Abstract
The transcription factor ThPOK (encoded by Zbtb7b) is well known for its role as a master regulator of CD4 lineage commitment in the thymus. Here, we report an unexpected and critical role of ThPOK as a multifaceted regulator of myeloid lineage commitment, differentiation and maturation. Using reporter and knockout mouse models combined with single-cell RNA-sequencing, progenitor transfer and colony assays, we show that ThPOK controls monocyte-dendritic cell versus granulocyte lineage production during homeostatic differentiation, and serves as a brake for neutrophil maturation in granulocyte lineage-specified cells through transcriptional regulation of lineage-specific transcription factors and RNA via altered messenger RNA splicing to reprogram intron retention.
This is a preview of subscription content, access via your institution
Access options
Access Nature and 54 other Nature Portfolio journals
Get Nature+, our best-value online-access subscription
$29.99 / 30 days
cancel any time
Subscribe to this journal
Receive 12 print issues and online access
$209.00 per year
only $17.42 per issue
Rent or buy this article
Prices vary by article type
from$1.95
to$39.95
Prices may be subject to local taxes which are calculated during checkout
Similar content being viewed by others
Data availability
scRNA-seq and CITE-seq raw and processed data/metadata are deposited as GEO accession number GSE163844. There are no data restrictions. Source data are provided with this paper.
References
Netherby, C. S. et al. The granulocyte progenitor stage is a key target of IRF8-mediated regulation of myeloid-derived suppressor cell production. J. Immunol. 198, 4129–4139 (2017).
Yanez, A., Goodridge, H. S., Gozalbo, D. & Gil, M. L. TLRs control hematopoiesis during infection. Eur. J. Immunol. 43, 2526–2533 (2013).
Manz, M. G. & Boettcher, S. Emergency granulopoiesis. Nat. Rev. Immunol. 14, 302–314 (2014).
Yanez, A., Megias, J., O’Connor, J. E., Gozalbo, D. & Gil, M. L. Candida albicans induces selective development of macrophages and monocyte derived dendritic cells by a TLR2 dependent signalling. PLoS ONE 6, e24761 (2011).
Basu, S. et al. ‘Emergency’ granulopoiesis in G-CSF-deficient mice in response to Candida albicans infection. Blood 95, 3725–3733 (2000).
Serbina, N. V., Hohl, T. M., Cherny, M. & Pamer, E. G. Selective expansion of the monocytic lineage directed by bacterial infection. J. Immunol. 183, 1900–1910 (2009).
Olsson, A. et al. Single-cell analysis of mixed-lineage states leading to a binary cell fate choice. Nature 537, 698–702 (2016).
He, X. et al. The zinc finger transcription factor Th-POK regulates CD4 versus CD8 T-cell lineage commitment. Nature 433, 826–833 (2005).
He, X. et al. CD4-CD8 lineage commitment is regulated by a silencer element at the ThPOK transcription-factor locus. Immunity 28, 346–358 (2008).
Park, K. et al. TCR-mediated ThPOK induction promotes development of mature (CD24−) γδ thymocytes. EMBO J. 29, 2329–2341 (2010).
Engel, I., Zhao, M., Kappes, D., Taniuchi, I. & Kronenberg, M. The transcription factor Th-POK negatively regulates Th17 differentiation in Vα14i NKT cells. Blood 120, 4524–4532 (2012).
Lee, H. O. et al. Disregulated expression of the transcription factor ThPOK during T-cell development leads to high incidence of T-cell lymphomas. Proc. Natl Acad. Sci. USA 112, 7773–7778 (2015).
Evrard, M. et al. Developmental analysis of bone marrow neutrophils reveals populations specialized in expansion, trafficking, and effector functions. Immunity 48, 364–379 (2018).
Xie, X. et al. Single-cell transcriptome profiling reveals neutrophil heterogeneity in homeostasis and infection. Nat. Immunol. 21, 1119–1133 (2020).
Blazek, K. et al. IFN-λ resolves inflammation via suppression of neutrophil infiltration and IL-1β production. J. Exp. Med. 212, 845–853 (2015).
Shrum, B. et al. A robust scoring system to evaluate sepsis severity in an animal model. BMC Res. Notes 7, 233 (2014).
