Loss-of-function Dnmt3a mutation leads to aberrant neutrophil migration

Abstract

Clonal hematopoiesis (CH), an age-related expansion of somatically mutated hematopoietic clones, is associated with increased risk of severe infections including COVID-19, yet the underlying mechanisms remain unclear. Here, we investigated the impact of Dnmt3a deficiency in a murine model of influenza A virus (IAV) pneumonia.

Dnmt3a-deficient mice exhibited increased pulmonary viral burden and reduced neutrophil accumulation in IAV-infected lungs despite comparable circulating neutrophil numbers. Functional analyses of neutrophils showed impaired chemotactic migration in vitro, while maturation, antimicrobial enzyme content, and metabolic capacity were unchanged. Transcriptomic profiling revealed downregulation of pathways involved in chemotaxis, cytokine signaling, and cellular activation, including reduced expression of Cxcr1. Supporting the translational relevance of these findings, proteomic analysis of plasma from individuals with germline DNMT3A mutations (Tatton-Brown-Rahman syndrome) revealed alterations in proteins associated with cell migration and cytoskeletal dynamics.

Collectively, our findings demonstrate that Dnmt3a loss compromises innate immune defense by impairing neutrophil migration in a cell-intrinsic manner, leading to ineffective pathogen clearance. This work provides mechanistic insight into how CH-associated mutations contribute to age-associated susceptibility to infection and highlights altered leukocyte trafficking as a potential therapeutic target in aging populations with CH.

Graphical Abstract

eTOC blurb

Zimmer et al report that loss of Dnmt3a disrupts innate immunity by impairing neutrophil chemotaxis. Dnmt3a deficiency reduces neutrophil recruitment to influenza-infected lungs, increasing viral burden. Multiomic analyses in mice and humans reveal downregulated chemotaxis pathways, linking clonal hematopoiesis–associated DNMT3A mutations to defective host defense.

Introduction

Infections remain a leading cause of morbidity, hospital admission, and healthcare expenditure worldwide, and sepsis is among the most common causes of death in intensive care units [1, 2]. Despite decades of research, sepsis mortality has improved only modestly, in part due to incomplete understanding of how host-intrinsic factors predispose to immune dysregulation [3], and how these factors change with age, a well-established risk factor for sepsis [4]. Emerging epidemiological data suggests that age-acquired mutations in the hematopoietic system may represent key, previously unrecognized drivers of sepsis susceptibility and outcome.

Clonal hematopoiesis (CH) is an age-associated condition in which hematopoietic stem and progenitor cells (HSPC) acquire somatic mutations that confer a competitive advantage, leading to the expansion of clonal populations of blood cells in otherwise healthy individuals [5]. The most frequently mutated genes encode the epigenetic regulators DNA Methyltransferase 3 Alpha (DNMT3A), Tet Methylcytosine Dioxygenase 2 (TET2) and Additional Sex Combs-Like 1 (ASXL1) [5]. Large sequencing studies have shown that detectable CH is rare before midlife but rises sharply with age, with reported incidences ranging from 10–20% to over 75% in individuals ≥70 years depending on sequencing depth [6, 7].

Initially recognized as a precursor state for hematologic malignancies, CH is now appreciated as a systemic risk factor for non-malignant, age-related diseases. Individuals with CH exhibit higher all-cause mortality, and CH has been linked to diverse chronic inflammatory conditions, including cardiovascular disease, chronic kidney disease, and neurodegenerative disorders such as Alzheimer’s disease [8, 9].

Recent epidemiologic data has extended this concept to infectious diseases. This relationship is bidirectional: infections have recently been identified as a driver of CH [10], while on the other hand, analyses of the UK Biobank and other large cohorts demonstrate that individuals with CH have an increased risk of bacterial and viral infections and sepsis, as well as increased severity and shorter survival after infection. This was first comprehensively described in coronavirus disease 2019 (COVID-19) [11–13] but applies to a variety of infections such as pneumonia [14], periodontitis [15], and pathogens such as Clostridioides difficile, Streptococcus, and Enterococcus [11]. In parallel, mosaic chromosomal alterations (mCAs) – larger chromosomal changes – have been found to increase the risk for diverse infections including sepsis, respiratory tract infections, digestive system infections, and genitourinary system infections, and for hospitalization due to COVID-19 [16]. Together, these findings position CH as a major, age-related determinant of host defense.

Mechanistic insights into how individual CH driver mutations influence host defense remain limited. Recent studies have begun to elucidate how loss of TET2 mutations affect innate immune responses during acute infection, especially by impairing neutrophil development and function [14, 17]. In contrast, despite being the most prevalent CH driver mutation, the impact of loss of function DNMT3A mutations on immune cell function during infection has not been systematically investigated. This gap is especially notable given evidence that loss of function DNMT3A mutations, but not TET2, are associated with reduced survival in COVID-19 [13].

