Background
Myelofibrosis (MF) is a chronic myeloproliferative neoplasm (MPN) that can be diagnosed as a primary disease (PMF) or secondary to polycythemia vera or essential thrombocythemia (PPV/PET-MF, also known as SMF). MF predominantly affects elderly patients with more than 65% of diagnosis occurring after 65 years of age [
1] and is clinically characterized by debilitating systemic symptoms, progressive splenomegaly, cytopenias, and overall reduced survival, mainly due to disease progression and leukemic transformation [
2‐
4].
The molecular pathogenesis of MF relates to mutations in three “driver” genes (namely:
JAK2,
CALR,
MPL) with around 60% of patients showing the
JAK2V617F mutation. Interestingly, 10% of patients are unmutated for the
JAK2,
CALR, and
MPL genes and are classified as triple-negative (TN) [
4]. Irrespective of “driver” mutations, a hyper-activation of the JAK-STAT pathway is observed in all MF patients [
5]. The molecular basis of TN remains mostly unknown, although a high molecular complexity has been previously described [
6] and rare, alternative, somatic mutations in both
JAK2 exon 14 and
MPL exon 10 have been previously annotated [
7,
8]. TN MF is associated with an aggressive clinical behavior characterized by a higher risk of developing anemia and thrombocytopenia, poorer outcomes in comparison with patients affected by the other MF molecular subtypes and a high rate of leukemic transformation [
9,
10].
Besides molecular alterations, a state of chronic inflammation involving the malignant hemopoietic stem/progenitor cells (HSPCs) and the non-malignant/malignant microenvironment has been indicated as the main contributor in MF initiation/clonal evolution [
11,
12]. Abnormal expression and activity of several cytokines involved in inflammation and immunoregulation are described in MF [
13,
14] and correlate with more severe marrow fibrosis [
15], worsening systemic symptoms [
16], and decreased survival [
17]. Importantly, this chronic inflammation may allow the neoplastic clone to gain a selective advantage [
18,
19] and the potential role of the interaction between the inflammaging process [
20] and niche effect has recently been raised up [
21].
Extracellular vesicles (EVs) have emerged as crucial actors in intercellular communication. Involved in a myriad of biological processes, including regulation of immunity and inflammation [
22,
23], they are released from a variety of cell types exerting pleiotropic effects. EVs present antigens and contain constituents from the cell of origin including microRNAs (miRNAs), transcription factors, mitochondria, nucleic acid, and microenvironmental signals that may contribute to the propagation of inflammation [
24‐
26]. Of interest, both damaged and functional mitochondria may be carried by EVs and they exert a role in immune regulation [
27‐
29]. Various hematological malignancies including MPN have been associated with increased numbers of circulating EVs [
30‐
33]. To date, the role of circulating EVs is still elusive in overall MF patients and their characterization remains fully obscure in the TN subclass.
In particular, the aggressive clinical behavior of TN MF patients is currently without any biological explanation and it is currently unclear the role of the HSPCs and/or the inflammatory microenvironment of TN MF patients. Here, we aim to identify a distinct signature of circulating CD34+ cells, cytokines, and EVs isolated from TN MF patients by comparatively evaluating the JAK2V617F-mutated counterparts.
Materials and methods
Patients
Peripheral blood (PB) was obtained from and MF patients (
n = 29) and age-matched healthy donors (HD;
n = 10). The clinical/laboratory characteristics of the patient cohorts are shown in Additional file
1: Table S1. In 15 pts., who received previous treatment (Hydroxyurea/Ruxolitinib), therapies had been discontinued for at least 3 months before sample collection.
CD34+ cells isolation
PB, anticoagulated with ethylenediaminetetraacetic acid (EDTA), was obtained from patients/controls. Mononuclear cells (MNC) were separated from MF and cord blood (CB) samples by stratification on Lympholyte-H 1.077 g/cm3 gradient (Gibco-Invitrogen, Milan, Italy), followed by red blood cell lysis for 15 min at 4 °C. MNCs were then processed on magnetic columns for CD34
+ cell isolation (mean purity 94% ± 5%) (MACS CD34 Isolation kit; Miltenyi Biotech, Bologna, Italy), as previously described [
19].
