Research ArticleMYELODYSPLASTIC SYNDROME

A variant erythroferrone disrupts iron homeostasis in SF3B1-mutated myelodysplastic syndrome

See allHide authors and affiliations

Science Translational Medicine  10 Jul 2019:
Vol. 11, Issue 500, eaav5467
DOI: 10.1126/scitranslmed.aav5467
  • Fig. 1 Differential gene expression in SF3B1MUT MDSs.

    (A) Biplot of the first two principal components (PCs) showing 45.4% of the variability within the data (PC1, x axis; PC2, y axis). (B) Volcano plot showing differentially expressed transcripts in SF3B1MUT BM MNCs compared to SF3B1WT BM MNCs. Fold change (FC) on the x axis and negative log10 of Benjamini-Hochberg (BH) corrected P values on the y axis. Dotted vertical and horizontal lines reflect the filtering criteria (FC < 0.5 or > 2.0; and BH-corrected P value < 0.05). The red dots represent differentially up-regulated transcripts, the blue dots represent differentially down-regulated transcripts, and the gray dots represent transcripts without differential expression. (C) GO enrichment analysis of differentially expressed genes with an absolute log2 (FC) >1 and a P value < 0.05 in SF3B1MUT MDS. The most down-regulated and up-regulated gene sets are represented. (D) Log2(FC) of the expression of 16 among the 96 genes of the IRON_ION_HOMEOSTASIS gene set (GO:005072) deregulated in SF3B1MUT MDS.

  • Fig. 2 Differential splice junctions in SF3B1MUTMDSs.

    (A) Hierarchical clustering and heat map of differential splice junctions between SF3B1MUT and SF3B1WT MDS samples. Values indicating percent usage of the differential splice junction versus all other junctions sharing the same ss are normalized as Z scores across patients and limited to a maximum of |Z| = 2. Rows are splice junctions with indicated types: acceptor (red), donor (blue), ambiguous (green), and differentially expressed canonical junction (gray). Columns are patients. (B) Venn diagram of the number of differential alternative 5′ donor and 3′ acceptor junctions in SF3B1MUT compared to SF3B1WT MDS. The overlapping area represents ambiguous junctions (n = 185). (C) Distances between the alternative (AG′) and canonical (AG) 3′ss within the 50 nt upstream of the AG plotted as a histogram. (D) Comparison of the expression of alternative junctions. The ratio AG′/AG′ + AG for each junction in SF3B1MUT (red) versus SF3B1WT (gray) MDS is shown. (E) Distribution of the biological functions of 63 genes affected by one or two aberrant 3′ss junctions located at <50 bases of the canonical 3′ss, an additional sequence multiple of 3 nt, a ratio AG′/(AG′ + AG) > 0.1 in SF3B1MUT samples, and an FC > 10. (F) Ratio of AG′/(AG′ + AG) in 26 genes whose expression was up- or down-regulated in patients with SF3B1MUT. (G) Sashimi plot of 3′ss canonical (*) and aberrant (**) junctions in the FAM132B/ERFE gene in three BM MNC samples, one SF3B1WT and two SF3B1K700E MDS. The ratio AG′/(AG′ + AG) is indicated in parentheses.

  • Fig. 3 SF3B1-dependent expression of 3′ss aberrant ERFE+12.