Muench, D. E. et al. Mouse models of neutropenia reveal progenitor-stage-specific defects. Nature 582, 109–114 (2020).
Kwok, I. et al. Combinatorial single-cell analyses of granulocyte-monocyte progenitor heterogeneity reveals an early uni-potent neutrophil progenitor. Immunity 53, 303–318.e5 (2020).
Amir el, A. D. et al. viSNE enables visualization of high dimensional single-cell data and reveals phenotypic heterogeneity of leukemia. Nat. Biotechnol. 31, 545–552 (2013).
Yamamoto, R. et al. Clonal analysis unveils self-renewing lineage-restricted progenitors generated directly from hematopoietic stem cells. Cell 154, 1112–1126 (2013).
DePasquale, E. A. K. et al. cellHarmony: cell-level matching and holistic comparison of single-cell transcriptomes. Nucleic Acids Res. 47, e138 (2019).
La Manno, G. et al. RNA velocity of single cells. Nature 560, 494–498 (2018).
Nutt, S. L. & Chopin, M. Transcriptional networks driving dendritic cell differentiation and function. Immunity 16, 942 (2020).
Sunami, Y. et al. BCL11A promotes myeloid leukemogenesis by repressing PU.1 target genes. Blood Adv. 6, 1827–1843 (2022).
Gruber, T. A. et al. An Inv(16)(p13.3q24.3)-encoded CBFA2T3-GLIS2 fusion protein defines an aggressive subtype of pediatric acute megakaryoblastic leukemia. Cancer Cell 22, 683–697 (2012).
Masetti, R., Bertuccio, S. N., Pession, A. & Locatelli, F. CBFA2T3-GLIS2-positive acute myeloid leukaemia. A peculiar paediatric entity. Br. J. Haematol. 184, 337–347 (2019).
Saha, S. et al. Transcriptomic analysis identifies RNA binding proteins as putative regulators of myelopoiesis and leukemia. Front. Oncol. 9, 692 (2019).
Diaz-Munoz, M. D. & Turner, M. Uncovering the role of RNA-binding proteins in gene expression in the immune system. Front. Immunol. 9, 1094 (2018).
Wong, J. J. et al. Orchestrated intron retention regulates normal granulocyte differentiation. Cell 154, 583–595 (2013).
Itskovich, S. S. et al. MBNL1 regulates essential alternative RNA splicing patterns in MLL-rearranged leukemia. Nat. Commun. 11, 2369 (2020).
Skoda, R. C. & Schwaller, J. Dual roles of EZH2 in acute myeloid leukemia. J. Exp. Med. 216, 725–727 (2019).
Sumter, T. F. et al. The High Mobility Group A1 (HMGA1) transcriptome in cancer and development. Curr. Mol. Med. 16, 353–393 (2016).
Iervolino, A. et al. hnRNP A1 nucleocytoplasmic shuttling activity is required for normal myelopoiesis and BCR/ABL leukemogenesis. Mol. Cell. Biol. 22, 2255–2266 (2002).
Hodson, D. J., Screen, M. & Turner, M. RNA-binding proteins in hematopoiesis and hematological malignancy. Blood 133, 2365–2373 (2019).
Basu, J. et al. Essential role of a ThPOK autoregulatory loop in the maintenance of mature CD4+ T cell identity and function. Nat. Immunol. 22, 969–982 (2021).
Sun, F., Xiao, G. & Qu, Z. Murine bronchoalveolar lavage. Bio Protoc. 7, e2287 (2017).
Gonçalves, R. & Mosser, D. M. The isolation and characterization of murine macrophages. Curr. Protoc. Immunol. 111, 14.1.1–14.1.16 (2015).
Venkatasubramanian, M., Chetal, K., Schnell, D. J., Atluri, G. & Salomonis, N. Resolving single-cell heterogeneity from hundreds of thousands of cells through sequential hybrid clustering and NMF. Bioinformatics 36, 3773 (2020).
Muench, D. E. et al. SKI controls MDS-associated chronic TGF-β signaling, aberrant splicing, and stem cell fitness. Blood 132, e24–e34 (2018).
