Introduction

In eukaryotic cells, C-mannosylation is a unique type of protein glycosylation for tryptophan (Trp) residues in secretory or membrane-bound proteins1,2. This posttranslational modification commonly occurs at the first Trp in the Trp-X-X-Trp/Cys (W-X-X-W/C) sequence and directly attaches a monomeric α-mannose to the substrate proteins via a specific C-mannosyltransferase in the endoplasmic reticulum3. The C-mannosyltransferase gene has been identified as DPY19 in Caenorhabditis elegans, and its mammalian gene homologues that have the mannosyltransferase activity are DPY19L1 and DPY19L33,4,5. The consensus W-X-X-W/C motif is frequently identified in two major superfamilies of substrate proteins: the cytokine receptor type I family and thrombospondin (TSP) type 1 repeat (TSR) superfamily1,2. The former superfamily includes the erythropoietin, thrombopoietin, and interleukin (IL)-21 receptors. The latter superfamily comprises TSP-1, F-spondin, R-spondins, a disintegrin-like and metalloproteinase with TSP type 1 motifs 13 (ADAMTS13), and some complement components1. Although C-mannosylation is thought to regulate the folding and secretion of these proteins, the biological functions of C-mannosylation in each substrate protein have not yet been fully elucidated.

Monomeric C-mannosyl Trp (CMW) is a proteolytic degradation product of C-mannosylated proteins produced in cells1,6. Although CMW is a stable compound in mammals, it is reportedly further degraded with an enrichment of a bacterial consortium to be utilized as a carbon source7. CMW was first identified in human urine using a high-resolution mass spectrometer8. In 2019, a novel assay was developed to measure CMW using ultra-performance liquid chromatography (UPLC) with hydrophilic interaction liquid chromatography, with precision and accuracy consistent with the European Medicines Agency guidelines9,10,11. Using this assay, we previously identified that the distribution of CMW in mouse tissues and its abundance in the ovaries and uterus9. Moreover, plasma CMW levels in patients with ovarian cancer are significantly higher than those in patients with borderline and benign tumors. These findings suggest that plasma CMW is a possible biomarker of ovarian cancer12. Moreover, our recent study showed that serum CMW levels were strongly correlated with chronic vascular complications and chronic kidney disease (CKD) in patients with type 2 diabetes10. The results were also consistent with previous studies showing the correlation between blood CMW and renal dysfunction13,14,15,16,17,18,19. Taken together, the quantification of serum CMW can serve as a diagnostic or prognostic tool for various human diseases.

The current study aimed to focus on assessing CMW in the human hematological disorders. We first assessed the association between serum CMW levels and clinical characteristics in healthy Japanese workers and found that platelet count positively correlated with serum CMW levels. Therefore, we further investigated the clinical relevance of C-mannosylation in the human hematological disorders. Our findings suggested that serum CMW level could be a novel biomarker in patients with thrombocytosis of myeloproliferative neoplasms (T-MPN).

Results

Measurement of the serum CMW levels in healthy workers

First, we measured serum CMW levels of 94 healthy workers and analyzed their correlation between clinical characteristics. The healthy workers were originally enrolled in the previously reported clinical study20. Among them, we analyzed the subjects who met the criteria as defined in materials and methods in current study. The precision and accuracy of the novel assay using UPLC with hydrophilic interaction liquid chromatography to quantify serum CMW levels meet the European Medicines Agency guidelines11,12. Figure 1 shows the representative elution patterns of serum CMW levels. CMW was detected on the basis of the fluorescence intensity (excitation at 285 nm/emission at 350 nm) peak at 4.8 min (Fig. 1A, arrow). Serum a (thin line) and b (dotted line) were from healthy workers with high and low platelet counts, respectively. The peak of CMW was confirmed by comparing it with that of chemically synthesized CMW added as a standard compound (Fig. 1B, arrow). CMW content was quantified by measuring the area of the CMW peak as described in materials and methods. Table 1 shows the clinical characteristics of the 94 healthy workers. The median serum CMW level of healthy employees was 217.9 nM. Serum CMW levels were not significantly correlated with sex, white blood cell (WBC) count, red blood cell (RBC) count, hemoglobin (Hb) level, and hematocrit (Ht) level. In accordance with a previous report, serum CMW levels significantly correlated with serum creatinine (Cr) level (p = 0.046). It was also strongly and positively correlated with age (p < 0.001) and platelet count (p = 0.016) (Table 1).