Functional characterization of CD34+ cells
Immunomagnetically isolated CD34
+ cells from MF patients or CB units were maintained in RPMI 1640 with 10% fetal bovine serum (FBS) with or without IL-1β (10 ng/mL, Thermo Scientific Pierce Biotechnology, Rockford, IL, USA), TNF-α (100 ng/mL, Thermo Scientific), IL-6 (10 ng/mL, Thermo Scientific), alone or in combination, for 24 h. In vitro survival was analyzed as previously described by apoptotic assay [
19,
34].
Clonogenic assay of CD34+ cells
MF/CB-derived CD34
+ cells were cultured in vitro to achieve hematopoietic cell differentiation and the formation of multi-lineage colony-forming units (CFU-Cs), including colony forming unit-granulocyte macrophage (CFU-GM) and Burst Forming Unit-erythroid (BFU-E) in the presence or the absence of IL-1β (1 ng/mL), TNF-α (100 ng/mL), IL-6 (10 ng/mL), alone or in combination, as previously described [
19,
34].
Migration assay of CD34+ cells
Migration of MF/CB purified CD34
+ cells were assayed towards a CXCL12 gradient (150 ng/ml, R&D) in transwell chambers (diameter 6.5 mm, pore size 8 μm; Costar; Corning), as previously described [
19,
34]. Specifically, 50 μl of RPMI 1640 plus 10% FBS containing 0,5 × 10
5 cells were added to the upper chamber and 150 μl of medium with or without CXCL12 ± IL-1β (1 ng/mL], TNF-α (100 ng/mL), IL-6 (10 ng/mL) (alone or in combination) were added to the bottom chamber.
The phenotype of circulating CD34+ cells
The phenotype of circulating CD34
+ cells was evaluated in PB from MF patients and in CB samples by conventional immunofluorescence, as previously described [
19,
34]. A minimum of 1 × 10
4 CD34
+ cells was acquired by flow cytometer BD Accuri C6 (Becton Dickinson). The analysis was performed excluding cellular debris in an SSC/FSC dot plot. The percentage of positive cells was calculated by subtracting the value of the appropriate isotype controls. The absolute number of positive cells/mL was calculated as follows: the percentage of positive cells × White Blood Cell count/100.
Gene expression profiling (GEP) of circulating CD34+ cells
GEP was performed on RNA samples of immunomagnetically isolated CD34
+ cells from MF patients using GeneChip Human Transcriptome Array 2.0 (Thermo Fisher Scientific), according to the manufacturer’s recommendations. Data quality control and normalization (signal space transformation robust multiple-array average, sst-RMA) and supervised analysis were carried out Transcriptome Analysis Console v4.0 software (Thermo Fisher Scientific). Fold-change absolute value ≥2 and
p ≤ 0.05 were used as a cut-off. The resulting genes were selected for functional annotation clustering, that was performed using David Bioinformatics Resources v6.8 (National Institute of Allergy and Infectious Diseases, NIH) [
35]. Gene set enrichment analysis (GSEA) was performed with GSEA v3.0 software (Broad Institute) [
36].
Collection of blood samples and isolation/enumeration of circulating EVs
Blood samples were collected into K2EDTA-containing collection tubes (Vacutainer® tubes, Becton Dickinson), plasma and EVs were prepared as previously described with minor modifications [
30,
37]. Briefly, platelet-poor plasma (PPP) was obtained (within 2 h from blood collection) after two consecutive centrifugations at 2500 g for 15 min at room temperature. The supernatant was subsequently ultracentrifuged at 100,000×g for 2 h at 4 °C using Optima L-90 K ultracentrifuge (Beckman Coulter) equipped with Type 50.2 Ti rotor. After centrifugation, pelleted EVs were resuspended and washed with Dulbecco’s PBS (DPBS; Sigma Aldrich). Finally, EVs were resuspended in saline buffer solution with 1% DMSO and stored at − 80 °C and/or used for further experiments. Isolated EVs were analyzed by nanoparticle tracking analysis (NTA), using the NanoSight LM10 system (NanoSight Ltd., Amesbury, UK), equipped with a 405 nm laser and with the NTA 2.3 analytic software, to define their dimension and profile.
