n e w s a nd v i e w s
Stop the executioners Andreas Wack
npg
© 2015 Nature America, Inc. All rights reserved.
Virus-triggered type I interferon induces the lysine methyltransferase Setdb2; this then generates repressive histone marks on the promoters of genes encoding molecules important for antibacterial immunity. This process can contribute to influenza virus–associated bacterial superinfection.
U
nlike experimental mice that live in specific pathogen-free conditions, humans walk the dirty streets of the real world. This means that the human immune system must cope with constant polymicrobial exposure, in contrast to the majority of infection models, in which a single pathogen is applied in an otherwise clean environment. In this issue of Nature Immunology, Schliehe et al. present mechanistic data describing the events underlying an experimental model of polymicrobial exposure and bacterial superinfection1. In real life, microbes do not form an orderly queue; instead, opportunistic and colonizing species often seize their chance to invade and proliferate while an immune response to another pathogen is ongoing. This is demonstrated perhaps most dramatically in bacterial superinfections following infection with influenza virus, a phenomenon observed as early as during the 1918 influenza pandemic, which killed far more people than the First World War itself. Eminent pathologists of the time were aware of the high frequency of superinfections during this public health emergency2, which prompted one of them, Louis Cruveilhier, to state in 1919: “If influenza condemns, additional infection executes.”3,4 The actual executioners in this deadly process were soon to be identified as Streptococcus pneumoniae, Staphylococcus aureus and other bacteria that are commonly found even in healthy people but can cause severe and lethal pneumonias in influenza virus–infected people4. Severe coinfections are not restricted to influenza pandemics or infections with highly pathogenic Division of Immunoregulation, Medical Research Council, National Institute for Medical Research, Mill Hill, London, UK. e-mail: [email protected]
6
strains of influenza virus but are an important component of morbidity and mortality in every influenza season, including those of moderate severity5. Studies have shown that the impressive synergy between influenza virus and respiratory bacteria observed in the clinic can also be modeled in mice. Broadly, the underlying mechanisms identified so far can be categorized into two sets6: virus-induced damage of lung epithelia that leads to impaired clearance of bacteria, or suppression of antibacterial immune responses by preceding infection with influenza virus. In the latter category, type I interferon–mediated suppression of interleukin 17 (IL-17) responses by canonical αβ T cells or γδ T cells present in the lung tissue7,8 or type I interferon–dependent suppression of macrophage or neutrophil action9–11 has been shown to mediate the impaired antibacterial response following infection with influenza virus. In particular, one of the mechanisms of diminished neutrophil responses after infection with influenza virus is a type I interferon– mediated decrease in the abundance of the neutrophil chemoattractant CXCL1 (KC)11. Although it is becoming increasingly clear what cell types and cytokines are involved in influenza virus–induced impairment of antibacterial immunity, the study by Schliehe et al.1 is among the first to attempt a description of the molecular mechanism underlying the suppression of the recruitment of neutrophils into the lungs. They show that one of the interferon-stimulated genes upregulated in the influenza virus–infected lungs is the lysine methyltransferase Setdb2. This enzyme is involved in generating the suppressive histone mark H3K9me3 by catalyzing tri methylation of histone 3 at Lys9. Cells from mice with a hypomorphic gene-trap construct of Setdb2 (Setdb2GT/GT mice) have much less Setdb2 protein and higher expression of a
range of genes, including Cxcl1, than do wild-type cells, upon stimulation with a variety of Toll-like receptor agonists. Chromatin immunoprecipitation indicates that Setdb2 binds to the Cxcl1 promoter upon cell stimulation, which leads to an increase in the repressive H3K9me3 mark at this promoter; this increase is found in macrophages from wild-type mice but not those from Setdb2GT/GT mice. In vivo, this translates into higher expression of CXCL1 induced by lipopolysaccharide or influenza virus and more inflammation in Setdb2GT/GT lungs than in wild-type lungs. Notably, this lack of repression of CXCL1 also shows protective effects when the authors assess influenza virus–bacteria1 coinfection. Setdb2GT/GT mice infected with influenza virus cope better with subsequent exposure to S. pneumoniae than do influenza virus–infected wild-type mice because they have better recruitment of neutrophils and, consequently, lower bacterial loads in the lungs (Fig. 1). The authors offer a molecular explanation for why stopping the bacterial executioners becomes difficult after a preceding infection with influenza virus: interferon-dependent upregulation of Setdb2 leads to a repressive histone mark on genes such as Cxcl1, which encode molecules that are important in the antibacterial response. This will reduce the recruitment of neutrophils and thus the ability of the host to control bacterial infection. However, it remains unclear why evolution has installed this interferon-induced suppressive mechanism in the first place. It might be part of a damage-limitation effort at the end of a strong inflammatory process associated with interferon-driven antiviral immune responses, and the bacteria, which either are already present in the lungs (in healthy carriers) or enter
volume 16 number 1 january 2015 nature immunology
n e w s a nd v i e w s
Influenza virus
S. pneumoniae 1
1
b Lung
6
2
IFN-α and IFN-β
Elimination of bacteria
Bacterial outgrowth
7 Neutrophil recruitment
CXCL1
Lung macrophage 2
5
PRR
Neutrophil
NF-κB Nucleus
PRR
IFN-α/βR
Less secretion of CXCL1
4 Setdb2↑ 3
6
Diminished neutrophil recruitment
Marina Corral Spence/Nature Publishing Group
a
npg
© 2015 Nature America, Inc. All rights reserved.