Figure 1
figure 1

Measurement of C-mannosyl tryptophan (CMW) in the serum samples from healthy workers. Serum CMW levels were quantified using ultra-performance liquid chromatography (UPLC). The fluorescence intensity (excitation at 285 nm/emission at 350 nm) and the identity of CMW were confirmed and quantified. (A) Serum a (thin line) and b (dotted line) were from healthy workers with high and low platelet counts, respectively. (B) Comparison of serum b alone and serum b supplemented with synthesized CMW. The arrows indicate the peak corresponding to CMW.

Full size image
Table 1 Correlation between serum C-mannosyl tryptophan (CMW) levels and clinical characteristics between healthy workers (n = 94).
Full size table

Next, we investigated the serum CMW levels of 55 healthy workers with normal blood cell counts (Supplementary Figure S1)21. A positive and significant correlation was observed only for age in this cohort (p < 0.01) (Supplementary Table S1).

Measurement of the serum CMW levels in patients with hematological disorders

Because of the positive correlation between the serum CMW level and platelet count in healthy participants, the correlation between the serum CMW level and hematological disorders was further investigated. We focused on essential thrombocythemia (ET), a rare hematological neoplasm characterized by persistent thrombocytosis with megakaryocytic hyperplasia of the bone marrow (BM), which can lead to thrombotic and hemorrhagic complications, and progression to overt myelofibrosis (MF) and leukemia22,23. The estimated annual incidence of ET varies from 0.59 to 2.53 cases per 100,000 inhabitants, with a prevalence of 24 cases per 100,000 individuals24,25. In this study, we defined T-MPN as newly diagnosed ET, previously diagnosed ET, and those with subsequently confirmed MF. Thirty-four patients with T-MPN and 52 patients with other hematological disorders (as control group) during outpatient follow-up were enrolled (Fig. 2 and Supplementary Table S2). Table 2 shows the clinical characteristics of these patients. These two groups did not significantly differ in terms of age, sex, Cr level, and lactate dehydrogenase (LDH) level. The T-MPN group had a significantly higher WBC count, platelet count, Hb level, and Ht level than the control group (p < 0.05). In contrast, the estimated glomerular filtration rate (eGFR) of the T-MPN group was lower than that of the control group (median, 58.1 vs. 68.1 mL/min/1.73 m2; p = 0.004) (Table 2). Interestingly, patients with T-MPN (n = 34) had significantly higher serum CMW levels than those with control (n = 52) (median, 457.7 vs. 329.2 nM; p < 0.01) (Fig. 3A).

Figure 2
figure 2

The consort flow diagram of study participants. Thrombocytosis of myeloproliferative neoplasms (T-MPN) group includes newly diagnosed essential thrombocythemia (ET), previously diagnosed ET, and those with subsequently confirmed myelofibrosis (MF). Control group represents patients with hematological disorders excluding T-MPN.

Full size image
Table 2 Comparison of clinical characteristics between patients with hematological disorders excluding thrombocytosis of myeloproliferative neoplasms (control group, n = 52) and patients with thrombocytosis of myeloproliferative neoplasms (T-MPN group, n = 34).
Full size table
Figure 3
figure 3

(A) Serum CMW levels in patients with control (n = 52) and patients with T-MPN (n = 34). The circles show CMW levels, and red line shows the medians. (B) Excluding patients with grade 3b and higher chronic kidney disease, the serum CMW levels of patients with control (n = 52) and those with T-MPN (n = 25). (C) The receiver operating characteristic (ROC) curve of serum CMW levels to the differential diagnoses for patients with T-MPN (n = 34). The area under curve (AUC) was 0.702; sensitivity, 0.529; and specificity, 0.846. Based on these values, the CMW cutoff value was 451.4 nM. *p < 0.05, **p < 0.01.

