To the Editor:
Somatic mutations in the splicing factor SF3B1 occur in about one-third of all myelodysplastic neoplasms (MDS) and define a subgroup of patients characterized by ring sideroblasts (RS), ineffective erythropoiesis, and an indolent disease course in lower-risk (LR) MDS [1]. They are typically heterozygous missense substitutions, most commonly (>50% in MDS) involving p.K700E (SF3B1 NM_012433.4: c.2098A>G (p.Lys700Glu), hereafter referred to as SF3B1K700E), and have been shown to induce mis-splicing of key genes throughout erythroid differentiation [2, 3]. Surprisingly, although SF3B1 mutations are known to target multipotent lymphomyeloid hematopoietic stem cells and clonally propagate to myeloid progenitors [4], their impact on mature immune cells remains largely unexplored. Clinically, SF3B1 mutations are associated with high response rates to the erythroid maturation agent luspatercept and lower response to immunosuppressive treatment (IST) [5,6,7].
In this study, we performed multiplex immunophenotyping in conjunction with machine learning-based analytical approaches on bone marrow (BM)/peripheral blood (PB) samples from newly diagnosed or disease-modifying treatment-naïve SF3B1mut or SF3B1wt MDS patients (experimental cohort: Supplementary Table S1; Fig. S1) and healthy donors (HD) to identify genotype-immunophenotype correlations. Initial gene expression profiling of 730 immune-related genes in SF3B1mut versus SF3B1wt MDS BM mononuclear cells (BM-MNCs) revealed a predominantly myeloid cell-related innate immune gene signature (e.g., CYBB, CSF1R) lacking signs of overt myeloid-driven inflammation (i.e. IL1B, CXCL5), whereas lymphoid-related genes were underrepresented (e.g., CD3D, CD79A) (Supplementary Table S4, Fig. S2). These results are consistent with the reported lower proportion of lymphocytes in BM [8], mild myeloid dysplasia [9], and our previous finding of significantly lower IL1B mRNA in BM monocytes from SF3B1mut LR-MDS [10]. As IL-1β protein levels in paired BM plasma samples were often below the detection limit, we could not determine whether lower mRNA levels correspond to lower cytokine levels.
Next, we conducted high-dimensional mass cytometry (CyTOF) on BM-MNCs and analyzed data using the Tracking Responders EXpanding (T-REX) algorithm to identify immunophenotypic differences associated with LR-MDS and SF3B1K700E LR-MDS in particular. As expected, LR-MDS (SF3B1mut and SF3B1wt) showed several immunophenotypic changes consistent with an activated immune response (Fig. 1A, Supplementary Fig. S3/S4), in particular specific clusters resembling terminally differentiated effector memory CD8+ T cells (TTE/TEMRA, cluster 1295), mature CD57+ NK cells (cluster 2495), CD27+ IgD− memory B cells (cluster 795), and γδ T cells with an exhausted immunophenotype (cluster 1395). LR-MDS exhibited dysregulated T-cell homeostasis, with fewer naïve CD4+ and CD8+ T cells, and memory phenotype skewing toward CD8+ effector memory (TEM) and TTE cells (Supplementary Fig. S4). This is consistent with progressive memory differentiation entailing loss of survival, which could contribute to impaired long-term antitumor immunosurveillance.
We then compared SF3B1K700E to SF3B1wt LR-MDS using the T-REX pipeline, which identified a SF3B1K700E-specific cluster comprising CD33+ CD14+ monocytes (cluster 495, p < 0.01) (Fig. 1B, Supplementary Fig. S5). Further analysis of CD33+ CD14+ BM-MNCs showed that a remarkable proportion of the monocytes in SF3B1K700E LR-MDS adopt a HLA-DRlow/neg phenotype (Fig. 1C). Importantly, retrospective analysis of diagnostic flow cytometry data (Fig. 1D–E) and external validation in two independent cohorts comprising combined 130 MDS (118 LR-MDS) patients (Supplementary Fig. S6) confirmed an increased frequency of HLA-DRlow/neg monocytes in SF3B1mut (both SF3B1K700E and SF3B1nonK700E) compared to SF3B1wt MDS and HD. The external data support our observation that staining cryopreserved BM-MNCs may underestimate the actual frequency of HLA-DRlow/neg monocytes. Additionally, we found a strong correlation between HLA-DRlow/neg monocyte frequencies in BM and PB (Supplementary Fig. S7). HLA-DRlow/neg monocytes in SF3B1mut MDS were classical monocytes (CM) based on the lack of CD16 surface expression (Supplementary Fig. S7). Analysis of longitudinal data from four SF3B1K700E MDS patients showed a consistently high frequency of HLA-DRlow/neg monocytes (Fig. 1F).