Muller, P. A. et al. Microbiota modulate sympathetic neurons via a gut-brain circuit. Nature 583, 441–446 (2020).
Esterházy, D. et al. Compartmentalized gut lymph node drainage dictates adaptive immune responses. Nature 569, 126–130 (2019).
Acknowledgements
This work was supported by NIH grant nos. R01 AI068907 (D.J.K.), R01 GM107179 (D.J.K.), R21 CA195356 (D.J.K.), R21AI164333 (J.B.), R01 HL122661 (H.L.G.) and P30 CA006927 (FCCC Comprehensive Cancer Center Core Grant). We are thankful for the assistance of M. B. Einarson in confocal imaging (R50 CA211479). We acknowledge the assistance of the following core facilities of the Fox Chase Cancer Center: Transgenic Mouse, Flow Cytometry, Cell Culture, DNA Sequencing, Genomic and Laboratory Animal facilities. We thank D. L. Wiest, G. Rall and E. Koltsova for critical comments and suggestions on the manuscript. We thank W. Baldwin III, for providing expertise with analysis of neutrophil morphology. We thank C. E. Canty for logistical support during the revision.
Author information
Authors and Affiliations
Contributions
J.B. and D.J.K. conceived the idea. J.B., D.J.K. and H.L.G. designed the experiments. J.B., A.O., K.F., E.K.T., K.C., E.N., P.C., D.L., L.G., X.H., N.S. and D.J.K. performed research. J.B., N.S., H.L.G. and D.J.K. analyzed the data. J.B. and D.J.K. wrote the manuscript. N.S. and H.L.G. reviewed and edited the manuscript.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Immunology thanks Renato Ostuni, Roxane Tussiwand and Andres Hidalgo for their contribution to the peer review of this work. L. A. Dempsey was the primary editor on this article and managed its editorial process and peer review in collaboration with the rest of the editorial team.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Extended data
Extended Data Fig. 1 ThPOK deficiency alters neutrophil and monocyte subset distribution.
a) Gating strategy to reveal neutrophil maturation and differentiation status, according to indicated markers. b) Cell-population percentages and absolute cell numbers of indicated neutrophil subsets from indicated organs of wt versus ThPOK-deficient mice (based on gating strategy in panel a). Data are displayed as mean ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Significant differences were determined by two-sided unpaired t test. c) Histological examination of flow sorted (Ly6g+ CD11b+) blood neutrophils. Cytospins of neutrophils from WT or ThPOK-/- mice were treated with Wright’s stain and examined with a light microscope. The experiment was repeated 3 times independently with similar results. d) FACS analysis of CD11b+ Ly6g-CD115+ F4/80- cells from indicated organs from WT and ThPOK-/- mice (left panels). Note increase in Ly6c+ classical monocytes in ThPOK-deficient compared to WT mice in BM. Plots showing total number of indicated gated populations in indicated organ in ThPOK-deficient compared to WT mice (right panels; same mice as in FACS panels). Error bars represent standard deviations. e) Plots showing frequency (top panels) or cell number (bottom panels) of T cells, NK cells, B cells, and DCs in BM, blood and spleen of ThPOK-deficient or WT mice. Significant differences between ThPOK-/- and WT mice were determined by unpaired T test with Welchs correction, and indicated by asterisks (* p < 0.05; ** p < 0.01; *** p < 0.001).
Extended Data Fig. 2 ThPOK deficiency perturbs selective neutrophil functions.