The phenotype of the isolated EVs
To phenotype EVs isolated from patients/HD, the MACSPlex Exosome Kit (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany) was utilized. It provides the detection of 37 surface epitopes plus 2 isotype controls as for manufacturer’s instructions. This approach allows semi-quantitative analysis of differential surface epitopes. The proportion of megakaryocyte (MK)- and platelets (PLT)-EVs in the isolated EVs samples was analyzed by flow cytometry as previously described [
30] using CytoFLEX (Beckman Coulter).
Where indicated, MitoTracker™ Red CMXRos (Thermo Fisher Scientific, Waltham, MA, USA) was used to stain mitochondria/mitochondria components in plasma EVs following the manufacturer’s instructions and the geometric mean of fluorescence intensity (MFI) was determined by flow cytometry.
Appropriate controls were used: single color stains, sample dilution of EVs in double-filtered PBS or their buffer, unstained EVs were used to determine the fluorescence background as well as the buffer with antibody/dye. Antibody/dye filtration was performed before staining using Ultrafree®-MC/Durapore®-PVDF centrifugal filter units.
Transmission electron microscopy (TEM) of the isolated EVs
TEM was performed on EVs isolated by ultracentrifugation, placed on 200 mesh nickel formvar carbon-coated grids (Electron Microscopy Science, Hatfield, PA, USA) and left to adhere for 20 min. The grids were then incubated with 2.5% glutaraldehyde containing 2% sucrose and after washings in distilled water the EVs were negatively stained with NanoVan (Nanoprobes, Yaphank, NK, USA) and observed using a Jeol JEM 1010 electron microscope (Jeol, Tokyo, Japan).
Western blot analysis
EV protein extracts were separated by Sodium Dodecyl Sulphate-PolyAcrylamide Gel Electrophoresis (SDS-PAGE, Bio-Rad) and transferred onto nitrocellulose membranes.
Membranes were incubated overnight with the following antibodies: rabbit anti-TOMM20 (AbCam, Cambridge, UK; ab56783) and goat anti-ß-tubulin (Santa Cruz Biotechnology) as control. Horseradish peroxidase (HRP) conjugated anti-rabbit Immunoglobuluin (Ig) G (GE Healthcare) and anti-goat IgG (Santa Cruz) were used as secondary antibodies. Enhanced chemiluminescence Prime (ECL™) reaction kit (GE Healthcare, Amersham, U.K.) was utilized for detection by ChemiDoc XRS+ System (Bio Rad) and Image J software was applied to perform signal quantification.
Analysis of the miRNAs cargo of the isolated EVs
miRNA expression of isolated EVs (10
9) from patients/HD was investigated after RNA extraction with the miRNeasy Micro kit (Qiagen, Milan, Italy) and TaqManTM Array Human MicroRNA A and B Cards (Applied Biosystems by Life Technologies, NY, USA) according to manufacturer’s protocol. This discovery phase was assessed on 3 HD and 6 MF. In the latter case, patients were equally divided in
JAK2V617F-mutated and TN. To discover significantly different expressed miRNAs, RT-qPCR validation assay was performed, as previously described, in the same samples used for the profiling and in an enlarged cohort of 10 HD, 6 TN and 10
JAK2V617F-mutated patients in total. Data were normalized with cel-miR-39 spiked-in at lysis step of RNA extraction. MiRNAs’ targets were investigated via KEGG analysis, using the mirPath software with DIANA tools [
38]. This analysis is based on the already validated targets reported in the Tarbase database.