To the best of our knowledge, the only other study directly investigating immunophenotypic features in BM of SF3B1mut MDS reported lower expression of CD11b, CD36, and CD64 on monocytes [8]. Another study found a higher frequency of thrombomodulin-expressing CM in MDS subtypes with <5% blasts and RS [11]. The association of SF3B1 mutations with lower monocyte surface HLA-DR expression identified here may be of clinical relevance, for example in view of the predicted poor response of SF3B1mut MDS to IST [6, 7]. Overall, the frequency of HLA-DRlow/neg monocytes showed no correlation with blood hemoglobin levels, the Revised International Prognostic Scoring System (IPSS-R) risk classifications, or the mutational burden of SF3B1 and co-mutated TET2 or DNMT3A (Supplementary Fig. S7). HLA-DRlow/neg monocyte frequencies were comparable between transfusion-dependent and –independent SF3B1mut LR-MDS (Supplementary Fig. S6). However, HLA-DRlow/neg monocytes have known immunoregulatory properties via multiple mechanisms, including effector T-cell inhibition, decreased antigen presentation, and defective dendritic cell maturation [12]. A possible scenario is that the early acquisition of SF3B1 mutations [13] and the presence of inflammation foster the emergence of HLA-DRlow/neg monocytes, which then contribute to counteract and balance inflammatory responses in established SF3B1mut MDS. In this context, T-REX also identified a cluster of naïve CD4+ T cells specific to SF3B1K700E LR-MDS with low expression of the co-stimulatory molecule CD27 (cluster 3895, MEM score CD27+1) (Fig. 1B, Supplementary Fig. S4/S5). Thus, although disease-related shifts in CD4+/CD8+ T-cell differentiation were noticeable irrespective of SF3B1 mutation status, naïve CD4+ T cells in SF3B1K700E LR-MDS displayed subtle immunophenotypic differences indicative of less recent activation.
As CD14+ monocytes lose HLA-DR expression, they become functionally deactivated, which can contribute to the transition to a more immunosuppressed state. To investigate whether this is the case for CM from SF3B1K700E LR-MDS, we studied their global gene expression profile using RNA-seq. Overall, we found 545 up- and 812 downregulated genes in the clonally involved CM from SF3B1K700E LR-MDS compared to HD (Supplementary Table S5). Importantly, these patients harbored an isolated K700E mutation and no confounding cytogenetic aberrations. Upregulated genes were enriched in genes involved in oxygen transport (e.g., HBB, HBA1/2), probably due to erythrocyte impurities or enhanced phagocytosis of damaged erythrocytes by CM in SF3B1K700E LR-MDS. Downregulated genes were significantly enriched in genes related to cytokine signaling, including cytokine receptors (e.g., IL6R, IL10RA, IL7R, TNFRSF1A), TREM1, signaling kinases (e.g., MAP3K7, MAP3K8, PIK3CG), and NF-κB signaling modulators (e.g., NFKBIB, IKBKG, RELA/B) (Fig. 2A). Ingenuity pathway analysis (IPA) of DEG identified enriched pathways pertaining to inflammatory cytokine signaling (i.e. NF-κB signaling, IL-6 signaling, acute phase response signaling, PI3K/AKT signaling) and inflammatory conditions (i.e. hepatic fibrosis signaling pathway) that could be affected in SF3B1K700E CM (Fig. 2B). Expression levels of the NF-κB targets IL1B and TNF were, however, variable between individual patients (Supplementary Fig. S8). Notably, IPA-based analysis of DEG in CM from SF3B1wt LR-MDS patients, of whom 2 out of 3 carried somatic mutations in TET2, brought to the fore different inflammatory pathways predicted to be more active compared to HD (Fig. 2B).
In addition, we analyzed alternative splicing in SF3B1K700E versus HD CM using rMATS (Supplementary Table S6). Among the more robust differentially spliced genes (DSG) were various genes previously reported as mis-spliced in SF3B1mut cells, such as BRD9, COASY, and TMEM214 (Fig. 2C, Supplementary Table S6). We could also confirm the previously reported cryptic 3’ splice site for MAP3K7 predicted to undergo nonsense-mediated RNA decay [14], along with decreased MAP3K7 transcript levels in SF3B1K700E CM (Supplementary Fig. S9, Table S5). We did not observe a clear association of the longer IRAK4 isoform with SF3B1K700E (Supplementary Fig. S9), as has been reported previously [15]. DSG were enriched in genes involved in the regulation of defense response and cytokine signaling, next to mRNA metabolism, apoptotic signaling, and mitotic cell cycle (Fig. 2C, Supplementary Table S7). Importantly, 369 out of the 834 DSG were also differentially spliced in SF3B1K700E versus SF3B1wt LR-MDS CM (Supplementary Table S8).