We hypothesized that loss of function DNMT3A mutations compromise effective host defense, thereby increasing susceptibility to infection. To test this hypothesis, we employed a Dnmt3a knockout mouse model and challenged mutant and control animals with acute influenza A virus (IAV) pneumonia. Using this approach, we sought to define how DNMT3A-dependent CH influences immune responses during viral infection and to identify the cellular pathways responsible for heightened disease susceptibility. This study demonstrates that Dnmt3a-deficient HSPCs produce neutrophils with defects in migration, thereby diminishing their ability to clear pathogens.

Methods

All procedures were performed in compliance with relevant laws and institutional guidelines. Detailed methods, materials, protocols, and references are provided in the Suppl. Data.

Mice.

Dnmt3a^fl/fl mice were generously provided by Dr. Margaret A. Goodell and crossed to Vav-iCre (JAX Stock No: 008610) [18–21] or Mx1-Cre mice (JAX Stock No: 003556) [22], all on a C57BL/6 background. Cre-negative littermates Dnmt3a^fl/fl;Vav-Cre⁻ and Dnmt3a^fl/fl;Mx1-Cre⁻ were used as wild-type (WT) controls. Dnmt3a deletion was induced in Dnmt3a^fl/fl;Mx1-Cre⁺ (Dnmt3a^Mx1) and Dnmt3a^fl/fl;Mx1-Cre⁻ mice by intraperitoneal injection of poly(I:C) (200 μg) three times weekly for two weeks and confirmed by polymerase chain reaction (PCR) four weeks later (Table S1). All mice were age- and sex-matched. Animals were housed under specific pathogen-free conditions at Baylor College of Medicine. All procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Baylor College of Medicine.

Influenza A virus (IAV) infection and viral quantification.

Influenza A/Puerto Rico/8/34 (H1N1; PR8) viral stocks were prepared and quantified as previously described [23]. For IAV pneumonia studies, mice were anesthetized and infected intranasally with 0.5× or 2× LD₅₀ (15 or 60 PFU). Naïve controls received PBS. Survival and weight loss were monitored daily. For immune and virologic analyses, mice were euthanized 40 h post-infection. Viral titers in lung homogenates were quantified by plaque assay on MDCK cells [24] or by quantitative RT-PCR targeting the PR8 nucleoprotein and normalized to lung weight [25].

Blood and lung immune profiling via flow cytometry.

Peripheral blood was collected via retroorbital bleed for complete blood counts and leukocyte isolation. Lungs were perfused and enzymatically digested, red blood cells were lysed, and cells were strained into single-cell suspensions. Blood and lung cells were stained at 4 °C (Panels see Suppl. Data). Absolute counts were normalized to lung weight.

Immunofluorescence and Immunohistochemistry.

Lungs were fixed in 4% paraformaldehyde, paraffin-embedded, sectioned, and subjected to immunohistochemistry or multiplex immunofluorescence using Opal^™ reagents (Table S2) [26, 27]. Imaging was performed on a Vectra Polaris system, with spectral unmixing and quantitative analysis conducted using inForm and QuPath software [28].

Neutrophil functional assays.

Neutrophils were elicited by intraperitoneal casein injection and enriched by magnetic negative selection [29]. Functional assays included measurement of neutrophil elastase, myeloperoxidase activity, NADPH levels, hydrogen peroxide production (± NOX2 inhibition), mitochondrial respiration by Seahorse XFp analysis, and chemotaxis using transwell assays [30, 31].

Bulk RNA sequencing.

Neutrophils were isolated by FACS-sorting from bone marrow and lungs of naïve and IAV-infected mice. RNA was extracted and sequenced using an ultra-low input pipeline. Reads were aligned to the mm10 genome, and differential gene expression was analyzed using DESeq2 [32]. Pathway enrichment was assessed using GSEA, Metascape, EnrichR, and ClusterProfiler. Data are deposited in GEO (GSE328797).

Human plasma proteomics.

Patient-derived plasma samples were collected after written informed consent under protocol H-40807 approved by the Institutional Review Board (IRB) of Baylor College of Medicine. Plasma from individuals with Tatton-Brown-Rahman-Syndrome and unaffected controls was analyzed following depletion of abundant proteins, tryptic digestion, and liquid chromatography–tandem mass spectrometry (LC–MS/MS). Protein quantification was performed using a label-free iBAQ approach, with statistical analysis conducted in R. Proteomics data are available via PRIDE (PXD077276).

Statistical analysis.

Statistical analyses were performed using GraphPad Prism v10. Details of tests and sample sizes are provided in figure legends.

