henhaobio
Sign in

PLoS PathogensImmunologyPeer-reviewed journal article

Transcriptomic timeseries links hepatic gene expression to an early and self-limited systemic response to enteric infection

Yuko HasegawaiD, Akina OsakiiD, Masataka SuzukiiD, Ting YaniD, Ian W. CampbelliD, Matthew K. WaldoriD

Abstract

The liver receives microbe and host signals from the intestine via the portal vein, thereby connecting the gut to systemic physiology. Homeostatic control of the timing of systemic responses is critical to prevent the expansion and dissemination of gut microbes and to mitigate untoward effects from prolonged systemic inflammation, however these mechanisms remain enigmatic. Here, to determine the role of the liver in coordinating systemic immune responses to enteric infection, matched measurements of global gene expression profiles were collected from the murine liver and intestinal epithelium throughout the course of enteric infection and clearance of Citrobacter rodentium, a mouse model of infectious colitis. These data revealed metabolic suppression in the liver during the peak of infection and a long-lived immune signaling pattern in the colon associated with CD4 and CD8 T cell infiltration that persisted beyond the clearance of infection. Furthermore, an early inflammatory signal was detected in the liver that resolved before the peak of disease and pathogen colonization. This self-limited, early signal depended on the pathogens virulence program and correlated with the timing of a corresponding systemic response, including circulating TNF- and IL-6, key mediators of acute-phase proteins. These results uncover the temporal pattern of hepatic changes in response to the course of intestinal infection and provide correlative evidence that an early pulse of gene expression in the liver coordinates and limits the duration of the systemic acute-phase protein response. Author summaryLocalized infections can trigger systemic inflammatory responses that help limit infection. However, prolonged systemic inflammation risks tissue damage and disease. Thus, the timing of systemic immune reactions is critical to the balance between immunity and damage. Yet, knowledge of the pathways that link gut-localized infections to systemic immune tone is limited. Since the liver is anatomically central to the connection between the gut and circulatory systems, we hypothesized that monitoring the kinetics of liver gene expression throughout the natural course of an intestinal bacterial infection would provide a valuable resource and yield insight into the control of systemic immune tone. Critically, we identified a burst of liver cytokine signaling that occurred and resolved before peak pathogen burden and disease in the colon and predicted the circulating inflammatory response. We propose that this early and self-limited signal from the liver coordinates the timing of the systemic response, ensuring it occurs early enough to promote immunity and resolve before causing tissue damage.

In-depth summary

Claude Sonnet 4.6 · xhigh effort · 8 Nov 2026 · from the full text of version 2 (27 Jul 2026) · CC BY 4.0

Matched liver and colon RNA-seq timeseries during Citrobacter rodentium infection in mice reveal an intimin-dependent hepatic cytokine burst at 3 days that tracks serum cytokine kinetics.

How gut-localized infection sets the timing of systemic immune tone is poorly charted, and prolonged signals such as IL-6 can cause tissue damage. This work pairs pathogen burden with organ-specific transcriptomes to show that liver expression, not colonic epithelial expression, tracks systemic cytokines.

Key results

  1. ShownLiver transcriptional responses to enteric infection arise earlier than colonic epithelial ones and return to baseline by 36 dpi, whereas colonic profiles stay altered after pathogen clearance.

    More DEGs in liver than colonic epithelium on days 1 and 3, colon exceeding liver from day 6; liver PC2 signal only at day 3; colonic burden peaked 6-10 dpi at ~10^9 CFU/g, while some animals had ~10^3 CFU/g in liver on days 6 and 10, generally <1/10^6 of colonic burden. · Enteric infection induces an early pattern of…; Figure legends

  2. SuggestedThe early (3 dpi) hepatic inflammatory program depends on intimin-mediated enteric infection, not on the inoculum bolus, and does not require EspF-dependent tight junction disruption.

    Δeae failed to expand and gave fewer PC2 marker transcripts by qPCR (control n=4, wt n=8, Δeae n=4) and RNA-seq; ΔespF (4.3-fold colonization defect) raised PC2 markers as high as wild type (control n=6, wt n=9, ΔespF n=9). qPCR increase in 4 out of 8 wt-infected animals. · Enteric infection induces an early liver gene…; Figure legends

  3. SuggestedTiming of liver cytokine/chemokine RNA predicts serum protein abundance, including an early set (TNF-α, IL-6) peaking at 3 dpi, whereas colonic epithelial RNA almost never does.

    46 serum cytokines profiled; early set (TNF-α, IL-6, IP-10, KC, MCP-1, MIP-1β) peaked at 3 dpi in liver RNA and serum; Pearson correlation (n=32, p<0.01 highlighted); in colonic epithelium only TIMP-1 significantly predicted serum abundance. · Enteric infection induces an early liver gene…; Figure legends

  4. ShownInfection transiently suppresses hepatic metabolic gene sets, mostly at peak colonization, with little comparable suppression in the colon.