Full size image

Based on a recent study, the serum CMW level of patients with type 2 diabetes who presented with eGFR < 60 mL/min/1.73 m2 (grade 3a or higher category based on the Kidney Disease Improving Global Outcome CKD guideline 2012) was significantly higher10. In this study, the serum CMW levels of 77 patients (52 with control and 25 with T-MPN) without grade 3b or higher CKD were reanalyzed. Even after excluding patients with moderate and severe renal dysfunction, there was a significant difference in the serum CMW level between the control and T-MPN groups (median, 329.2 vs. 395.3 nM; p = 0.039) (Fig. 3B).

In addition, we compared serum CMW levels in between healthy workers and T-MPN patients with grade 3b or higher CKD excluded. We found that serum CMW levels of T-MPN group were significantly higher than those of healthy group (median, 218.0 vs. 395.3 nM; p < 0.001) (Supplementary Figure S2).

Correlation analysis between the clinical characteristics and serum CMW levels in patients with myeloproliferative neoplasms with thrombocytosis (T-MPN)

Approximately 90% of patients with ET show three driver mutations (JAK2, CALR, or MPL) that are often mutually exclusive and linked to the phenotype and prognosis22,23. Supplementary Table S3 shows the mutation types and treatments of T-MPN. In terms of genetic mutations, 23 patients showed the JAK2V617F mutation, 9 patients did the CALR mutations, and 1 patient did the MPLW515L mutation. There was no significant difference in the serum CMW level between patients showing JAK2V617F mutation and those showing CALR mutation (median, 430.7 vs. 776.4 nM, p = 0.228).

The clinical characteristics correlated with serum CMW levels in patients with T-MPN were investigated. The results showed no correlation between serum CMW and laboratory findings, including platelet count. Meanwhile, the WBC count, Hb, Ht, LDH levels, Cr levels, and eGFR correlated with serum CMW levels (Table 3). Among these, low Hb, low Ht, and high LDH levels, which were involved in the progression to MF, were associated with ET severity26.

Table 3 Correlation between serum CMW levels and clinical characteristics in 34 patients with T-MPN.
Full size table

Next, the optimal cutoff point for serum CMW levels for the differential diagnosis of T-MPN was examined. Based on receiver operating characteristic (ROC) curve analysis, the optimal cutoff value for diagnosing T-MPN was 451.4 nM (area under the curve [AUC], 0.702; p = 0.002; sensitivity, 0.529; and specificity, 0.846) (Fig. 3C). Based on this cutoff value, patients with T-MPN were stratified into a CMW low-level group (17 patients with low CMW levels [≤ 452 nM]) and a CMW high-level group (17 patients with high CMW levels [> 452 nM]). The clinical characteristics of the two groups were compared (Table 4). The high-level CMW group had a significantly higher LDH and Cr levels than the low-level CMW group. Furthermore, the CMW high-level group had a significantly lower Hb and Ht levels than the CMW low-level group (median, 14.2 vs. 8.7 g/dL and median, 43.6 vs 26.9%, respectively; p < 0.001).

Table 4 Comparison of the CMW low-level (≤ 452 nM, n = 17) and CMW high-level (> 452 nM, n = 17) groups in patients with T-MPN.
Full size table

Correlation analysis between myelofibrosis and the serum CMW levels in patients with T-MPN