3
4
5
NF-κB
DNA CXCL1
H3
Less CXCL1
Figure 1 Infection with influenza virus reduces the immune response to secondary bacterial infection by a Setdb2-dependent mechanism. (a) Infection with S. pneumoniae (1) leads to the recognition of bacterial ligands by pattern-recognition receptors (PRRs) (2), which triggers activation of NF-B and its translocation to the nucleus (3). NF-B-mediated induction of Cxcl1 (4) leads to the production and secretion of CXCL1 and to the recruitment of neutrophils (5), which control bacterial spread and eliminate bacteria (6). (b) During coinfection, infection with influenza virus (1) leads to the production of type I interferons (INF- and IFN-β) by infected airway epithelial cells (2) and other cells, which induces Setdb2 production in lung macrophages (3) and other cells. Recognition of the secondary infection with S. pneumoniae (4) activates the NF-B pathway, but NF-B-mediated induction of Cxcl1 is reduced by the repressive histone mark H3K9me3 established by Setdb2 at the Cxcl1 promoter (5). Reduced production of CXCL1 leads to reduced recruitment of neutrophils (6) and bacterial outgrowth due insufficient bacterial control (7). IFN-/βR, receptor for IFN- and IFN-β.
during this phase, take advantage of the downregulation of the immune response to invade and expand. Alternatively, repressing Cxcl1 and, as the authors suggest, other genes regulated by the transcription factor NF-κB, might be part of setting up the appropriate antiviral immune response by suppressing functions geared more toward antibacterial responses. This would mean the bacteria do not attack when the immune response is shutting down; instead, they strike when the immune response is tailored to a different type of challenge. An interesting detail in this study by Schliehe et al. is that they add the synthetic lipopeptide and Toll-like receptor 2 agonist Pam3CSK4 to the growing list of bacterial ligands able to elicit type I interferons1. Pam3CSK4 and other bacterial ligands, such as lipopolysaccharide or the dinucleotide CpG, known to induce type I interferons, should therefore lead to upregulation of Setdb2 in bacterial infection, which would limit the recruitment of neutrophils. Despite this, the effect of Setdb2 appears to be most prominent in coinfection rather than following infection with S. pneumoniae alone. It will be interesting to determine if the lack of difference between wild-type and Setdb2GT/GT mice in their anti–S. pneumoniae response
is a question of kinetics and timing, if findings will be similar for infection with other bacteria, including Gram-negative bacteria, and whether the induction and action of Setdb2 require a strong type I interferon stimulus such as that induced by virus but not by bacteria. Several exciting questions are raised by these studies. Is Cxcl1 one of only few targets of Setdb2, or is the transcriptional response to bacterial infection modulated more globally by Setdb2 activity? How long-lived is the effect of this modulation? It has been postulated that previous infections can condition responsiveness to unrelated microbes12 (a phenomenon sometimes referred to as ‘trained immunity’), and the mechanism shown here would certainly be a very plausible one to explain longterm changes. Alternatively, if these effects are of short duration, then other chromatinmodifying processes could lead to long-term conditioning of the lung immune response. Apart from macrophages, are other cell types in the lungs also affected by Setdb2 activity? The lack of Setdb2 expression in the absence of type I interferon signaling in mice deficient in the receptor for interferon-α (IFN-α) suggests that it is expressed mainly in cells of the immune system, whereas epithelial cells in the lungs might still rely on type III interferons to induce