Results

Loss of Dnmt3a increases viral burden in acute influenza A virus (IAV) infection.

To investigate the impact of Dnmt3a loss of function mutation on acute viral infection outcome, we used a mouse model of acute IAV pneumonia. Dnmt3a^fl/fl;Vav-Cre⁺ (Dnmt3a^Vav) and Dnmt3a^fl/fl;Vav-Cre⁻ (WT) mice were intranasally challenged with a 0.5 LD₅₀ dose of H1N1-PR8 (15 PFU) or PBS and monitored daily. Overall mortality and weight loss after infection were not significantly different in Dnmt3a^Vav mice compared to WT controls (Figure 1A–B). Since Vav-Cre insertion is known to disrupt the COMM domain containing 10 (Commd10) gene [33], a regulator of NF-kB signaling, we repeated these experiments using Dnmt3a^fl/fl;Mx1-Cre⁺ (Dnmt3a^Mx1) versus Dnmt3a^fl/fl;Mx1-Cre⁻ (WT) mice, all induced with polyinosinic:polycytidylic acid (poly(I:C)) ~1 month prior to influenza challenge. Similar to Dnmt3a^Vav mice, we found no difference in mortality or weight loss when comparing IAV infection in Dnmt3a^Mx1 versus WT mice (Figure S1A–B). Thus, homozygous knockout of Dnmt3a did not affect mortality or weight loss after IAV infection at this dose.

Figure 1. Loss of function Dnmt3a mutation increases viral burden in IAV-infected lungs.

(A-D) Influenza Challenge with PBS or IAV in WT and (A, B, D) Dnmt3a^Vav or (C) Dnmt3a^Mx1 mice. (A) Kaplan-Meier curve of survival probability. (B) Changes in percent body mass relative to starting body weight (baseline) on the day of challenge. WT PBS, n=3; WT IAV, n=10; Dnmt3a^Vav PBS, n=3; Dnmt3a^Vav IAV, n=11. (C) Viral load in lung tissue measured by qRT-PCR. WT IAV, n= 14; Dnmt3a^Mx1 IAV, n=25. (D) Viral load in lung tissue measured by plaque assays. WT IAV, n = 17; Dnmt3a^Vav IAV, n = 28. Statistics: (A) Log-rank Mantel-Cox test (B) 2-way ANOVA with Tukey’s multiple testing. Mean ± SD are shown. (C-D) Mann-Whitney test. Median ± IQ range are shown. (C-D) Outliers were excluded (Q = 5%). (C-D) Data are of 2-3 experiments combined. ns=nonsignificant, *p<0.05. Relevant significant comparisons are shown. Abbreviations: WT = Wild type, IAV = Influenza A virus, PFU = plaque-forming unit

Next, viral load was determined by qPCR detection of viral H1N1PR8-NP in the lung. We found a dramatic increase in viral copy number in Dnmt3a^Mx1 mice compared to controls (Figure 1C). Consistent with this finding, plaque assays in Dnmt3a^Vav mice showed a similar trend of more PFU in the Dnmt3a^Vav mice compared to WT mice (Figure 1D). Overall, the finding that Dnmt3a-deficient mice harbored a higher viral load than WT mice suggests that immunity is impaired in the setting of loss of function Dnmt3a mutation.

Loss of function Dnmt3a mutation reduces neutrophil accumulation in IAV infected lungs.

To elucidate the reason for the elevated viral burden in Dnmt3a-deficient mice, we harvested lungs of IAV-infected Dnmt3a^Mx1 and WT mice to characterize immune populations in the lung. We performed flow cytometric analysis of hematopoietic cells from perfused lungs to quantify immune cell populations including CD4⁺ and CD8⁺ T cells, B cells, inflammatory macrophages, and neutrophils (Figure 2A–F), as well as NK cells and eosinophils (Figure S1C–D). Upon infection, we observed a reduction in T cells (CD4+ and CD8+) in infected WT and Dnmt3a^Mx1 mice relative to their uninfected counterparts, with no significant differences between genotypes. In contrast, neutrophil responses differed markedly between genotypes. Infected WT mice showed an expected robust increase in neutrophils (CD11b⁺ Ly6G⁺) both in absolute numbers and as a percentage of CD45⁺ leukocytes, whereas Dnmt3a^Mx1 mice showed no increase in neutrophils in the lungs upon IAV infection (Figure 2E–F).

Figure 2. Loss of function Dnmt3a mutation reduces neutrophil accumulation in IAV-infected lungs.