In the diagnosis of primary MF (PMF), the revised 2016 World Health Organization classification emphasizes the European consensus for evaluating the BM fibrosis grade27,28. According to this consensus, BM fibrosis is classified into four levels, from grades 0 to 3 (MF-0 to MF-3)27. Patients with PMF who presented with MF-0 and MF-1 (MF-0/1) BM fibrosis were diagnosed with prefibrotic primary myelofibrosis (prePMF). Patients with MF-2 and MF-3 (MF-2/3) BM fibrosis were diagnosed with overt PMF27,28. Patients with ET commonly have a favorable prognosis. However, some patients have a poor prognosis due to progressive MF followed by leukemic transformation29,30,31. Therefore, studies on serum biomarkers to evaluate MF in patients with ET are challenging. In this study, BM biopsy specimens were obtained during the same period as serum CMW levels were measured in 18 patients with T-MPN. Among them, 11 patients were pathologically diagnosed with MF-0/1, and 7 patients were diagnosed with MF-2/3. By monitoring the fluorescence intensity using UPLC, very high peak of CMW level was detected in the patient with MF-2, compared with the patient with MF-0 (Fig. 4A).

Figure 4
figure 4

(A) Comparison of representative serum samples from T-MPN patient with MF-2 (thin line) versus T-MPN patient with MF-0 (dotted line). The serum CMW levels were quantified using UPLC. The arrow indicates a peak corresponding to CMW. (B) Serum CMW levels of patients with MF-0/1 (n = 11) and with MF-2/3 (n = 7). The circles show CMW levels, and red line shows the medians. (C) Serum CMW levels of patients with MF-0 (n = 6) and with MF-1/2/3 (n = 12). (D) The ROC curve of serum CMW levels for diagnosing T-MPN with MF-1/2/3. The AUC was 0.875; sensitivity, 0.750; and specificity, 1.000. Based on these values, the CMW cutoff value was 517.9 nM. ns; not significant. **p < 0.01.

Full size image

Although not significantly, the serum CMW levels of patients with MF-2/3 were higher than those of patients with MF-0/1 (median, 847.4 vs. 430.7; p = 0.150) (Fig. 4B). Furthermore, the serum CMW levels of patients with MF-1/2/3 were significantly higher than those of patients without MF-0 (median, 811.9 vs. 315.2; p = 0.010) (Fig. 4C). Based on the ROC analysis, the optimal cutoff value for diagnosing T-MPN with MF-1/2/3 was 517.9 nM (AUC, 0.875; p = 0.011; sensitivity, 0.750; and specificity, 1.000) (Fig. 4D).

Discussion

Our results revealed that serum CMW levels were positively correlated with platelet count, but not with WBC count and Hb level, in healthy workers. These results suggested that protein C-mannosylation and the degradation product CMW might be associated with platelet production and function. Indeed, Sasazawa et al. reported that thrombopoietin receptor (c-Mpl), a pivotal molecule for megakaryocyte proliferation leading to platelet production, is C-mannosylated at the four C-mannosylation-consensus sites, and the all C-mannosyl modifications are required for c-Mpl-mediated JAK-STAT signaling32. This also suggested a functional link between protein C-mannosylation and platelet pathophysiology. To investigate the involvement of protein C-mannosylation in hematological diseases with platelet dysregulation, the present study evaluated the serum CMW levels of patients with MPN who presented with persistent thrombocytosis and characterized their clinical features.

Among various hematological disorders, T-MPN was associated with significantly high serum CMW levels. Importantly, patients with MPN without grade 3 or higher CKD had significantly higher serum CMW than those with other hematological disorders. Meanwhile, the quantification of serum CMW levels had high specificity (0.846), but low AUC (0.702) and sensitivity (0.529) for the differential diagnosis of T-MPN. Collectively, the correlation between serum CMW and T-MPN was validated, but the CMW level might not be suitable to be used as an independent diagnostic biomarker for T-MPN.