    KEGG GSEA: oxidative phosphorylation, fatty acid metabolism, drug metabolism and steroid hormone biosynthesis reduced in liver, mainly at 6 and 10 dpi, back to baseline by 17 to 36 dpi; generally not in colon (adjusted p<0.01 cutoff for plotted pathways). · Enteric infection transiently diminishes hepatic…; Figure legends

  5. ShownColonic infection leaves a durable immune imprint: delayed innate cell influx, long-lived CD4 and CD8 T cell residence (primarily memory phenotype in the epithelial fraction), and persistent interferon and inflammatory pathway activation, unlike the liver.

    Colonic CD4 T cells >30-fold and CD8 T cells >10-fold, peaking 17 dpi and persisting to at least 30 dpi; innate cells peaked 17 dpi in colon vs 10 dpi in liver (n=5/group, two-way ANOVA); colonic pathways elevated to at least 36 dpi. · Enteric infection results in a long-lasting…; Figure legends

Methods, data and reproducibility

Female C57BL/6J mice were orally infected with C. rodentium and sampled over time for CFU burden, bulk RNA-seq of liver and colonic epithelium (DESeq2, PCA, GSEA), flow cytometry and a multiplex serum cytokine panel. Δeae and ΔespF mutants tested virulence dependence.

  • Bulk RNA-seq timeseries of liver and colonic epithelium, days 0 to 36 after C. rodentium infection · Mus musculus (C57BL/6J) · n = 4 per time point (Fig. 6B legend); 32 in correlation analysis · GSE326680; PRJNA1445395
  • Serum multiplex panel of 46 cytokines from the same timecourse animals · Mus musculus (C57BL/6J) · n = 5 per day (Fig. 6A legend)
  • Spectral flow cytometry of immune cells from liver, colonic lamina propria and colonic epithelium · Mus musculus (C57BL/6J) · n = 5 per group
  • Liver qPCR of PC2 marker genes at 3 dpi with wild type or Δeae C. rodentium · Mus musculus (C57BL/6J) · n = control 4, wt 8, Δeae 4
  • Liver qPCR at 3 dpi with wild type or ΔespF C. rodentium · Mus musculus (C57BL/6J) · n = control 6, wt 9, ΔespF 9
Methods & tools
STAR v2.7.3a, featureCounts (subread v2.0.0), DESeq2 v1.36.0, clusterProfiler v4.4.4, msigdbr v7.5.1, KEGG metabolism pathways, Combat-seq (sva), Trim Galore v0.6.6, PAM clustering (cluster v2.1.4), FlowJo v10.8, GraphPad Prism 10, Eve Technologies multiplex immunoassay
Software & environment
STAR v2.7.3a, DESeq2 v1.36.0, subread v2.0.0, clusterProfiler v4.4.4, msigdbr v7.5.1, cluster v2.1.4, Trim Galore v0.6.6, FlowJo v10.8, GraphPad Prism 10, ggplot2 v4.0.2; mm10 genome, GENCODE M25 annotation.
Reproducibility gaps
Supplementary Tables 1 and 2 (DEG and pathway results), antibody and primer tables, and per-test statistics are in figure legends or supplements not reproduced in the supplied text.
Limitations (6)
  • The evidence is correlational: no direct experiment connects the hepatic events to blood protein levels, and blocking delivery of the signal to the liver was not tested. · stated by the authors
  • What signals, possibly originating in the cecum, activate the early liver response has not been identified, so the upstream trigger remains open. · stated by the authors
  • The 3 dpi PC2 response was heterogeneous (3 of 4 animals in RNA-seq; 4 of 8 by qPCR), possibly because a brief window was missed in some mice. · stated by the authors
  • The colonic epithelial fraction may contain intraepithelial immune cells and lamina propria contamination, and resident lamina propria cells were not profiled transcriptomically. · stated by the authors
  • The Δeae RNA-seq comparison combined a separate animal batch with four reused wild-type timeseries samples and relied on Combat-seq batch correction, which may leave residual batch effects. · noted in this summary
  • Only female C57BL/6J mice aged 8-10 weeks were used, so sex- and strain-generality of the hepatic kinetics is untested. · noted in this summary

Read this if you study gut-liver axis inflammation, or want a matched multi-organ timeseries design linking tissue transcriptomes to serum cytokine kinetics.

Citations
0
FWCI
0.00
Field percentile
top 48%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Similarity Map
SelectedSame fieldCross-field
THISWinkle '21Zhang '23Hackstein '22Konijnenbur '17Mishra '25Kulkarni '24
65% semantic matchImmunology
J. A. Van Winkle, S. T. Peterson, E. A. Kennedy+11 more·bioRxiv·Jun 2, 2021