nature immunology volume 16 number 1 january 2015
Setdb2 expression13. It will be interesting, however, to assess the role of Setdb2 in other cells of the immune system and in airway epithelial cells, which are known to respond to infection with influenza virus by producing a range of cytokine and chemoattractants, including CXCL1. Interestingly, Schliehe et al. show that the scavenger receptor Marco is induced more in macrophages from Setdb2GT/GT than in those from wild type mice1. Suppression of Marco has been described as an IFN-γ-dependent mechanism of an impaired antibacterial immune response in coinfection14. Therefore, the gene encoding Marco could be another target of Setdb2 that is induced not only by type I interferons but also by IFN-γ, as the authors show here1. Other promising areas of future research would be to determine what guides Setdb2 to the promoters of its target genes and which transcription factors and other chromatin-associated factors are involved upstream and downstream of Setdb2. Also, the possibility that proteins other than histone H3 are methylated by Setdb2 needs to be explored. In conclusion, the colds that beset humans or the pneumonias that confine people to bed during adulthood are only the latest in a lifelong string of infections that the respiratory immune system has had to deal with. These 7
n e w s a nd v i e w s infections possibly hit during a damp winter during which other microbes are never far away and some are already present in the lungs. The human immune system has certainly evolved under strong pressure to integrate past and simultaneous microbial encounters, to avoid overshooting as well as insufficient or inappropriate responses. Schliehe et al. delineate a novel molecular mechanism by which an ongoing immune response affects a subsequent one1. More studies like this will be needed to figure out how this sometimes undesired interference
works, not least to find better prophylactic and therapeutic interventions to stop the bacterial executioners in severe coinfection. COMPETING FINANCIAL INTERESTS The author declares no competing financial interests. 1. Schliehe C. et al. Nat. Immunol. 16, 67–74 (2015). 2. Cruveilhier, L. Ann. Inst. Pasteur 33, 448–461 (1919) 3. Trinchieri, G. J. Exp. Med. 207, 2053–2063 (2010). 4. Morens, D.M. et al. J. Infect. Dis. 198, 962–970 (2008). 5. Shrestha, S. et al. Sci. Transl. Med. 5, 191ra184 (2013).
6. McCullers, J.A. Nat. Rev. Microbiol. 12, 252–262 (2014). 7. Kudva, A. et al. J. Immunol. 186, 1666–1674 (2011). 8. Li, W. et al. J. Virol. 86, 12304–12312 (2012). 9. Navarini, A.A. et al. Proc. Natl. Acad. Sci. USA 103, 15535–15539 (2006). 10. Nakamura, S. et al. J. Clin. Invest. 121, 3657–3665 (2011). 11. Shahangian, A. et al. J. Clin. Invest. 119, 1910–1920 (2009). 12. Goulding, J. et al. Proc. Am. Thorac. Soc. 4, 618–625 (2007). 13. Crotta, S. et al. PLoS Pathog. 9, e1003773 (2013). 14. Sun, K. & Metzger, D.W. Nat. Med. 14, 558–564 (2008).
A surprising role for TLR7
npg
© 2015 Nature America, Inc. All rights reserved.
Michael M Lederman Ligation of the Toll-like receptor TLR7 in human CD4+ T cells elicits an anergic state that may contribute to CD4+ T cell hyporesponsiveness after infection with human immunodeficiency virus type 1 and may also enhance propagation of this virus.