EV co-culture studies
Circulating CD34+ cells isolated from CB or MF patients were seeded in 96 wells plate and co-cultured over-night with plasma-derived EVs from HD or MF patients, respectively. EV concentration was calculated using Nanosight (Malvern) and EV concentration for co-cultures was determined by dose titration. Cells were then collected, counted, and stained with Annexin V and PI to detect the apoptotic rate by flow cytometry, as above described.
Plasma levels measurement of selected circulating cytokines
We measured the cytokine plasma levels of patients/HD by ELISA, according to the manufacturer’s instructions. PPP was obtained (within 2 h from blood collection) after two consecutive centrifugations at 2500 g for 15 min at room temperature. The plasma was then collected and stored at − 80 °C until quantification. In particular, the Human Thrombopoietin Quantikine ELISA Kit was provided from R&D Systems (Minneapolis, Minnesota, USA) and CXCL12 ELISA kit from Krishgen ByoSistems (Ashley CT, Whittier, CA, USA). The CiraplexTM immunoassay kit / Human 9-Plex Array (Aushon BioSystems, Billerica, MA, USA) was used for the measurement of various cytokines including IL-1β and TNF-α.
Mutation analysis
JAK2V617F allele-burden was assessed in granulocyte DNA with ipsogen JAK2 MutaQuant Kit (Qiagen, Marseille, France) 505 on 7900 HT Fast Real-Time PCR System (Applied Biosystem, Monza, Italy). CALR exon 9 sequencing was performed by Next Generation Sequencing (NGS) approach with GS Junior (Roche-454 platform; Roche Diagnostics, Monza, Italy); analysis was performed with AVA Software (GRCh38 as referenced). Rare CALR mutations identified by NGS were confirmed by Sanger sequencing. MPL mutations were investigated by ipsogen MPLW515K/L MutaScreen Kit (Qiagen) and by Sanger sequencing (for MPLS505N and other secondary exons 10 mutations).
Statistical analysis
Numerical variables have been summarized by their median and range, and categorical variables by count and relative frequency (%) of each category. Comparisons of quantitative variables between groups of patients were carried out by the nonparametric Wilcoxon rank-sum test. The differences between the groups were analyzed with Mann Whitney, Kruskal Wallis, one-way ANOVA tests as appropriate. Since miRNA fold change values and cytokines values were not normally distributed as computed by Shapiro-Wilk test, we performed Mann Whitney and Kruskal Wallis as non-parametric tests. All p values were considered as statistically significant when ≤0.05. Statistical analyses were performed using Graphpad (Graphpad Software Inc., La Jolla, CA) and SPSS software (PASW Statistics for Windows, Version 18.0. Chicago, IL).
Discussion
MF remains an incurable and critical disease [
40] and the lack of studies on MF, particularly on the TN subtype, has not conferred advances in the treatment landscape in the last years. In the present study, through an in-depth analysis of HSPCs and key microenvironmental factors including cytokines and EVs, we compared TN and
JAK2V617F-mutated patients providing novel insights in the pathogenesis of MF. As clearly reported by us [
19] and others [
41], the inflammatory microenvironment plays a crucial role in MF pathogenesis; however, few works deserved specific attention to the different molecular subgroups of MF patients.
Firstly, we demonstrated that the in vitro hemopoietic phenotype, survival, and function of circulating CD34+ cells are significantly altered in TN patients. Of note, we found an increase in the absolute number of circulating CD34+ cells in TN patients as a tool that may predict a more aggressive disease. These data were also consistent with the downregulation reported by GEP in selected genes involved in cell adhesion processes. Therefore, the enumeration of circulating CD34+ cells is highly relevant not only for improving diagnosis and risk assessment but also for evaluating response to potentially novel therapeutic approaches.