a) Neutrophil surface marker expression (n = 3) (Top 1st panel from left), phagocytosis [data expressed as geometric Mean Fluorescent Intensity (MFI) ± SEM (n = 3)] (Top 2nd panel), PMA-mediated ROS production using dihydrorhodamine 123 (DHR) [data are shown as geometric means ± SEM (n = 4)] (Top 3rd panel), Myeloperoxidase release (n = 5) (Top 4th panel), NET associated elastase release (n = 5 for wt and n = 3 for ThPOK-/- neutrophil, respectively) (Bottom left panel), and cytokine/chemokine release upon in vitro stimulation (N = 6) (Bottom right panel). Data are representative of three independent experiments. Significance was calculated by two-sided unpaired t-test with Welch’s correction (*p < 0.05; ***p < 0.001). b) Absolute numbers of total neutrophils or indicated subsets from airpouch lavage of wt versus ThPOK-deficient mice, 14hrs after injection of E. coli or PBS into airpouch (top left, top 2nd from left, and bottom left panels). Same mice were used to assess phagocytosis by pHrodo™ green E. coli incorporation [geometric MFI ± SEM (n = 3)] (bottom 2nd from left panel), ROS production [geometric means ± SEM (n = 3)] (top right panel), Myeloperoxidase release (n = 3) (bottom right panel), NET associated elastase release (n = 3) (center right panel), and Citrullinated H3 (CitH3) release (Top 2nd from right panel) for airpouch lavage, serum and plasma samples from the same mice. Significance was calculated by two-sided unpaired t-test with Welch’s correction (*p < 0.05; **p < 0.01; ***p < 0.001). c) Frequency (bottom panels) or cell number (top panels) of indicated subsets in WT or ThPOK-deficient mice 16hrs after sepsis induction or normal saline administration. Significant differences were determined by one way ANOVA posthoc Bonferroni’s multiple comparison test (*p < 0.05; **p < 0.01; ***p < 0.001). d) Same mice as in panel c were assessed for ROS production [geometric means ± SEM (n = 4)] (top right panel), Myeloperoxidase release (n = 4) (bottom left panel), NET associated elastase release (n = 4) (center left panel), and Citrullinated H3 (CitH3) release (top left panel). Assays were performed for blood serum and plasma samples from the same mice (top, middle and bottom panels in center). e) Cytokine/chemokine levels in peritoneal lavage, blood serum and plasma of ThPOK-/- and WT mice 14hrs after sepsis induction (N = 4). Significance was calculated by one way ANOVA posthoc Bonferroni’s multiple comparisons test (* p < 0.05; **p < 0.01; ***p < 0.001; **** p < 0.0001). f) H&E staining of tissue sections from indicated organ 16hrs after sepsis induction. Images shown are representative of 6 mice/genotype. The experiment was repeated 3 times with similar results.
Extended Data Fig. 3 ThPOK expression and effect of ThPOK deficiency on gene expression.
a) GFP expression by indicated subsets from ThPOK-GFP reporter mice. GFP expression level is color-coded according to heat-scale below each plot. b) Gating strategy to reveal granulocyte-monocyte lineage development (according to REF. 17). Red arrows indicate sequential gates. c) Unsupervised clustering (viSNE) of cells in the modified-GMP gate (upper left). Other panels show expression of indicated surface marker superimposed on the same clusters. d) GFP protein (top), GFP mRNA (center), and endogenous ThPOK mRNA levels, in indicated progenitor subsets sorted according to gating strategy in (b), as well as indicated control T cell populations from transgenic ThPOK-GFP reporter mice. e) Competitive reconstitution assay with WT and ThPOK − /− BM cells (same experiment as depicted in Fig. 1j) (n = 6, per genotype). Relative contribution to indicated BM populations was assessed at 6 months post-transplant. Data are normalized for only donor cells, so that the sum of WT and KO donor cells equals 100%. Two sided unpaired T test with Welch’s Correction was performed. Data are displayed as mean ± SEM (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001). f) Cell cycle analyses of indicated gated subsets in WT and ThPOK -/- BM 20 h after a single in vivo BrdU pulse. Representative FACS plots of BM subsets from PBS-injected WT controls (BrdU-), or WT and ThPOK-/- mice that received BrdU, as indicated (right panels). Cell cycle stages were defined by flow cytometry based on BrDU intensity and DNA content (7-AAD staining), as indicated. Bar graphs at left summarize the results for indicated progenitor populations, expressed as means ± SEM (n = 3 independent animals per genotype) and are representative of two experiments. Two sided unpaired t-test with Welch’s Correction was performed (*p < 0.05; **p < 0.01; *** p < 0.001).