    (A) Induction of ERFE+12 by expressing SF3B1K700E in human SF3B1WT UT-7/EPO cell line. Cells were transfected with a pLVX plasmid encoding a synthetic SF3B1WT or SF3B1K700E cDNA. Nontransfected (NT) UT-7/EPO cells are shown as a control. The canonical ERFE and aberrant ERFE+12 transcripts were detected by capillary electrophoresis of fluorescent PCR products. The x axis represents molecular size (nt for nucleotides) of PCR amplicons, and the y axis represents relative fluorescent units (RFU). The peak at 150 nt corresponds to the canonical transcript, whereas the peak at 162 nt refers to the alternative transcript due to cryptic AG′ usage. (B) Analysis of alternative AG′ and canonical AG usage of ERFE and ENOSF1 minigenes transfected into murine CRISPR-Cas9 SF3B1WT (clone 9.2) and SF3B1K700E (clone 5.13H) G1E-ER4 cells by fluorescent PCR. Transfected and NT parental G1E-ER4 cells are shown as controls. The peak at 135 nt corresponds to the transcript generated by a canonical AG usage, whereas the peak at 147 nt refers to the alternative transcript due to cryptic AG′ usage. (C) Detection of ERFE+12 depends on the presence of an SF3B1 mutant in MDS. BM MNC RNAs from three patients with SF3B1 mutations (SF3B1K700E, SF3B1H622Q, and SF3B1N626D), three with mutations in other splice genes (U2AF1Q157P, SRSF2P95H, and ZRSR2H191Y), and three with IDH1R132C, ASXL1dupG646, or no mutations were analyzed. ERFE+12/ERFE+12 + ERFEWT ratios are indicated. (D) Quantification of ERFEWT and ERFE+12 transcripts by RT-qPCR in BM samples depicted in (C). Results are expressed as normalized ratio quantities (NRQ) to ACTB and B2M housekeeping genes. (E) Detection of ERFEWT and ERFE+12 transcripts in SF3B1MUT or SF3B1WT erythroid progenitors or basophilic/polychromatic erythroblasts (Baso/polyE) in culture in comparison with BM MNC. (F) Analysis of ERFE transcripts in SF3B1WT diseases with ineffective erythropoiesis. The image shows peripheral blood (PB) MNC from one patient with MDS-RS with TET2 and SRSF2 and no SF3B1 mutations, samples from three patients with congenital sideroblastic anemias (two BM samples with ALAS2 mutation and one PB sample with GLRX5 mutation), and one PB sample from a patient with severe β-thalassemia (β-Thal). Samples were analyzed by capillary electrophoresis of ERFE PCR products. One PB sample from a patient with SF3B1K700E MDS was used as a positive control.

  • Fig. 4 Identification of ERFEVPFQ peptide by mass spectrometry and hepcidin repression by recombinant ERFEVPFQ protein.

    (A) Amino acid sequence of ERFEWT and ERFEVPFQ peptides. VPFQ (red), collagen domain (blue). (B) Identification of a specific ERFEVPFQ peptide in erythroblast cell lysates by mass spectrometry using nano liquid chromatography coupled with a Q Exactive Plus mass spectrometer. The calculated peptide mass was 3227.650782 from 3+ ion observed at mass/charge ratio (m/z) 1076.890870 with measured Δ = 3.6 ppm, Mascot score = 17, expectation value = 0.069. HY, hydroxylated proline residues. (C) SDS–polyacrylamide gel electrophoresis and Coomassie blue staining of purified ERFEWT and ERFEVPFQ in reducing and nonreducing conditions. (D) Hep3B and HepG2 hepatocellular carcinoma cells were treated with purified recombinant human ERFEWT or ERFEVPFQ (2 μg/ml) for 16 hours. HAMP was quantified by RT-qPCR and normalized to HPRT. Data shown are means ± SEM of three independent experiments and represent an FC of hepcidin mRNA expression in ERFE-treated compared to untreated (CTRL) cells. Two-tailed Student t test for P values; ***P < 0.0001; **P < 0.001.

  • Fig. 5 Increased plasma concentration of ERFE in patients with SF3B1MUT MDS.

    Quantitative analysis was performed in plasma samples collected from 20 non–blood donor healthy volunteers (black), 156 patients with MDS including 94 SF3B1MUT (red) and 62 SF3B1WT (gray) representing the training cohort (A to D), and 55 patients with MDS including 42 SF3B1MUT (red) and 13 SF3B1WT (gray) representing the validation cohort (E to H). The graphs show quantification of erythroferrone (A and E), ferritin (B and F), hepcidin (C and G), and hepcidin/ferritin ratio (D and H). Results are expressed as medians and interquartile ranges (IQRs). The boxplots represent the median and the first and third quartiles, and the whiskers represent the lowest and the highest values still within the 1.5 IQR of the lower and upper quartiles. Mann-Whitney for P values.

  • Fig. 6 Erythroid cell-restricted expression of ERFE+12.

    (A) May-Grünwald-Giemsa–stained cytospins of erythroid and granulocytic precursors obtained from liquid culture of BM SF3B1WT or SF3B1K700E CD34+ progenitors. (B) Quantification of ERFEWT and ERFE+12 in erythroid and granulocytic precursors by RT-qPCR. Results are expressed as NRQ ± SEM to ACTB and B2M housekeeping genes. (C) ERFE+12 is absent in SF3B1MUT CLL. BM MNCs from a patient with SF3B1K700E CLL + MDS or PB MNCs from patients with SF3B1T663I CLL, SF3B1WT CLL, and SF3B1K700E MDS were collected for analysis by capillary electrophoresis of ERFE and MAP3K7 fluorescent PCR products. ERFE+12 was detected as a 162-nt fragment and MAP3K7+20 was detected as a 170-nt fragment. PB samples from one patient with an SF3B1WT CLL and one patient with an SF3B1MUT MDS are shown as controls. (D) ERFE+12 expression is restricted to SF3B1MUT myeloid lineage. CD19+CD5 B cells, CD19+CD5+ B CLL cells, CD3+ T cells, and myeloid cells were sorted from the BM MNC fraction of an SF3B1K700E CLL + MDS, an SF3B1K700E MDS, and an SF3B1WT MDS and from the PB MNCs of an SF3B1T663I CLL. ERFE and MAP3K7 transcripts were analyzed by fluorescent PCR. (E) The sequencing of SF3B1 was performed on cDNA of each cell fraction. +, mutated; WT, wild type; NA, not available.