Regarding T-MPN in general, ET has a more favorable prognosis compared with other MPNs. However, patients with ET have a shorter survival time (approximately 20 years) than the general population33. Meanwhile, the cumulative incidence rates of fibrotic progression and blast transformation are 2.8–10.3% and 3.8–5.3%, respectively33,34. In patients with ET, progression to MF is a determining factor for survival outcomes, with a median survival of approximately 6–8 years30,31,35. Some studies have investigated the effects of prognostic factors on progression to post-ET MF22,23,30,35,36,37. Hb level < 10 g/dL, peripheral blood blasts ≥ 1%, and platelet count < 100 × 109/L are prognostic factors for post-ET MF30,35. Recently, a retrospective large-scale cohort study revealed that several independent risk factors, including splenomegaly, smoking, anemia, elevated LDH levels, and greater red blood cell distribution width, were independent factors for the early predictors of post-ET MF-free survival38. In addition to the MF-predictors in ET, serum CMW may also be a useful indicator when combined with the assessment of anemia and elevated blood LDH levels. Therefore, in the high CMW group, both anemia and elevated LDH levels should indicate disease severity associated with fibrotic progression and blast transformation in patients with ET.

In general, a BM biopsy is recommended by physicians for patients with suspected disease progression. However, BM biopsy should be performed by specialists because it is invasive and can cause pain and other complications. Therefore, exploring serum biomarkers of disease severity is particularly valuable for detecting the progression to MF during the early stages of long-term follow-up in patients with T-MPN. Previous studies have revealed altered inflammatory cytokine levels in patients with MPN, and IL-8 and IL-2 receptor levels are associated with PMF prognosis39,40,41. However, most studies on inflammatory cytokine profiles have not focused on ET severity. A recent study revealed that chemokine (C-X-C motif) ligand 1/growth-regulated oncogene-α, epidermal growth factor, and eotaxin were considered potential biomarkers of disease progression based on cytokine profiling in patients with ET42. Further analysis found that serum CMW level (cutoff 517.9 nM) had high accuracy (AUC: 0.875; sensitivity: 0.750; specificity: 1.000) in predicting BM fibrosis (MF-1/2/3) between T-MPN patients. Our findings suggest a stronger correlation between serum CMW level and fibrotic progression than between serum CMW level and diagnosis in patients with T-MPN. Ozono et al. found that the plasma tumor growth factor (TGF)-β1 levels of recipient mice transplanted with BM cells were significantly higher than those of mice carrying the JAK2V617F mutation that developed MF43,44. This pleiotropic cytokine is involved in the differentiation of neoplastic fibrocytes from neoplastic monocytes44. In our previous study, C-mannosylated TSP peptide attenuated the synaptogenesis of neurons via astrocytic TGF-β signaling45, suggesting the functional link between protein C-mannosylation and TGF-β signaling. Taken together, according to our data on the close association between serum CMW levels and fibrotic progression, C-mannosylation might have a functional role in disease progression via dysregulated TGF-β signaling. Hence, serum CMW level can be used as a biomarker to predict fibrotic progression in ET.

Although the precise mechanism underlying CMW upregulation in the blood of patients with T-MPN has not yet been determined, it may be correlated with a platelet activation-induced increase in the degradation of C-mannosylated proteins, such as TSP-1. TSP-1, a member of the TSP subfamily, is an extracellular matrix protein functioning in cell proliferation, cell adhesion, angiogenesis, and platelet aggregation46,47,48,49,50. In fact, TSP-1, a well-characterized C-mannosylated protein rich in platelets, is involved in the pathological status of PMF51,52,53. Moreover, circulating platelets are progressively cleared by macrophages and neutrophils in patients with ET54. Furthermore, CMW was produced in vitro by incubating mouse liver-derived lysosomal fractions with recombinant TSP-1 protein6. These findings suggest that activation and degradation of platelets in T-MPN can induce enhanced degradation of C-mannosylated proteins, such as TSP-1, leading to the upregulation of CMW production in T-MPN. However, further investigations are required to clarify the precise mechanism underlying the upregulation of CMW in T-MPN pathophysiology.

This study has a few limitations. First, only a small sample size was included, and this study was retrospectively conducted at a few centers in Wakayama, Japan. Second, BM biopsies were not performed in some patients with ET at initial diagnosis, and these patients were finally diagnosed based on the clinical judgment of the attending physician. Third, the direct pathogenesis of the contribution of C-mannosylation to fibrotic progression in patients with ET was not assessed. To address the query if C-mannosylation is cause or result of fibrotic progression, it is needed to analyze the animal model of MPN such as the mice carrying the JAK2V617F mutation43,44.