Specifically, TN patients show ex vivo
/in vitro increased apoptosis within the CD34
+ cell compartment with reduced hemopoietic function, revealing a stronger dependence on their microenvironment in comparison to the
JAK2V617F-mutated CD34
+ cells. Indeed, several genes involved in apoptotic and proliferation mechanisms are down-regulated in TN CD34
+ cells, confirming the presence of deregulated survival pathways and an intrinsic disadvantage of TN CD34
+ cells in terms of survival and proliferative intracellular signals. Among the differentially expressed genes, we observed an upregulation of tumor protein 53-induced nuclear protein 1 (
TP53INP1) in TN CD34
+, which is over-expressed during stress responses including inflammation [
42].
TP53INP1 is a pro-apoptotic gene and its upregulation might partly explain the increased apoptotic rate of the TN CD34
+ cells compartment in vitro. Consistent with this finding we observed KRAS signature enriched in TN CD34
+ cells. It is noteworthy that KRAS overexpression confers an adverse prognosis in cytogenetically normal acute myeloid leukemia (AML) [
43] and somatic activation of oncogenic KRAS in hematopoietic cells initiates a rapidly fatal myeloproliferative disorder [
44]. To sum up, in the present study, the GEP of TN HSPCs might highlight their vulnerability and might explain the more aggressive disease compared with the
JAK2V617F MF molecular subtypes. Furtherly, we previously reported that pro-inflammatory cytokines (e.g. IL1-β, IL6, TNF-α) promoted the survival, clonogenic capacity, and migration of CD34
+ isolated from MF patients, mostly
JAK2V617F mutated [
19]. By contrast, in the current study, we discovered that CD34
+ cells from TN patients were insensitive to specific pro-inflammatory cytokines, suggesting their dependence on other factors from their microenvironment. Thus, it remains a matter of discussion whether the defective response to cytokines of TN CD34
+ cells is related to altered apoptosis pathway, as reported by GEP analysis, “exhaustion” driven by previous inflammatory status, or to the dependence on “to be defined” factors from the microenvironment.
Notably, comparing the cytokine profiling between two subtypes of patients, TN patients show a milder inflammatory phenotype suggesting that therapies targeting the cytokine storm might be ineffective in this subset of MF patients. Moreover, only TPO plasma levels were lower in TN patients compared to the
JAK2V617F-mutated patients, suggesting TPO as a biomarker of poor prognosis in MF. In accordance,
Seiki Y et al. [
45] reported relatively low TPO levels in Myelodysplastic syndromes (MDS) patients and that low levels of TPO were associated with poor prognosis and progression to AML. For the first time, we also reported higher TPO levels in female MF compared to males. Interestingly,
Barraco et al. [
46] observed that female MF patients had a better prognosis with slower disease progression. Consequently, all these data point out that an inflammatory microenvironment is closely associated with the
JAK2V617F mutation but not to the TN counterparts. Whether this is due to the presence of a normally activated JAK/STAT pathway in TN patients remains a matter of speculation.
Based on our in vitro results, we then sought to determine whether the functional behavior of TN CD34
+ cells might be influenced by circulating EVs as signals from the microenvironment. Here, for the first time, we analyzed the phenotype, the mitochondrial content, and the miRNA cargo profile of the isolated circulating EVs from MF patients/HD. Since most circulating EVs are of megakaryocyte and platelet origin, we demonstrated that the MK-EVs were significantly reduced in both
JAK2V617F-mutated and TN patients. This finding confirms the occurrence of megakaryocyte abnormalities in MF; conversely, the PLT-EVs were decreased in TN patients. Of note, a previous study demonstrated that TPO may promote platelet aggregation [
47]; therefore, it is likely that the increased level of circulating TPO in the
JAK2V617F-mutated patients may partly contribute to the platelet activation and, as a consequence, to PLT-EVs release.