  • Fig. 7 ERFE+12 expression as a marker of clonal erythropoiesis and survival.

    Fluorescent PCR was performed at screening and evaluation in (A) 10 paired samples from patients with SF3B1MUT MDS enrolled in the GFM-Retacrit-2013 clinical trial (four nonresponding and six responding patients) and (B) 14 paired samples from patients with SF3B1MUT MDS enrolled in the GFM-LenEpo-2008 clinical trial (eight nonresponding and six responding patients). Peak heights of ERFE+12 and ERFEWT signals were integrated as ERFE+12/ERFE+12 + ERFEWT ratios. Percent variations of ratios are indicated (right) as medians and IQR (25 to 75%). Mann-Whitney test for P values. (C) Overall survival according to ERFE+12/ERFE+12 + ERFEWT ratio shown as a Kaplan-Meier curve. A threshold of positivity of 0.008 was determined by ROC analysis. Log-rank test for P value. ns, not significant.

  • Table 1 Erythroferrone, hepcidin, and SF3B1 mutation as independent predictors of hyperferritinemia in low transfusion burden patients with MDS.

    Erythroferrone, hepcidin, sTfR, number of RBC units per 8 weeks, and SF3B1 mutation were evaluated in the group of 60 patients receiving less than 4 RBC units per 8 weeks on the basis of ferritin with a cutoff value of 300 ng/ml. Parameters are indicated as means and 95% confidence intervals (95% CI) or ranges. For univariate analysis, Mann-Whitney and χ2 tests were used to compare the variables between SF3B1MUT and SF3B1WT MDS. Multivariate logistic regression analysis was performed for variables with P value < 0.1 in univariate analysis.

    ParametersFerritin < 300 ng/mlFerritin ≥ 300 ng/mlUnivariateMultivariate
    n = 19n = 41P valueP value
    Erythroferrone (ng/ml),
    mean (95% CI)
    36.7 (26.8–46.6)72.7 (55.4–90.0)0.0050.002
    Hepcidin (ng/ml), mean
    (95% CI)
    17.2 (11.2–23.2)35.9 (25.9–45.9)0.013<0.0001
    sTfR (ng/ml), mean (95% CI)1.21 (0.99–1.43)1.42 (1.18–1.65)0.484
    Number of RBC units/
    8 weeks, mean (range)
    0.2 (0–2)0.6 (0–3)0.104
    SF3B1 mutation yes, n (%)3 (15.8)22 (53.6)0.0060.023

Supplementary Materials

  • stm.sciencemag.org/cgi/content/full/11/500/eaav5467/DC1

    Materials and Methods

    Fig. S1. Genomic characteristics of the training cohort of 156 patients with MDS.

    Fig. S2. Identification of in-frame alternative transcripts in SF3B1MUT MDS samples by RNA sequencing.

    Fig. S3. Allele-specific investigation of mutant SF3B1 causal relationship to aberrantly spliced ERFE+12 transcript using Degron-KI strategy.

    Fig. S4. Massive use of FAM132B/ERFE cryptic junction in one MDS-RS with a bi-allelic alteration of the SF3B1 gene.

    Fig. S5. Identification of ERFE peptide by LC MS/MS.

    Fig. S6. Correlation of plasma ERFE concentrations with ERFE+12 transcript, WHO classification, and iron homeostasis parameters.

    Fig. S7. Biological parameters of low transfusion burden patients with MDS in the training cohort.

    Fig. S8. Plasma ERFE, ferritin, and hepcidin concentrations at enrollment in 59 patients with MDS according to the response to EPO.

    Fig. S9. Erythroid-specific expression of ERFE variant transcript.

    Fig. S10. Variations of ERFE+12 transcript expression SF3B1MUT low-risk MDS after therapy.

    Fig. S11. Variations in plasma concentration of ERFE protein in SF3B1MUT low-risk MDS after therapy.