(A-F) Influenza Challenge with PBS or IAV in WT and Dnmt3a^Mx1 mice. Flow cytometric quantification of immune populations (n=3-16 per group) in lung. Populations were defined as (A) CD4⁺ T cells (Thy1.2+ CD4+), (B) CD8⁺ T cells (Thy1.2+ CD8+), (C) B cells (B220+), (D) Inflammatory Macrophages (CD11b+ Ly6C⁺ SSC^low), (E-F) Neutrophils (CD11b+ Ly6G⁺). Statistics: 2-way ANOVA with Tukey’s multiple testing. Mean ± SD are shown. Outliers were excluded (Q = 5%). ns=nonsignificant, *p<0.05, **p < 0.01, ****p < 0.0001. Relevant significant comparisons are shown. Abbreviations: WT = Wild type, IAV = Influenza A virus.

To determine if the deficit of neutrophils in lungs of infected Dnmt3a-deficient mice is due to an overall decrease in neutrophil numbers, we quantified immune populations in the peripheral blood by flow cytometry. In contrast to the lung, infection led to a slight increase in neutrophil frequency in the blood of Dnmt3a^Mx1 mice compared to uninfected controls (Figure 3A–B). While CD8+ T cells were somewhat diminished in blood of infected Dnmt3a^Mx1 mice, we did not identify any other significant differences in peripheral blood immune cell populations (Figure 3C–F). These findings indicate that the differences in neutrophils observed within the lung of Dnmt3a-deficient mice are not caused by a lack of circulating neutrophils, but rather by their ability to accumulate at the site of infection.

Figure 3. Loss of function Dnmt3a mutation does not prevent neutrophil accumulation in the blood following IAV infection.

(A-F) Influenza Challenge with PBS or IAV in WT and Dnmt3a^Mx1 mice. Flow cytometric quantification of immune populations (n=10-17 per group) in peripheral blood. Populations were defined as (A-B) Neutrophils (CD11b+ Ly6G⁺) (C) CD4⁺ T cells (Thy1.2+ CD4+), (D) CD8⁺ T cells (Thy1.2+ CD8+), (E) B cells (B220⁺), (F) Monocytes (CD11b+ Ly6G⁻ SSC^low). Statistics: 2-way ANOVA with Tukey’s multiple testing. Mean ± SD are shown. Outliers were excluded (Q = 5%). Data are of 2 experiments combined. ns=nonsignificant, *p<0.05. Relevant significant comparisons are shown. Abbreviations: WT = Wild type, IAV = Influenza A virus.

Loss of function Dnmt3a mutation does not impede neutrophil effector functions and metabolism.

Effective host defense during viral pneumonia relies on a complex interplay between diverse immune cell populations that coordinate viral clearance and inflammatory responses[34, 35]. Neutrophils are among the first leukocytes recruited to the lung upon infection and contribute to pathogen control through phagocytosis and the release of antimicrobial effector molecules. Besides their classical antimicrobial functions, neutrophils are increasingly recognized as important modulators of both innate and adaptive immune responses, for example through cytokine production and the recruitment of additional immune cells [36]. Further, neutrophils contribute to host protection during viral infections, including influenza, through early viral control, production of reactive oxygen species, and neutrophil extracellular trap (NET) formation [37].

Given the clear neutrophil phenotype observed in Dnmt3a-deficient mice, we assessed whether the higher viral load could be due to defective neutrophil function. Lung neutrophils were characterized for differences in maturation markers (CXCR4, CD62L, CD49d and CD69) using flow cytometry. We identified no differences in the expression of these surface markers between the genotypes following infection, suggesting Dnmt3a-deficient neutrophils are capable of efficient maturation (Figure S2A–D).

Neutrophil elastase (NE) and myeloperoxidase (MPO) are key enzymes in neutrophil-mediated pathogen killing. NE supports microbial killing and clearance by degrading bacterial proteins and extracellular matrix components and contributes to neutrophil extracellular trap (NET) formation. MPO catalyzes the conversion of hydrogen peroxide (H₂O₂) and chloride to hypochlorous acid (HOCl), which gives it a central role in neutrophil oxidative burst–dependent killing. To assess whether loss of function Dnmt3a mutation alters protease content of neutrophils in vivo, we performed immunohistochemistry on perfused lung sections from naïve and IAV-infected Dnmt3a^Vav and WT mice (Figure 4A; S3A). Consistent with our flow cytometry data, lungs of infected Dnmt3a^Vav mice contained fewer neutrophils than WT controls (Figure 4B; S3B). Despite lower absolute neutrophil counts in infected Dnmt3a^Vav lungs compared to infected WT, the proportions of NE⁺ neutrophils were similar between genotypes (Figure 4C). Likewise, the fraction of MPO⁺ neutrophils did not differ between groups (Figure S3C), indicating that loss of function Dnmt3a mutation does not diminish abundance of key antimicrobial enzymes in neutrophils of infected lungs.

