Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2012 May 22:8:584.
doi: 10.1038/msb.2012.17.

Multi-layered stochasticity and paracrine signal propagation shape the type-I interferon response

Ulfert Rand  1 , Melanie Rinas, Johannes Schwerk, Gesa Nöhren, Melanie Linnes, Andrea Kröger, Michael Flossdorf, Kristóf Kály-Kullai, Hansjörg Hauser, Thomas Höfer, Mario Köster
Affiliations

Multi-layered stochasticity and paracrine signal propagation shape the type-I interferon response

Ulfert Rand et al. Mol Syst Biol. .

Abstract

The cellular recognition of viruses evokes the secretion of type-I interferons (IFNs) that induce an antiviral protective state. By live-cell imaging, we show that key steps of virus-induced signal transduction, IFN-β expression, and induction of IFN-stimulated genes (ISGs) are stochastic events in individual cells. The heterogeneity in IFN production is of cellular-and not viral-origin, and temporal unpredictability of IFN-β expression is largely due to cell-intrinsic noise generated both upstream and downstream of the activation of nuclear factor-κB and IFN regulatory factor transcription factors. Subsequent ISG induction occurs as a stochastic all-or-nothing switch, where the responding cells are protected against virus replication. Mathematical modelling and experimental validation show that reliable antiviral protection in the face of multi-layered cellular stochasticity is achieved by paracrine response amplification. Achieving coherent responses through intercellular communication is likely to be a more widely used strategy by mammalian cells to cope with pervasive stochasticity in signalling and gene expression.

PubMed Disclaimer

Conflict of interest statement

The authors declare that they have no conflict of interest.