T
LR7 is one of several Toll-like receptors (TLR3, TLR7, TLR8 and TLR9) that recognize microbial nucleic acid sequences. TLR7 and TLR8, which recognize singlestranded RNA, are distributed broadly among myeloid and other cells and are characteristically expressed in endosomal compartments, where their engagement with microbial sequences is thought to take place. In this issue of Nature Immunology, DominguezVillar et al. demonstrate a surprising role for TLR7 in host-virus relationships1. In a paper fraught with novelty, the authors provide evidence that engagement of TLR7 expressed in primary human CD4+ T cells can induce anergy. They also find that activation of a TLR7-dependent mechanism in CD4+ T cells promotes the replication of human immunodeficiency virus type 1 (HIV-1) within these cells—a hitherto unexpected role for this microbial sensor. The authors first report that incubation of human CD4+ T cells with various TLR7 agonists such as imiquimod impairs their proliferation and the expression of various cytokines in response to costimulation via the invariant signaling protein CD3 and coreceptor CD28. They do not obtain a similar effect with TLR8 agonists, and this effect of TLR7 stimulation is abrogated by silencing of TLR7-encoding RNA; therefore, the proanergic effect seems to be a specific property of TLR7 ligation. Michael M. Lederman is with Case Western Reserve University, University Hospitals Case Medical Center, Cleveland, Ohio, USA. e-mail: [email protected]
8
There is an emerging literature suggesting that non-synchronized activation of the T cell antigen receptor and costimulation results in calcium flux and induction of anergy through activation of the transcription factor NFAT1 (ref. 2). Dominguez-Villar et al. now show that engagement of TLR7 in CD4+ T cells results in induction of calcium flux and activation of the transcription factor NFATc2 via dephosphorylation that then drives the expression of a host of anergy-related genes1. This effect is abrogated by silencing of the gene encoding NFATc2. The engagement of TLR7 with imiquimod also results in phosphorylation of the transcription factor NF-κB and the signaling molecules IRAK4 and p38 but lower basal expression of the signaling molecule Jnk and less phosphorylation of Jnk and its target Jun (a component of the transcription factor AP-1) induced by the phorbol ester PMA plus ionomycin. Collectively these results offer two potential means by which stimulation via TLR7 diminishes the immunological responsiveness of CD4+ T cells: one by the induction of various anergy-related genes, and the other by interference with AP-1-dependent signals. To try to place their findings into a more physiological context, the authors seek to understand whether the hyporesponsive state of CD4+ T cells seen in HIV infection could be explained by that TLR7 effect. The authors confirm earlier reports that infection of CD4+ T cells with HIV-1 in vitro diminishes their ability to produce the cytokines IL-2 and IFN-γ after restimulation. Productive infection is apparently not
necessary for this anergic effect, as cells that have not supported expression of an HIV-1 reporter gene seem to be as compromised in these assays as are their neighboring productively infected cells. It might be a stretch to contend that this phenomenon drives the dysfunction of CD4+ T cells in untreated HIV infection, as an even smaller minority of CD4+ T cells seem to be HIV-1 infected in vivo. However, in the setting of unrestrained viral infection that is linked to both clinical evidence and laboratory evidence of CD4+ T cell dysfunction, it may be possible that many CD4+ T cells are in fact exposed to virus, viral proteins or viral sequences that do not reflect completion of the viral life cycle but still may engage the coreceptor CD4, TLR7 or other innate sensors3. The authors then report an unexpected role for engagement of TLR7 that facilitates the infection of CD4+ T cells with HIV-1, as silencing of the gene encoding TLR7 or blockade of this innate receptor with the inhibitor IRS661 attenuates infection. Here too a role for NFATc2 is suggested, as silencing of the gene encoding NFATc2 or blockade of NFATc2 itself also attenuates HIV-1 replication. The authors conclude that HIV-1 uses the anergic state induced by ligation of TLR7 to support its own propagation. As best as can be seen, their systems use broad activation of T cells to render target cells susceptible to in vitro infection, but still it is fascinating to propose that HIV has ‘learned’ to utilize for its own propagation in CD4+ T cells an innate sensor that is typically used by dendritic cells, monocytes and
volume 16 number 1 january 2015 nature immunology
Stop the executioners Andreas Wack
npg
© 2015 Nature America, Inc. All rights reserved.
Virus-triggered type I interferon induces the lysine methyltransferase Setdb2; this then generates repressive histone marks on the promoters of genes encoding molecules important for antibacterial immunity. This process can contribute to influenza virus–associated bacterial superinfection.