    Fig. S12. Kaplan-Meier curve for the analysis of OS according to ERFE+12/ERFEWT + ERFE+12 ratio.

    Table S1. Clinical and biological characteristics of the training cohort of 156 patients with MDS according to SF3B1 status.

    Table S2. Differential expression of ERFE transcripts according to the mutational status of SF3B1, TET2, and DNMT3A genes by RNA sequencing.

    Table S3. Primer sequences used for fluorescent PCR and RT-qPCR.

    Table S4. Expression of total ERFE, ERFEWT, and ERFE+12 transcripts by fluorescent PCR and RT-qPCR in nine MDS BM samples.

    Table S5. Clinical and biological characteristics of the validation cohort of 55 patients with MDS according to SF3B1 status.

    Table S6. Clinical and biological characteristics of the cohort of 59 patients with MDS in the GFM-Retacrit-2013 cohort according to response status.

    Table S7. Clinical and biological characteristics of the cohort of 90 patients with MDS with survival data according to SF3B1 and ERFE+12 status.

    Data file S1. Differentially expressed transcripts between patients with SF3B1MUT and SF3B1WT MDS by RNA sequencing and custom code.

    Data file S2. Differentially expressed 5′ and 3′ junctions between patients with SF3B1MUT and SF3B1wT MDS by RNA sequencing.

    Data file S3. In-frame and differentially expressed 3′ cryptic ss junctions in SF3B1MUT MDS.

    References (59, 60)

  • The PDF file includes:

    • Materials and Methods
    • Fig. S1. Genomic characteristics of the training cohort of 156 patients with MDS.
    • Fig. S2. Identification of in-frame alternative transcripts in SF3B1MUT MDS samples by RNA sequencing.
    • Fig. S3. Allele-specific investigation of mutant SF3B1 causal relationship to aberrantly spliced ERFE+12 transcript using Degron-KI strategy.
    • Fig. S4. Massive use of FAM132B/ERFE cryptic junction in one MDS-RS with a bi-allelic alteration of the SF3B1 gene.
    • Fig. S5. Identification of ERFE peptide by LC MS/MS.
    • Fig. S6. Correlation of plasma ERFE concentrations with ERFE+12 transcript, WHO classification, and iron homeostasis parameters.
    • Fig. S7. Biological parameters of low transfusion burden patients with MDS in the training cohort.
    • Fig. S8. Plasma ERFE, ferritin, and hepcidin concentrations at enrollment in 59 patients with MDS according to the response to EPO.
    • Fig. S9. Erythroid-specific expression of ERFE variant transcript.
    • Fig. S10. Variations of ERFE+12 transcript expression SF3B1MUT low-risk MDS after therapy.
    • Fig. S11. Variations in plasma concentration of ERFE protein in SF3B1MUT low-risk MDS after therapy.
    • Fig. S12. Kaplan-Meier curve for the analysis of OS according to ERFE+12/ERFEWT + ERFE+12 ratio.
    • Table S1. Clinical and biological characteristics of the training cohort of 156 patients with MDS according to SF3B1 status.
    • Table S2. Differential expression of ERFE transcripts according to the mutational status of SF3B1, TET2, and DNMT3A genes by RNA sequencing.
    • Table S3. Primer sequences used for fluorescent PCR and RT-qPCR.
    • Table S4. Expression of total ERFE, ERFEWT, and ERFE+12 transcripts by fluorescent PCR and RT-qPCR in nine MDS BM samples.
    • Table S5. Clinical and biological characteristics of the validation cohort of 55 patients with MDS according to SF3B1 status.
    • Table S6. Clinical and biological characteristics of the cohort of 59 patients with MDS in the GFM-Retacrit-2013 cohort according to response status.
    • Table S7. Clinical and biological characteristics of the cohort of 90 patients with MDS with survival data according to SF3B1 and ERFE+12 status.
    • References (59, 60)

    [Download PDF]

    Other Supplementary Material for this manuscript includes the following:

    • Data file S1 (Microsoft Excel format). Differentially expressed transcripts between patients with SF3B1MUT and SF3B1WT MDS by RNA sequencing and custom code.
    • Data file S2 (Microsoft Excel format). Differentially expressed 5′ and 3′ junctions between patients with SF3B1MUT and SF3B1wT MDS by RNA sequencing.
    • Data file S3 (Microsoft Excel format). In-frame and differentially expressed 3′ cryptic ss junctions in SF3B1MUT MDS.

Stay Connected to Science Translational Medicine

Navigate This Article