Figures

Figure 1
Figure 1
Quantitative and temporal heterogeneity of IFN-β induction. A BAC-based reporter construct in which the IFN-β gene is replaced by TurboGFP was integrated into murine NIH3T3 fibroblasts. A cell clone with a stable integration of the BAC and representative response towards NDV infection was used (error bars represent ±s.d. of triplicates). (A) Induction of IFN-β–tGFP expression upon NDV infection. IFN-β reporter cells were infected with NDV for 1 h. Expression of tGFP was determined 24 h post-infection by flow cytometry. Representative dot plots are shown for 10, 20, und 40 HAU/ml NDV. (B) IFN-β reporter reflects endogenous IFN production. IFN-β–tGFP expression frequencies after infection with 40 HAU/ml NDV were detected at various time points by flow cytometry. Frequencies were plotted against time points post-infection (black circles) and compared with titres of type-I IFN in the supernatant (grey rhombs). (C) IFN-β expression frequency increases with viral titre. Reporter cells infected with increasing concentrations of NDV (HAU/ml) were subjected to flow cytometry 24 h post-infection. Frequency of IFN-β–tGFP expression (circles) following infection with 1, 2, 5, 10, 20, 40, 80, and 100 HAU/ml and the geometric mean of their fluorescence intensity (triangles) are presented. Source data is available for this figure in the Supplementary Information.
Figure 2
Figure 2
Viral replication is necessary but not sufficient to induce IFN-β expression. (A) Fractional IFN-β expression among productively infected cells. Reporter cells were infected with 40 HAU/ml NDV for 1 h. IFN-β–tGFP reporter expression and intracellular NDV HN protein was measured by flow cytometry at indicated time post-infection. Dot plots show IFN-β–tGFP expression among productively infected (NDV HN+) cells at indicated time post-infection. (B) Separate kinetics of viral replication and IFN-β expression. Frequency of IFN-β–tGFP (black circles) and NDV HN expression (grey squares) over time. (C) Unresponsiveness is not caused by the absence of inducing viral RNA. NDV-infected (80 HAU/ml) IFN-β–tGFP reporter cells were separated into GFP+ and GFP− fractions. Total RNA was isolated and transfected into naive IFN-β–tGFP reporter cells (lower graphs). RNA from non-infected cells served as a control (upper graph). The frequency of IFN-β–tGFP-expressing cells 20 h after transfection is presented. Source data is available for this figure in the Supplementary Information.
Figure 3
Figure 3
Temporal variability in cellular IFN-β induction. (A) IFN-β expression onset in single cells. Variability of response timing is virus-independent. IFN-β–tGFP reporter cells infected for 1 h with indicated concentrations of NDV or transfected with poly I:C at given concentrations were subjected to time-lapse microscopy (15 min picture intervals). Distribution of tGFP expression onset over time (scatter plot, n=456 (NDV), n=140 (poly I:C)) and CVs are shown. (B) Synchronous activation of NF-κB and IRF-7. NIH3T3 cell clone stably expressing the fusion proteins IRF-7–CFP and NF-κB/p65–YFP were infected with 80 HAU/ml NDV for 1 h and subjected to time-lapse microscopy. Fluorescence pictures for CFP and YFP were taken every 20 min. Subcellular localization of IRF-7–CFP (left column) and p65–YFP (right column) at indicated time after infection. The diagram shows relative nuclear fluorescence for IRF-7–CFP and p65–YFP from sister cells. (C) Synchronicity is independent of response time. IRF-7–CFP and p65–YFP initial nuclear translocation were determined in individual cells and plotted against each other (n=65). Coloured dots indicate the frequency of data points. (D) Expression delay of an individual cell. NIH3T3 cell clone stably expressing IRF-7–CFP together with IFN-β–tGFP were infected with 80 HAU/ml NDV. Fluorescence pictures for CFP and GFP were taken every 20 min. Subcellular localization of IRF-7–CFP (left column) and IFN-β–tGFP expression (right column) at indicated time after infection. Graphs show relative IRF-7–CFP nuclear fluorescence and tGFP intensity. Tsig: time interval between infection and IRF-7–CFP nuclear translocation. Tgen: time interval between IRF-7–CFP nuclear translocation and onset of IFN-β–tGFP gene expression. (E) Response variation at distinct stages of IFN induction. The starting times for IRF-7–CFP nuclear translocation were plotted against the times of IFN-β–tGFP expression for individual cells (n=315). Source data is available for this figure in the Supplementary Information.
Figure 4
Figure 4
Temporal variability of signalling events in sister cells reveals stochasticity. NIH3T3 cell clone stably expressing IRF-7–CFP or IRF-7–TagRFP together with IFN-β–tGFP were infected with 80 HAU/ml NDV for 1 h (A–D, n=38 sister-cell pairs) or transfected with poly I:C (5 μg/ml) (E–H, n=36 sister-cell pairs) and subjected to time-lapse microscopy (20 min interval). Sister-pair analysis was carried out for IRF-7 nuclear translocation and IFN-β expression onset. Coloured dots indicate the frequency of data points. (A, E) Time of IRF-7 nuclear translocation onset (Tsig) in sister-cell pairs. (B, F) Time intervals between IRF-7 nuclear translocation and IFN-β–tGFP expression onset (Tgen) among sister cells. (C, G) Time elapsed from cell division to IRF-7 nuclear translocation (Tsig−Tdiv) of sister cells. (D, H) Time elapsed from cell division to IFN-β–tGFP expression (Tsig+Tgen−Tdiv). Source data is available for this figure in the Supplementary Information.
