and disease progression 3. The apparent absence of TET2 variants in our series may thus stem from the restricted exon coverage rather than their true biological absence. Comprehensive sequencing of the entire coding region will therefore be required to delineate the full TET2 mutational landscape.

MYD88 mutation

We identified three patients with MPNs—two with PV and one with ET—harboring the rare rs4988457 (c.1944C>G) MYD88 variant. Consistent with previous reports, MYD88 mutations are exceedingly uncommon in BCR-ABL–negative MPNs. In large sequencing series, including a cohort of 303 individuals, no MYD88 alterations were detected 59, with the exception of a single ET case complicated by Waldenström’s macroglobulinemia (WM), in which a MYD88 variant was observed 19. Accordingly, our study is among the first to evaluate MYD88 mutations in BCR-ABL-negative MPNs and in MDS/MPN overlap syndromes. Although the pathobiological significance of rs4988457 remains undefined, MYD88 aberrations have been implicated in multiple other hematologic malignancies, including T-cell lymphomas 60, activated B-cell–type diffuse large B-cell lymphoma (ABC-DLBCL) 61, chronic lymphocytic leukemia (CLL) 62, and AML 63.