Figure 4: Loss of function Dnmt3a mutation does not impede neutrophil effector functions.

(A-F) Influenza challenge with PBS or IAV in WT and Dnmt3a^Vav mice. (A-C) Immunofluorescence staining of lung sections. WT PBS, n = 1; WT IAV, n = 3; Dnmt3a^Vav PBS, n = 1; Dnmt3a^Vav IAV, n = 3. (A) Representative images. DAPI (nuclei), a-SMA (Opal 690, vascular smooth muscle), Ly6G (Opal 520, neutrophils); NE (Opal). Bar represents 100 μm. (B) Quantification of lung neutrophils (Ly6G⁺). (C) Percentage of lung neutrophils that are NE⁺. (D-F) Quantification of (D) neutrophil elastase, (E) myeloperoxidase, and (F) hydrogen peroxide (H₂O₂), in intraperitoneally elicited neutrophils of Dnmt3a^Vav and WT mice in vitro (n=3-6 per group). Statistics: (B-C, F) 2way ANOVA with Tukey’s multiple testing. Mean ± SD are shown. (D-E) unpaired two-tailed Student’s t test with Welch’s correction. Mean ± SD are shown. (D-F) Data shown are representative of 2-3 independent experiments. (B-C) Values are averages of 10 regions. ns=nonsignificant, ****p < 0.0001. Relevant significant comparisons are shown. Abbreviations: WT = Wild type, IAV = Influenza A virus, NE = Neutrophil elastase, H₂O₂ = Hydrogen peroxide.

To test quantitative differences in effector enzymes, intraperitoneally (i.p.)-elicited neutrophils were stimulated with phorbol 12-myristate 13-acetate (PMA) to induce degranulation and oxidative responses. Supernatants were assayed for NE activity using a fluorescence-based assay kit. NE secretion was comparable across genotypes (Figure 4D). Similarly, MPO activity was not significantly different in Dnmt3a^Vav neutrophils compared to WT (Figure 4E). Because MPO requires H₂O₂ to generate HOCl and downstream oxidants, we assessed upstream reactive oxygen species (ROS) production. We observed a small, but significant increase in H₂O₂ secretion in Dnmt3a^Vav neutrophils relative to WT (Figure 4F). Pharmacologic inhibition of NADPH oxidase (NOX2) reduced H₂O₂ levels to near baseline in both genotypes, confirming NOX2 as the predominant source of H₂O₂ in these assays (Figure S3D). Notably, residual H₂O₂ levels remained modestly elevated in Dnmt3a^Vav neutrophils under NOX2-inhibition, indicating a potential contribution of additional, NOX2-independent sources of H₂O₂ generation. However, intracellular nicotinamide adenine dinucleotide phosphate (NADPH) content, the substrate utilized by NOX2, did not differ between Dnmt3a^Vav and WT neutrophils (Figure S3E).

The ability of neutrophils to adapt to increased energy needs is one of the main prerequisites for maintained neutrophil function. To determine whether loss of function Dnmt3a mutation alters metabolic capacity, we performed Seahorse extracellular flux analysis on activated neutrophils isolated from the peritoneum after casein induction. The spare respiratory capacity (mitochondrial reserve) did not differ between Dnmt3a^Vav and WT neutrophils (Figure S3F), suggesting comparable ability to meet energetic demands. Altogether, these data indicate that, despite differences in neutrophil accumulation within the lung, loss of function Dnmt3a mutation does not result in functional impairment of neutrophil metabolism or effector function.

Dnmt3a loss of function suppresses chemotactic transcriptional pathways.

To investigate the mechanisms underlying reduced neutrophil accumulation in the lungs of infected Dnmt3a-deficient mice, we performed bulk RNA sequencing and differential gene expression analyses on neutrophils isolated from the bone marrow of naïve mice and from the bone marrow and lungs of Dnmt3a^Vav and WT mice two days after IAV infection. Across all conditions and compartments, Dnmt3a^Vav neutrophils showed distinct transcriptional changes affecting immune signaling and functional response programs when compared to wildtype.

Pathway analysis of differentially expressed genes (DEGs) between Dnmt3a^Vav and WT bone marrow neutrophils following IAV infection showed downregulation of pathways associated with cell activation, immune signaling, chromatin organization, stimulus-responsive transcription, and stress response pathways (Figure S4A). These pathways are closely coupled to activation-dependent signaling states and their reduced expression indicates an overall decrease in transcriptional responsiveness to inflammatory signals in neutrophils of Dnmt3a^Vav mice following infection, consistent with prior studies [38]. Further examination of affected gene regulatory networks identified changes in Runx1-dependent transcription, along with gene sets associated with cell cycle and proliferative signaling, processes linked to hematopoietic differentiation and neutrophil maturation (Figure S4B).