Figure 5
Figure 5
Bimodal antiviral response towards IFN. A BAC-based reporter construct containing mCherry fused to the C-terminus of the chromosomal IRF-7 gene was integrated into NIH3T3 cells. Experiments were performed with a cell clone exhibiting a stable integration of the BAC and a representative response towards IFN. (A) Binary dose- and time-dependent IRF-7–mCherry expression. IRF-7–mCherry reporter cells were stimulated with the indicated concentrations of IFN-β, mCherry expression was determined by flow cytometry. (B, C) Bimodality of IRF-7 expression is reflected in ISG transcription and antiviral protection. (B) Reporter cells were treated with 500 U/ml IFN-β for 16 h and subjected to FACS. IRF-7–mCherry positive (2) and negative (1) populations indicated by the shaded areas were separated. RNA was prepared from both populations and analysed by qRT–PCR for expression of the indicated ISGs and Rps9 as a control. The results were normalized to β-actin mRNA and are shown as fold increase of untreated reporter cells. (C) Reporter cells treated with 500 U/ml IFN-β for 8 and 24 h were infected with 80 HAU/ml NDV. In all, 20 h after infection, cells positive (blue) or negative (red) for NDV HN (left column, intracellular antibody staining) were analysed for IFN-stimulated IRF-7–mCherry expression (right column) by flow cytometry.
Figure 6
Figure 6
A multi-scale mathematical model of IFN induction and response. (A) Scheme of the state transitions of an individual cell. The model describes a population of individual cells with virus replication, induction of IFN genes through RIG-I signalling to NF-κB and IRFs, induction of ISGs (including IRF-7) by IFN, and cell communication via secreted IFN. The possible state transitions of an individual cell with the propensity functions w are shown. The colour code indicates nuclear translocation of IRFs/NF-κB (cyan nucleus), induction of IFN (green cytoplasm), and induction of IRF-7 (red cytoplasm), as imaged experimentally. Induction of IFN-β (dashed arrow wI+) is explicitly modelled as a multi-step process to match the available experimental data (see (C) below); all other transitions are modelled as single steps (solid arrows). (B) Model simulation of NF-κB/IRF nuclear translocation times versus experimental data. The distribution computed with the model (black line) matches the data (histogram) obtained for high infection dose (80 HAU/ml). (C) Model simulation of IFN-β–tGFP onset times versus experimental data. The distribution computed with the model (black line) matches the data (histogram) obtained for high infection dose (80 HAU/ml). (D) Comparison between model and experimental data for high-dose NDV infection. The model reproduces the observed dynamics of viral load (as measured by HN expression), IFN-β–tGFP induction, extracellular IFN titre and IRF-7–mCherry expression (solid lines, model; dots, experimental data; infection dose 40 HAU/ml). Note that the smooth model curves in (B, C) are mean values obtained by simulating 104 cells. Source data is available for this figure in the Supplementary Information.
Figure 7
Figure 7
Paracrine amplification of the IFN response is predicted by the model and verified experimentally. (A) Comparison between model and experiment for low infection doses. For low viral doses yielding sparse infection (1, 2, and 5 HAU/ml), the model predicts IFN response kinetics of the correct magnitude (solid lines, model; dots, experimental data). (B) Paracrine propagation of the signal. The rarity of IFN-β-producing cells at low infection doses reveals strong paracrine propagation of the IFN response because up to ∼40 times as many cells respond with IRF-7–mCherry expression (IFN responders). The agreement between model (squares) and experimental data (black dots) is remarkable, given that the model was calibrated only for a single condition (40 HAU/ml, marked with blue square) while the red squares represent predictions that were subsequently tested. (C) Co-culture of IFN-β–tGFP (green) and IRF-7–mCherry (red) cells illustrates paracrine communication. IFN-β–tGFP reporter cells were infected with 40 HAU/ml NDV for 1 h. Then IRF-7–mCherry cells were added at same density and cells were subjected to time-lapse microscopy. Merged fluorescent pictures for IFN-β–tGFP and IRF-7–mCherry at selected time points post-infection are shown. Source data is available for this figure in the Supplementary Information.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Apostolou E, Thanos D (2008) Virus infection induces NF-κB-dependent interchromosomal associations mediating monoallelic IFN-β gene expression. Cell 134: 85–96 - PubMed
    1. Brennan K, Bowie AG (2010) Activation of host pattern recognition receptors by viruses. Curr Opin Microbiol 13: 503–507 - PubMed
    1. Brierley MM, Fish EN (2005) Stats: Multifaceted regulators of transcription. J Interferon Cytokine Res 25: 733–744 - PubMed
    1. Chen S, Short JAL, Young DF, Killip MJ, Schneider M, Goodbourn S, Randall RE (2010) Heterocellular induction of interferon by negative-sense RNA viruses. Virology 407: 247–255 - PMC - PubMed
    1. Childs K, Stock N, Ross C, Andrejeva J, Hilton L, Skinner M, Randall R, Goodbourn S (2007) mda-5, but not RIG-I, is a common target for paramyxovirus V proteins. Virology 359: 190–200 - PubMed

Publication types

MeSH terms