U
nlike experimental mice that live in specific pathogen-free conditions, humans walk the dirty streets of the real world. This means that the human immune system must cope with constant polymicrobial exposure, in contrast to the majority of infection models, in which a single pathogen is applied in an otherwise clean environment. In this issue of Nature Immunology, Schliehe et al. present mechanistic data describing the events underlying an experimental model of polymicrobial exposure and bacterial superinfection1. In real life, microbes do not form an orderly queue; instead, opportunistic and colonizing species often seize their chance to invade and proliferate while an immune response to another pathogen is ongoing. This is demonstrated perhaps most dramatically in bacterial superinfections following infection with influenza virus, a phenomenon observed as early as during the 1918 influenza pandemic, which killed far more people than the First World War itself. Eminent pathologists of the time were aware of the high frequency of superinfections during this public health emergency2, which prompted one of them, Louis Cruveilhier, to state in 1919: “If influenza condemns, additional infection executes.”3,4 The actual executioners in this deadly process were soon to be identified as Streptococcus pneumoniae, Staphylococcus aureus and other bacteria that are commonly found even in healthy people but can cause severe and lethal pneumonias in influenza virus–infected people4. Severe coinfections are not restricted to influenza pandemics or infections with highly pathogenic Division of Immunoregulation, Medical Research Council, National Institute for Medical Research, Mill Hill, London, UK. e-mail: [email protected]
6
strains of influenza virus but are an important component of morbidity and mortality in every influenza season, including those of moderate severity5. Studies have shown that the impressive synergy between influenza virus and respiratory bacteria observed in the clinic can also be modeled in mice. Broadly, the underlying mechanisms identified so far can be categorized into two sets6: virus-induced damage of lung epithelia that leads to impaired clearance of bacteria, or suppression of antibacterial immune responses by preceding infection with influenza virus. In the latter category, type I interferon–mediated suppression of interleukin 17 (IL-17) responses by canonical αβ T cells or γδ T cells present in the lung tissue7,8 or type I interferon–dependent suppression of macrophage or neutrophil action9–11 has been shown to mediate the impaired antibacterial response following infection with influenza virus. In particular, one of the mechanisms of diminished neutrophil responses after infection with influenza virus is a type I interferon– mediated decrease in the abundance of the neutrophil chemoattractant CXCL1 (KC)11. Although it is becoming increasingly clear what cell types and cytokines are involved in influenza virus–induced impairment of antibacterial immunity, the study by Schliehe et al.1 is among the first to attempt a description of the molecular mechanism underlying the suppression of the recruitment of neutrophils into the lungs. They show that one of the interferon-stimulated genes upregulated in the influenza virus–infected lungs is the lysine methyltransferase Setdb2. This enzyme is involved in generating the suppressive histone mark H3K9me3 by catalyzing tri methylation of histone 3 at Lys9. Cells from mice with a hypomorphic gene-trap construct of Setdb2 (Setdb2GT/GT mice) have much less Setdb2 protein and higher expression of a
range of genes, including Cxcl1, than do wild-type cells, upon stimulation with a variety of Toll-like receptor agonists. Chromatin immunoprecipitation indicates that Setdb2 binds to the Cxcl1 promoter upon cell stimulation, which leads to an increase in the repressive H3K9me3 mark at this promoter; this increase is found in macrophages from wild-type mice but not those from Setdb2GT/GT mice. In vivo, this translates into higher expression of CXCL1 induced by lipopolysaccharide or influenza virus and more inflammation in Setdb2GT/GT lungs than in wild-type lungs. Notably, this lack of repression of CXCL1 also shows protective effects when the authors assess influenza virus–bacteria1 coinfection. Setdb2GT/GT mice infected with influenza virus cope better with subsequent exposure to S. pneumoniae than do influenza virus–infected wild-type mice because they have better recruitment of neutrophils and, consequently, lower bacterial loads in the lungs (Fig. 1). The authors offer a molecular explanation for why stopping the bacterial executioners becomes difficult after a preceding infection with influenza virus: interferon-dependent upregulation of Setdb2 leads to a repressive histone mark on genes such as Cxcl1, which encode molecules that are important in the antibacterial response. This will reduce the recruitment of neutrophils and thus the ability of the host to control bacterial infection. However, it remains unclear why evolution has installed this interferon-induced suppressive mechanism in the first place. It might be part of a damage-limitation effort at the end of a strong inflammatory process associated with interferon-driven antiviral immune responses, and the bacteria, which either are already present in the lungs (in healthy carriers) or enter
volume 16 number 1 january 2015 nature immunology
n e w s a nd v i e w s
Influenza virus
S. pneumoniae 1
1
b Lung
6
2
IFN-α and IFN-β
Elimination of bacteria
Bacterial outgrowth
7 Neutrophil recruitment
CXCL1
Lung macrophage 2
5
PRR
Neutrophil
NF-κB Nucleus
PRR
IFN-α/βR
Less secretion of CXCL1
4 Setdb2↑ 3
6
Diminished neutrophil recruitment
Marina Corral Spence/Nature Publishing Group
a
npg
© 2015 Nature America, Inc. All rights reserved.