Recently, it has been described the existence and function of intact mitochondria or mitochondrial constituents transferred through EVs [
48]. Consistently, platelet EVs may transport mitochondria as a mechanism to mediate inflammation [
49] or immune cell regulation [
28]. Conversely, mitochondria, either naked or encapsulated by EVs, may trigger inflammation, may interact with innate immune signaling pathways and may regulate cell metabolism, apoptosis, and various pathophysiological situations [
50]. Here, we observed the presence of mitochondrial outer membrane protein TOMM20 in EVs from plasma and, interestingly, EVs from MF patients carry the highest level of MitoTracker MFI when compared with the healthy counterparts. These results, suggesting the presence of mitochondria in the EVs of our study, might be also in contrast to the TEM analysis that did not detect any intact mitochondria. However, although MitoTracker Red has been utilized as a membrane potential-sensitive dye, mitochondrial staining has several limitations [
51,
52] and up to date has not been fully explored in EVs. Therefore, our data support the hypothesis that the MitoTracker dye might be associated with mitochondrial components rather than with respiring mitochondria in EVs. Despite the need for further studies, this finding might play a role in the Darwinian selection of cancer cells with the potential of fueling tumor cell maintenance and proliferation. As far as the increased prevalence in elderly patients is concerned, the role of inflammaging, i.e. the low chronic level of the inflammatory status associated with the aging process [
20], could be another important risk factor for the change of microenvironment, making the cells more susceptible to undergo transformation, particularly for
JAK2V617F-mutated patients. In this perspective, the additional role of mitochondria circulation, a phenomenon likely related to the PB increase of mtDNA with aging [
53], and their possible transfer into other cells should be further addressed. To our knowledge, the mitochondria content of EVs is for the first time reported in MF patients and this might represent a novel clinically relevant therapeutic target that needs further investigations.
Recently, it has been also described that the miRNA cargo of EVs might have a prognostic role and can be used to evaluate disease progression [
54,
55]. Comparing MF patients and HD, we showed that a distinct miRNA profiling characterizes the MF patients EVs with upregulation of miR-34a-5p, − 127-3p, and − 212-3p. In accordance,
Bianchi et al. [
56] found miR-34a-5p upregulation in PMF CD34
+ hematopoietic progenitor cells, demonstrating that its overexpression favors the megakaryocyte and monocyte commitment of CD34
+ cells. Therefore, our data suggest that the alteration in MF megakaryocytes might be also due to EV-driven signals and not only to intrinsic mechanisms encountered in CD34
+ cells. It is therefore not surprising that miR-34a-5p expression is positively associated with the
JAK2V617F allele burden. Of note, miR-212-3p expression was negatively associated with the
JAK2V617F allele burden, in line with its role as tumor suppressor observed in AML [
57]. Additionally, we also confirmed the upregulation of miR-127-3p, another miR highly expressed in MF CD34
+ from MF patients as previously reported [
58]. Specifically, we observed a high miR-127-3p expression in the EVs isolated from the
JAK2V617F-mutated MF patients. Similarly to miR-34a [
59], miR-127 has been associated with DNA damage response and with cellular senescence [
60].
Notably, the common target of the identified miRNAs, i.e. PABPC1, plays a central role in mRNA processing binding the poly(A) tail of mRNA. PABPC1 deregulation by miRNAs interaction could contribute to making susceptible the cells of the microenvironment/niche to MF pathogenesis being known its role in carcinogenesis [
61]. Thus, it is interesting the hypothesis of potential epigenetic modulators which, circulating in the blood via EVs, may also transport “potential cancer signals” far from the original site of cell development, accordingly to the theory of “inflammaging and garb-aging” [
62].
Importantly, comparing the two groups of patients, only miR-361-5p expression was significantly upregulated in the TN EVs. Previous studies produced somewhat controversial results related to miR-361 [
63]. Increased expression of miR-361 was detected in AML suggesting that miR-361 dysregulation might be required to impair differentiation program in hemopoiesis and leukemia [
63]. It has also become apparent that miR-361 can downregulate the mRNA expression of IL-6 and IL-8 in endometrial cancer cells through targeting TWIST [
64]. This finding suggests a potential anti-inflammatory role of miR-361-5p, which is in line with the mild pro-inflammatory profile observed in TN patients.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.