Based on the RNA-seq data pointing to dysregulated cytokine signaling in SF3B1K700E CM, we then assessed their cytokine secretion following in vitro stimulation with the Toll-like receptor 4 agonist lipopolysaccharide (LPS). We found that CM with a heterozygous mutation in SF3B1 (VAF ∼ 0.4) responded to LPS stimulation with adequate secretion of pro- (TNF, IL-1β, IL-6, IP-10, MCP-1) and anti-inflammatory cytokines (IL-10, IL-1RA) (Fig. 2D), except for one patient with an extremely high mutation burden (VAF = 0.86) (Fig. 2D). This patient exhibited the highest basal mRNA levels of TNF and IL-6 (Supplementary Fig. S8), which only marginally increased with LPS stimulation. Altogether, at the level of secreted cytokines, we did not observe markedly hyperactivated NF-κB signaling in SF3B1K700E CM following LPS exposure, although we can confirm mis-splicing and reduced mRNA expression of MAP3K7, previously linked to enhanced NF-kB activity [14]. In interpreting our findings, it is important to acknowledge the small sample size for functional assays as a limitation of our study. Therefore, further research with larger sample sizes will be required to address the functional and stimulation context-dependent deficits resulting from mis-splicing of the identified genes.
Phenotypically, the HLA-DRlow/neg CM resemble monocytic myeloid-derived suppressor cells (M-MDSCs), but markers associated with M-MDSC biology were not enriched in SF3B1K700E CM (Supplementary Fig. S8). However, HLA-DRlow/neg CM from one SF3B1K700E LR-MDS patient with co-mutations in TET2 and DNMT3A had a less stimulatory effect on the proliferative capability of autologous CD4+ T cells compared to their HLA-DRhigh counterparts (Supplementary Fig. S10). In light of this, the conversion to HLA-DRlow/neg CM may prevent excessive inflammatory reactions in the tissue driven partly by disproportionate T-cell activation. Further studies comparing HLA-DRlow/neg and HLA-DRhigh CM from patients with an isolated SF3B1K700E mutation will help to clarify their respective roles in the inflammation process.
Data availability
RNA-seq data are publicly available at GEO under accession number GSE236535.
References
Huber S, Haferlach T, Meggendorfer M, Hutter S, Hoermann G, Baer C, et al. SF3B1 mutated MDS: Blast count, genetic co-abnormalities and their impact on classification and prognosis. Leukemia. 2022;36:2894–902.
Lieu YK, Liu Z, Ali AM, Wei X, Penson A, Zhang J, et al. SF3B1 mutant-induced missplicing of MAP3K7 causes anemia in myelodysplastic syndromes. Proc Natl Acad Sci USA. 2022;119:e2111703119.
Clough CA, Pangallo J, Sarchi M, Ilagan JO, North K, Bergantinos R, et al. Coordinated missplicing of TMEM14C and ABCB7 causes ring sideroblast formation in SF3B1-mutant myelodysplastic syndrome. Blood. 2022;139:2038–49.
Mortera-Blanco T, Dimitriou M, Woll PS, Karimi M, Elvarsdottir E, Conte S, et al. SF3B1-initiating mutations in MDS-RSs target lymphomyeloid hematopoietic stem cells. Blood. 2017;130:881–90.
Platzbecker U, Germing U, Götze KS, Kiewe P, Mayer K, Chromik J, et al. Luspatercept for the treatment of anaemia in patients with lower-risk myelodysplastic syndromes (PACE-MDS): a multicentre, open-label phase 2 dose-finding study with long-term extension study. Lancet Oncol. 2017;18:1338–47.
Stahl M, DeVeaux M, De Witte T, Neukirchen J, Sekeres MA, Brunner AM, et al. The use of immunosuppressive therapy in MDS: clinical outcomes and their predictors in a large international patient cohort. Blood Adv. 2018;2:1765–72.
Zhang Q, Haider M, Al Ali NH, Lancet JE, Epling-Burnette PK, List AF, et al. SF3B1 mutations negatively predict for response to immunosuppressive therapy in myelodysplastic syndromes. Clin Lymphoma Myeloma Leuk. 2020;20:400–406.e2.
Duetz C, Westers TM, in ’t Hout FEM, Cremers EMP, Alhan C, Venniker-Punt B, et al. Distinct bone marrow immunophenotypic features define the splicing factor 3B subunit 1 (SF3B1)-mutant myelodysplastic syndromes subtype. Br J Haematol. 2021;193:798–803.
Malcovati L, Stevenson K, Papaemmanuil E, Neuberg D, Bejar R, Boultwood J, et al. SF3B1-mutant MDS as a distinct disease subtype: a proposal from the International Working Group for the Prognosis of MDS. Blood. 2020;136:157–70.