Extended Data Fig. 4 ThPOK loss does not affect MDP development.
a) Bar graphs showing absolute cell numbers for CMP, GMP, MEP and CD34lo subsets among gated cKit+ Sca1- Lin- BM cells from ThPOK-deficient or WT mice (same mice as Fig. 2a, b) (n = 5 independent animals per genotype). Error bars represent SEM. Significant differences between ThPOK-/- and WT mice were determined by two-sided unpaired T test with Welch’s correction, and indicated by asterisks (* p < 0.05; ** p < 0.01; *** p < 0.001). b) FACS analysis of CD135 and CD115 expression by gated CMPs from WT and ThPOK-/- mice (left panels). Bar graph at right shows % of indicated gated subsets (n = 5 independent animals per genotype). c) FACS analysis of CD11b, CD115, Ly6c, and Ly6g expression by gated cGMP (conventional GMP) (grey histograms) and c-Kit+ CD34lo CD16hi Lin- (green histograms) BM subsets from WT mice. d) FACS analysis of CD16/32 and CD34 expression by gated Lin- c-Kit+ Sca1- BM cells from ThPOPK+/- mice. CMP, GMP, MEP and CD34lo subsets are marked. e) CFU assay of FACS-sorted Lin- BM cells, from indicated WT or ThPOK-deficient mice. Error bars represent SEM. Significant differences between ThPOK-/- and WT mice were determined by two-sided multiple multiple unpaired T with Welch’s correction, and indicated by asterisks (* p < 0.05; ** p < 0.01; *** p < 0.001). Note that ThPOK-/- Lin- progenitors exhibited a substantial (** p < 0.01) increase in CFU-GM myeloid colony production even after secondary plating. f) FACS analysis of CD11b, Thy1, CD41 and Ter119 lineage marker expression by WT and ThPOK-deficient cells after secondary CFU assay (same experiment as in panel d).
Extended Data Fig. 5 Single-cell transcriptome analysis of ThPOK-/- and WT BM populations.
a) UMAP plot of curated cKit+ progenitor scRNA-Seq cell populations that serves as the reference for cellHarmony alignment analyses in these studies (see Methods), b) Flow cytometry selection of c-Kit+ BM C-GMP progenitors (gated cells denoted in blue). c) Unsupervised clustering UMAP plot of ~26,000 combined C-GMP CITE-Seq captured mRNA profiles (WT and ThPOK-/-) following analysis with the software ICGS2. Indicated cell-population labels are those automatically assigned by ICGS2 (AltAnalyze BioMarker database). d) UMAP plot of all WT and ThPOK-/- cells from panel c aligned to the reference described in panel a. e) UMAP plot from panel d showing the expression levels of ThPOK mRNA in the wild-type CITE-Seq populations. f) Cell-population percentage of distinct BM progenitor populations detected by CITE-Seq from WT and ThPOK-/- mice. g, h) Heatmap of relative normalized (TotalVI) CITE-Seq antibody derived-tag (ADT) intensities for cellHarmony cKit+ aligned cell populations. Panels (g) and (h) displays cells from WT and ThPOK-/- BM progenitors, respectively. i-p) Gene Ontology enrichment analyses from the software GO-Elite, for each of the indicated cell population differential expression analysis comparisons (all ThPOK-/- vs. WT).
Extended Data Fig. 6 ThPOK−/− impacts global splicing and cellular process networks.
(a, b) UMAP plot of single-cell Fluidigm RNA-Seq analysis of prior profiled (a) wild-type hematopoietic progenitors and (b) Ly6c-ThPOK-/- GMPs aligned to cKit+ CITE-Seq. Consistent alignment of cell populations in the UMAP space indicates a lack of apparent batch effects. c) Gene Ontology enrichment analysis of ThPOK-/- dependent alternative splicing events in proNeu1 cells, for splicing events with the opposite pattern of exon inclusion in proNeu1 versus MultiLin (discordant = more immature splicing) or (d) with the same pattern (concordant = promoting neutrophil specification associated splicing). (e-f) Validation of deregulated gene expression in ThPOK-/- versus WT progenitors for important transcripts as observed in CITE seq data: e) qRTPCR analyses of indicated flow-sorted progenitors (pooled from 4 mice/genotype) for monocyte/DC genes (Irf8, Zeb2, Runx1), granulocyte lineage genes (Cebpe, Pde4d, Cd63 and Nkg7) and RNA binding protein genes (Ddx3x and Srsf5). The experiment was repeated 3 times independently with similar results. f) Expression ratio of Hmga1 alternate transcript versus reference transcript in indicated progenitor populations (pooled from 4 mice/genotype). g) Representative confocal micrographs of EZH2 protein (top), DAPI (middle), and H3K27Me3 (bottom) staining in indicated subsets. Each experiment was reproduced twice and significant differences between ThPOK-/- and WT mice were determined by two-sided unpaired T test with Welch’s correction, and indicated by asterisks (* p < 0.05; ** p < 0.01; *** p < 0.001). h) Heat maps showing relative mRNA expression of indicated genes in WT progenitors (left panel), or in ThPOK-/- versus WT progenitor subsets (right panel) for genes that are relatively up-regulated in any ThPOK-/- progenitor population and that are known Ezh2 targets in WT immune cells by ChIP-seq (GSE181873). Note that in WT mice most of these genes show marked stage-specific regulation.