Similarly, analysis of neutrophils isolated from the lung after IAV challenge again identified downregulation of pathways involved in cytokine signaling, cell growth and proliferation, as well as neutrophil effector functions including neutrophil degranulation, indicating reduced transcriptional support for neutrophil expansion and activation in Dnmt3a^Vav neutrophils compared to WT (Figure 5A). Interestingly, analysis of affected gene regulatory networks within the lung identified significant changes in the mTOR pathway, Kit-mediated signaling as well as chemotaxis (Figure 5B). Together, these pathways integrate extracellular cues to regulate neutrophil activation, chemotaxis, and responsiveness to chemokines. Reduced expression of these pathways indicates decreased transcriptional support for neutrophil maturation, activation, and directed migration.

Figure 5. Dnmt3a loss of function suppresses chemotactic transcriptional pathways.

(A-B) Bulk RNA-seq analysis on Ly6G⁺ CD11b⁺ cells sorted from lung of naive and IAV infected WT and Dnmt3a^Vav mice. WT PBS, n = 3; WT IAV, n = 4; Dnmt3a^Vav PBS, n = 4; Dnmt3a^Vav IAV, n = 4. (A) Gene ontology (GO) pathway analysis of differentially expressed genes (DEGs). (B) Gene regulatory network analysis. Abbreviations: WT = Wild type, IAV = Influenza A virus.

Of note, at the individual gene level, the second most significantly downregulated gene in IAV infected bone marrow after Commd10 was Cxcr1, a central regulator of neutrophil chemotaxis (Figure 6A). The expression of Cxcr1 was downregulated in Dnmt3a^Vav neutrophils across all conditions: in naïve bone marrow, infected bone marrow, and infected lungs (Figure 6B; S5A–B). Additional genes involved in neutrophil trafficking, including Cxcr5, Cxcl1 and Cxcl3, were likewise downregulated in Dnmt3a^Vav neutrophils compared to WT (Figure 5B; 6A). Together, the transcriptomic profile points to a cell-intrinsic migration defect in Dnmt3a-deficient neutrophils through disrupted chemokine signaling that is evident at baseline and maintained during infection.

Figure 6. Loss of function Dnmt3a mutation impairs neutrophil chemotaxis.

(A-B) Differentially expressed genes (DEGs) from bulk RNA-seq analysis on Ly6G⁺ CD11b⁺ cells sorted from bone marrow and lung of PBS and IAV infected WT and Dnmt3a^Vav mice. WT PBS, n = 3; WT IAV, n = 4; Dnmt3a^Vav PBS, n = 4; Dnmt3a^Vav IAV, n = 4. (A) Volcano plot of DEGs in Dnmt3a^Vav vs WT IAV infected neutrophils isolated from bone marrow. (B) Cxcr1 expression relative to housekeeper gene Gapdh, presented as normalized read counts. (C) Transwell migration assay showing the number of migrated neutrophils upon stimulation with human IL-8. BM neutrophils were isolated from bone marrow of WT, WT^Vav (Vav-Cre⁺ Dnmt3a^WT) or Dnmt3a^Vav mice. Statistics: (B-C) 2way ANOVA with Tukey’s multiple testing. Mean ± SD are shown. (C) Data shown are from one experiment representative of 2 independent experiments. ns=nonsignificant, *p<0.05, **p < 0.01, ****p < 0.0001. Abbreviations: WT = Wild type, BM = bone marrow, IAV = Influenza A virus, IL-8 = interleukin 8.

Dnmt3a loss of function impairs neutrophil migration in vitro.

To test whether these transcriptional changes translated into functional defects, we performed transwell migration assays with neutrophils from bone marrow of Dnmt3a^Vav and WT mice using interleukin-8 (IL-8), a known neutrophil chemokine, as the chemoattractant [31]. Dnmt3a^Vav neutrophils showed significantly impaired migration compared to WT controls (Figure 6C). Because Commd10 is affected by the Vav-Cre insertion, we further tested Dnmt3a^WT Vav-Cre⁺ (WT^Vav) neutrophils as an additional control. These mice carry the Vav-Cre insertion with the resulting deletion of Commd10 while the Dnmt3a locus is unedited. WT^Vav neutrophils migrated comparably to WT and significantly more than Dnmt3a^Vav neutrophils (Figure 6C), confirming that the migration defect results from Dnmt3a loss rather than Commd10 deletion. In summary, transcriptomic analysis and in vitro migration assays indicate a cell-intrinsic migration defect in Dnmt3a-deficient neutrophils driven in part by reduced Cxcr1 expression.