Schneider M, Rolfs C, Trumpp M, Winter S, Fischer L, Richter M, et al. Activation of distinct inflammatory pathways in subgroups of LR-MDS. Leukemia. 2023;37:1709–18.
van Leeuwen-Kerkhoff N, Westers TM, Poddighe PJ, de Gruijl TD, Kordasti S, van de Loosdrecht AA. Thrombomodulin-expressing monocytes are associated with low-risk features in myelodysplastic syndromes and dampen excessive immune activation. Haematologica. 2020;105:961–70.
Mengos AE, Gastineau DA, Gustafson MP. The CD14+HLA-DrlO/NEG monocyte: an immunosuppressive phenotype that restrains responses to cancer immunotherapy. Front Immunol. 2019;10:1147.
Todisco G, Moura PL, Hellström-Lindberg E. Clinical manifestations of clonal hematopoiesis: what has SF3B1-mutant MDS taught us? Semin Hematol. 2022;59:150–5.
Lee SCW, North K, Kim E, Jang E, Obeng E, Lu SX, et al. Synthetic lethal and convergent biological effects of cancer-associated spliceosomal gene mutations. Cancer Cell. 2018;34:225–241.e8.
Choudhary GS, Pellagatti A, Agianian B, Smith MA, Bhagat TD, Gordon-Mitchell S, et al. Activation of targetable inflammatory immune signaling is seen in myelodysplastic syndromes with SF3B1 mutations. Elife. 2022;11:e78136.
Acknowledgements
We would like to thank Ivonne Habermann (University Hospital Carl Gustav Carus, Dresden, Germany) for excellent technical support and all technicians in the flow cytometry laboratories for processing samples and acquiring raw data. We would also like to thank Dr. Sebastian Stasik (Department of Internal Medicine I, University Hospital Carl Gustav Carus, Faculty of Medicine, TU Dresden, Dresden, Germany) for help with variant calling from RNA-seq data, Caroline E. Roe (Mass Cytometry Center of Excellence, Vanderbilt University, Nashville, TN, USA) for help with the T-REX pipeline, and Dr. Falk Heidenreich (Department of Internal Medicine I, University Hospital Carl Gustav Carus, Faculty of Medicine, TU Dresden) for constructive feedback and discussions. Moreover, we would like to thank Dr. Jörg Lehmann and Claudia Müller for providing and assisting with the FLEXMAP 3D™ system (Fraunhofer Institute for Cell Therapy and Immunology, Leipzig, Germany). We also thank the King’s College London research facilities and the University College London/Department of Computer Science HPC facility for support. We would also like to thank all the patients for contributing to this study. This work was supported by the transCampus funding program between King’s College London and TU Dresden (UP), the DKTK partner site Dresden (UP), the European Union - Transcan 7 Horizon 2020 - EuroMDS project #20180424 (MGDP), the Next Generation EU - NRRP M6C2 - Investment 2.1 Enhancement and strengthening of biomedical research in the NHS (MGDP), and the AIRC Foundation (Associazione Italiana per la Ricerca contro il Cancro, Milan, Italy) projects #22053 (to MGDP) and #26537 (to VS). SK and RADR are supported by the CRUK City of London Centre Award [CTRQQR-2021/100004] at KCL.
Author information
Authors and Affiliations
Contributions
SW, MSchn, UO, GM, ER, MGR, SB, BP, DC, RADR, JAT, NS, and AT contributed to the collection of data. UO analyzed clinical FCM data of the experimental cohort. ER, MGR, SB, and BP analyzed FCM data of the validation cohorts. DC performed bioinformatics analysis of NanoString and RNA-seq data. EG and CL analyzed and visualized RNA-seq data. SW and MSchn analyzed and visualized data. SW, MSchn, JV, KS, AP, MGDP, VS, MSchm, and SK interpreted the results. UP and SK designed the project. UP and KS recruited MDS patients and contributed to clinical care. SW and SK conceptualized the work and wrote and edited the manuscript. All authors were involved in the review of the work and approval of the final version of the manuscript.
Corresponding author
Ethics declarations
Competing interests
SW, MS, UO, GM, ER, MGR, SB, BP, DC, RADR, EG, CL, JV, JT, NS, KS, AP, AT, MGDP, VS, MS, and UP declare no competing financial interests. SK has received research support and honoraria from Novartis (advisory board, speakers bureau), Alexion (speakers bureau), Beckman Coulter (speakers bureau), MorphoSys (research support), and Pfizer (speakers bureau). None of these are relevant to the current work.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Winter, S., Schneider, M., Oelschlaegel, U. et al. Mutations in the splicing factor SF3B1 are linked to frequent emergence of HLA-DRlow/neg monocytes in lower-risk myelodysplastic neoplasms. Leukemia 38, 1427–1431 (2024). https://doi.org/10.1038/s41375-024-02249-z
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/s41375-024-02249-z