Extended Data Fig. 7 ThPOK-/- progenitor transfer assays.
a) Gating strategy for isolation of GMPs and MDPs for in vitro and in vivo transfer studies. b) FACS analysis of Ly6g, CD11b, F4-80, CD115, Ly6c, and CD43 in blood of mice reconstituted with sorted GMPs from BM of WT or ThPOK-/- mice. Note substantial population of F4-80hi CD11b+ Ly6g- CD115+ macrophages in ThPOK-deficient versus WT mice (middle panels).
Extended Data Fig. 8 ThPOK-/- MDP exhibit neutrophil differentiation bias in vitr.
a) CFU assay of FACS-sorted CMP (Lin- Kit+ Sca-CD34+ CD16/32-Flt3+ CD115-) cells from WT and ThPOK KO mice, as indicated. Centre line refers to mean. b) Representative images of colony morphology after culture of WT or ThPOK KO MDPs in methyl cellulose (left), or of dissociated cells after Giemsa staining (right). Note increased neutrophil frequency in ThPOK-/- versus WT MDP cultures. c) Bar graphs indicate absolute cell number of indicated donor-derived gated myeloid subsets (n = 5 independent animals) in blood after in vivo MDP transfer (gated as in Fig. 6f). Error bars represent standard deviations (centre line refers to mean). Significant differences between ThPOK-/- and WT mice were determined by two-sided unpaired t test with Welch’s correction, and indicated by asterisks (* p < 0.05; ** p < 0.01; *** p < 0.001).
Supplementary information
Supplementary Table 1
Single-cell RNA-seq dataset metrics.
Supplementary Table 2
Pairwise statistical comparison of ThPOK−/− and wild-type cell-state.
Supplementary Table 3
Genes ThPOK−/− versus WT up-regulated genes in neutrophils; and genes ThPOK−/− versus WT downregulated genes in DC subsets.
Supplementary Table 4
ThPOK−/− KO ChIP–seq annotated peaks.
Supplementary Table 5
Pairwise statistical comparison of alternative splicing in differentiation states.
Supplementary Table 6
Pairwise statistical comparison of ThPOK−/− and wild-type cell-state alternative splicing.
Supplementary Table 7
143 genes overlapping alternate splicing and any differentially expressed genes in any cell-type ThPOK−/− versus wild-type up-/downregulated genes.
Supplementary Table 8
Antibody list (Methods).
Source data
Source Data Fig. 1
Statistical source data.
Source Data Fig. 2
Statistical source data.
Source Data Fig. 4
Statistical source data.
Source Data Fig. 5
Statistical source data.
Source Data Fig. 6
Statistical source data.
Source Data Extended Data Fig. 1
Statistical source data.
Source Data Extended Data Fig. 2
Statistical source data.
Source Data Extended Data Fig. 3
Statistical source data.
Source Data Extended Data Fig. 4
Statistical source data.
Source Data Extended Data Fig. 6
Statistical source data.
Source Data Extended Data Fig. 8
Statistical source data.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Basu, J., Olsson, A., Ferchen, K. et al. ThPOK is a critical multifaceted regulator of myeloid lineage development. Nat Immunol 24, 1295–1307 (2023). https://doi.org/10.1038/s41590-023-01549-3
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/s41590-023-01549-3