Mediators of neutrophil migration are impaired in humans with DNMT3A mutation

Tatton-Brown-Rahman syndrome (TBRS) is a human overgrowth syndrome caused by germline mutations in DNMT3A [39]. Notably, TBRS has been associated with altered myeloid homeostasis, characterized by neutrophilia and an increased risk for myeloid neoplasms. To understand whether neutrophil migration may also be impacted in humans with loss-of-function mutations in DNMT3A, we performed proteomic analysis of plasma samples from individuals with TBRS. Plasma samples from 30 TBRS patients aged 4-40 years were compared to plasma samples from 12 unaffected siblings aged 2-46 years. Samples were drawn from patients who were in their usual state of health, without known infection or acute illness. Targeted proteomic analysis revealed proteins important in neutrophil migration, including matrix metalloproteinase 2 (MMP2) and several proteins important in actin cytoskeleton dynamics (PFN1, CFL1, TPM4) to be enriched in WT plasma compared with TBRS. In contrast, proteins important in cell-cell adhesion, tight junctions, and barrier function (CDH5, FCN3, KRT5) were more abundant in TBRS plasma compared to WT controls (Figure 7A–B). Notably, KRT5 has also been implicated in regulation of macrophage chemotactic cytokines CCL2 and MIP [40], suggesting that DNMT3A mutations may also interfere with myeloid cell migration by dampening chemotactic cytokines. Altogether, analysis of plasma from TBRS and control patients supports a role for DNMT3A in myeloid migration via modulation of cell-cell adhesion, chemotactic cytokines, and matrix metalloproteinase 2.

Figure 7. Mediators of neutrophil migration are altered in humans with DNMT3A mutation.

(A-B) Mass spectrometry proteomic analysis of plasma samples from individuals with Tatton-Brown-Rahman-Syndrome (TBRS) compared to unaffected controls. TBRS, n=30; controls, n= 12. (A) Volcano-Plot showing differentially abundant plasma proteins in TBRS patients relative to control individuals. 11/330 proteins significant at p<0.05 and 2-fold cut-off. (B) Heatmap showing relative abundance of differentially abundant proteins in TBRS vs. control plasma based on targeted analysis. Each column represents an individual participant. Abbreviations: TBRS = Tatton-Brown-Rahman-Syndrome (TBRS); log(iBAQ) = log of the intensity based absolute quantification.

Discussion

CH has been linked to increased susceptibility to severe infections [11], but how CH-associated mutations affect anti-pathogen immunity remains unknown. Here, we show that loss of function Dnmt3a mutation weakens the innate immune defense by impairing neutrophil migration.

The innate immune system plays a central role in early defense against viral infections such as IAV by detecting viral components through pattern-recognition receptors and releasing cytokines that mediate effective pathogen clearance by adaptive immune cells and other mechanisms [41, 42]. Ineffective pathogen clearance is associated with increased disease severity and mortality in both disease models and clinical studies [43] [41, 44]. Indeed, we observed an increased viral load in the lungs of Dnmt3a-deficient mice paired with reduced neutrophil numbers in the infected lung. Surprisingly, our study did not show a difference in survival upon IAV challenge. The severity and rapid progression of lethal IAV infection may have limited the ability to detect subtle differences in host defense mechanisms that could become more apparent in sublethal or alternative infection models [45] [46].

CH-associated mutations have been shown to influence hematopoiesis and immune function by altering the differentiation and transcriptional programming of immune cells. In both humans and mice, loss of Dnmt3a leads to increased hematopoietic stem cell self-renewal and clonal expansion, with relatively subtle or context-dependent effects on peripheral blood composition. While DNMT3A-deficient human HSPCs show neutrophil skewing in vitro [47], epidemiologic data and mouse models lack a clear peripheral blood phenotype [48], possibly pointing towards a slower, marrow-restricted myeloid skewing [49]. Here we did not observe significant differences in circulating neutrophil counts between Dnmt3a-deficient and WT mice, but the ability of neutrophils to accumulate at sites of infection was severely impaired. We also observed a slight decrease in CD8+ T cells upon infection in Dnmt3a-deficient mice potentially consistent with the known role of Dnmt3a in expansion and polarization of T cells during infection [50]. However, these differences in CD8+ T cells were relatively minor. Overall, our data suggest that increased infection susceptibility is driven by an inability of innate immune cells to efficiently traffic to sites of infection.