In conclusion, this study first measured the serum CMW levels of healthy controls and patients with T-MPN using a recently developed assay. The results showed that the serum CMW level could be a potential serum biomarker for predicting ET severity and prognosis. Present studies evaluating C-mannosylation using this novel assay for progression to overt MF in ET may provide promising future directions.

Materials and methods

Patients and study design

Healthy workers were originally enrolled in the previous clinical study20. Briefly, the study enrolled 99 healthcare workers aged ≥ 19 years who were employees of Wakayama City Medical Association and had the criteria but did not have the exclusion criteria including liver and renal dysfunction, as defined previously20. Their serum samples were collected in March 202120. Among them, 5 subjects were excluded owing to diabetes mellitus, because diabetic nephropathy affects serum CMW level10. For the study of Supplementary Figure S1, we employed the subjects who have normal blood cell counts, according to hematological parameter references among the Japanese population as previously reported21. As for the control group, 335 patients who visited the Wakayama Medical University Hospital from February 2023 to July 2023 were enrolled in this study. One hundred fifty-four patients diagnosed hematological disorder, 52 patients with given consent and collected their serum were final study control population. From February 2023 to February 2024, 53 patients who attended in the Wakayama Medical University Hospital as ET or newly diagnosed ET. Among these, 34 patients were finally enrolled in the study as T-MPN population. Serum samples were collected from the patients during the same period. ET was diagnosed using the 2017 World Health Organization classification and diagnostic criteria for ET or the diagnostic and classification criteria that were in effect at the time27. The diagnosis of BM fibrosis was based on the European consensus28. We retrospectively reviewed the medical records of all the participants to collect data on age, sex, and blood examination results, including complete blood count and serum biochemical test findings. Information on the mutational status and treatment of the 34 patients with T-MPN was also collected. Previous studies have identified somatic mutations, such as JAK2, CALR, and MPL55,56.

This retrospective study was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of Wakayama Medical University (approval number: 3540, approved on June 15, 2022) and Wakayama City Medical Association Seijinbyo Center (no. 202103–1). All the participants provided written informed consent.

Assessment of serum CMW levels

C2-α-C-mannosyl-L-Trp was synthesized as a standard molecule in accordance with previous studies9,10,12,57. Serum CMW was assessed using UPLC with hydrophilic interaction liquid chromatography, as previously described9,10,12,57. The precision and accuracy of the assay for quantifying serum CMW levels met the European Medicines Agency guidelines10,11. The serum fraction was isolated from the blood and was stored frozen at − 80 °C. To measure CMW levels, human serum samples were thawed and diluted with an extraction solution comprising methanol, acetonitrile, and formic acid (50:49.9:0.1). Subsequently, these were centrifuged at 12,000 × g for 15 min at 4 °C, and 140 μl of the supernatant was then filtered using a 0.20-µm polytetrafluoroethylene syringe filter. Thereafter, the collected samples were analyzed using a Waters Acquity UPLC H-Class system (Waters Corporation, Milford, MA, USA) with an Acquity UPLC BEH Amide column, a photodiode array detector, and a fluorescence detector, as previously described9. The CMW was quantified by measuring the fluorescence (excitation at 285 nm/emission at 350 nm)9. The Empower 3 software (Waters Corporation, Milford, MA, USA) was used to collect and process all data obtained.

Statistical analyses

Statistical analyses were performed using GraphPad Prism 10 (GraphPad Software Inc., San Diego, CA, USA) or the EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria)58. Statistical comparisons between groups were performed using the Pearson’s correlation coefficient, Mann–Whitney U test, and Fisher’s exact test, as appropriate. ROC curves were constructed, and the AUC was calculated to evaluate the utility of CMW levels in the diagnosis of T-MPN and the identification of progression to MF. p-values of < 0.05 indicated statistically significant differences.