In support of this model, our transcriptional analyses indicate that loss of function Dnmt3a mutation profoundly alters neutrophil gene programs linked to activation, migration, and tissue recruitment. As a major differentially expressed network, we found that genes regulated by Runx1 were significantly downregulated in neutrophils from bone marrow of IAV-infected Dnmt3a^Vav mice compared to WT. Runx1 has recently been defined as a key regulator of early neutrophil maturation, required especially during the bone marrow–to–blood transition. Importantly, Runx1-deficiency resulted not only in delayed neutrophil maturation but also in impaired chemotaxis and reduced recruitment to inflamed tissues [51]. In addition to Runx1-associated networks, our data indicate that loss of function Dnmt3a mutation suppresses mTOR–PI3K–Akt–dependent transcriptional pathways that integrate cytokine and chemokine signals to regulate neutrophil cytoskeletal remodeling and directed chemotaxis, as well as aspects of effector function [52, 53].

Notably, Cxcr1 was one of the most differentially expressed genes, and this downregulation was conserved across conditions and tissues in the mouse. These transcriptional changes were accompanied by impaired neutrophil chemotaxis both in vivo and in vitro. Strikingly, Quin et al. recently reported a very similar phenotype in Tet2-deficient mice during Streptococcus pneumoniae infection, with impaired recruitment of neutrophils to the lung along with reduced expression of migration-associated genes [14]. In addition, they observed lower circulating levels of the chemokines CXCL1 and CXCL5 in CH carriers from the UK biobank [14]. Given the phenotypic similarities between Dnmt3a and Tet2 mutations, these findings support a broader role for CH-associated epigenetic modifiers in the regulation of neutrophil migration. However, the precise molecular mechanisms by which DNMT3A and TET2 mediate these effects remain unclear, in particular whether the observed transcriptional changes arise from direct alterations in DNA methylation or reflect indirect effects. While emerging studies suggest that DNMT3A may have functions beyond its canonical role as a DNA methyltransferase, the absence of DNA methylation data in our study limits definitive mechanistic conclusions and highlights an important area for future investigation.

Comparison of plasma proteomics from TBRS patients lacking DNMT3A and unaffected siblings did not show reductions in the same proteins, such as CXCR1, seen in the mouse transcriptomics. This is not surprising since the human proteomics were performed on cell-free plasma at steady state conditions, in which chemokine receptors would not be expected. Nonetheless this human study revealed a reduction in proteins such as MMP2 involved in cellular migration. Interestingly, one of the proteins noted to be increased in the plasma of TBRS patients was ficolin-3 (FCN3), a potent complement activator that plays an important role in tumor immune evasion by regulating immune cell infiltration [54]. Overall, while there was no direct correlation between the cell-intrinsic transcriptional differences seen in mouse neutrophils and the plasma proteomic differences seen in human, both datasets point toward impairment in innate immune cell migration in the setting of Dnmt3a mutation.

One limitation of our study is the use of Vav-Cre mice, which were recently reported to harbor a loss of function mutation in Commd10 [33], a regulator of NF-κB signaling, which potentially may alter inflammatory responses independently of Dnmt3a. However, comprehensive phenotyping by the International Mouse Phenotyping Consortium (http://www.mousephenotype.org/) did not reveal any significant immune or systemic alterations associated with the Commd10 variant [55]. To address potential concerns related to the Vav-Cre background, we reproduced our key in vivo findings in Mx1-Cre mice, thereby strengthening our conclusions by demonstrating consistency across two independent genetic systems.

Aging profoundly alters the immune system, a process often referred to as inflammaging, which is characterized by chronic inflammation and impaired immune regulation [56]. These age-associated immune changes contribute to both clonal expansion of mutant HSPC clones [6, 57–59] and increased susceptibility to infections [60–62]. One limitation of our study is that it was conducted in young mice, which may not fully recapitulate the complex interplay between inflammation, aging, CH and infections observed in older hosts.

Taken together, our data demonstrate that loss of function Dnmt3a mutation in the hematopoietic system increases susceptibility to influenza A infection through impaired neutrophil migration. The risk of influenza-related mortality rises sharply with age; individuals over 65 account for more than 90% of influenza-associated deaths in the United States [63]. Yet the biological links between aging, inflammation, impaired innate immunity, and CH remain only partially understood. Given the high prevalence of DNMT3A-mutant CH in aging populations, and with the number of individuals over 60 expected to double between 2020 and 2050 [64], defining how CH-associated mutations alter host defense will become increasingly important. Our study adds to the understanding of how specific CH mutations reshape immune effector functions and may ultimately inform targeted strategies to restore immune competence in affected individuals.

Highlights.

1. Dnmt3a deficiency increases viral burden in a mouse model of influenza infection

2. Dnmt3a loss impairs neutrophil migration and recruitment to the infected lung

3. Dnmt3a-mutant neutrophils show impaired CXCR1 expression and chemotaxis

4. Defects in neutrophil migration occur without loss of neutrophil effector function

5. Humans with DNMT3A mutation show reduced migration-related plasma proteins