3
4
5
NF-κB
DNA CXCL1
H3
Less CXCL1
Figure 1 Infection with influenza virus reduces the immune response to secondary bacterial infection by a Setdb2-dependent mechanism. (a) Infection with S. pneumoniae (1) leads to the recognition of bacterial ligands by pattern-recognition receptors (PRRs) (2), which triggers activation of NF-B and its translocation to the nucleus (3). NF-B-mediated induction of Cxcl1 (4) leads to the production and secretion of CXCL1 and to the recruitment of neutrophils (5), which control bacterial spread and eliminate bacteria (6). (b) During coinfection, infection with influenza virus (1) leads to the production of type I interferons (INF- and IFN-β) by infected airway epithelial cells (2) and other cells, which induces Setdb2 production in lung macrophages (3) and other cells. Recognition of the secondary infection with S. pneumoniae (4) activates the NF-B pathway, but NF-B-mediated induction of Cxcl1 is reduced by the repressive histone mark H3K9me3 established by Setdb2 at the Cxcl1 promoter (5). Reduced production of CXCL1 leads to reduced recruitment of neutrophils (6) and bacterial outgrowth due insufficient bacterial control (7). IFN-/βR, receptor for IFN- and IFN-β.
during this phase, take advantage of the downregulation of the immune response to invade and expand. Alternatively, repressing Cxcl1 and, as the authors suggest, other genes regulated by the transcription factor NF-κB, might be part of setting up the appropriate antiviral immune response by suppressing functions geared more toward antibacterial responses. This would mean the bacteria do not attack when the immune response is shutting down; instead, they strike when the immune response is tailored to a different type of challenge. An interesting detail in this study by Schliehe et al. is that they add the synthetic lipopeptide and Toll-like receptor 2 agonist Pam3CSK4 to the growing list of bacterial ligands able to elicit type I interferons1. Pam3CSK4 and other bacterial ligands, such as lipopolysaccharide or the dinucleotide CpG, known to induce type I interferons, should therefore lead to upregulation of Setdb2 in bacterial infection, which would limit the recruitment of neutrophils. Despite this, the effect of Setdb2 appears to be most prominent in coinfection rather than following infection with S. pneumoniae alone. It will be interesting to determine if the lack of difference between wild-type and Setdb2GT/GT mice in their anti–S. pneumoniae response
is a question of kinetics and timing, if findings will be similar for infection with other bacteria, including Gram-negative bacteria, and whether the induction and action of Setdb2 require a strong type I interferon stimulus such as that induced by virus but not by bacteria. Several exciting questions are raised by these studies. Is Cxcl1 one of only few targets of Setdb2, or is the transcriptional response to bacterial infection modulated more globally by Setdb2 activity? How long-lived is the effect of this modulation? It has been postulated that previous infections can condition responsiveness to unrelated microbes12 (a phenomenon sometimes referred to as ‘trained immunity’), and the mechanism shown here would certainly be a very plausible one to explain longterm changes. Alternatively, if these effects are of short duration, then other chromatinmodifying processes could lead to long-term conditioning of the lung immune response. Apart from macrophages, are other cell types in the lungs also affected by Setdb2 activity? The lack of Setdb2 expression in the absence of type I interferon signaling in mice deficient in the receptor for interferon-α (IFN-α) suggests that it is expressed mainly in cells of the immune system, whereas epithelial cells in the lungs might still rely on type III interferons to induce
nature immunology volume 16 number 1 january 2015
Setdb2 expression13. It will be interesting, however, to assess the role of Setdb2 in other cells of the immune system and in airway epithelial cells, which are known to respond to infection with influenza virus by producing a range of cytokine and chemoattractants, including CXCL1. Interestingly, Schliehe et al. show that the scavenger receptor Marco is induced more in macrophages from Setdb2GT/GT than in those from wild type mice1. Suppression of Marco has been described as an IFN-γ-dependent mechanism of an impaired antibacterial immune response in coinfection14. Therefore, the gene encoding Marco could be another target of Setdb2 that is induced not only by type I interferons but also by IFN-γ, as the authors show here1. Other promising areas of future research would be to determine what guides Setdb2 to the promoters of its target genes and which transcription factors and other chromatin-associated factors are involved upstream and downstream of Setdb2. Also, the possibility that proteins other than histone H3 are methylated by Setdb2 needs to be explored. In conclusion, the colds that beset humans or the pneumonias that confine people to bed during adulthood are only the latest in a lifelong string of infections that the respiratory immune system has had to deal with. These 7
n e w s a nd v i e w s infections possibly hit during a damp winter during which other microbes are never far away and some are already present in the lungs. The human immune system has certainly evolved under strong pressure to integrate past and simultaneous microbial encounters, to avoid overshooting as well as insufficient or inappropriate responses. Schliehe et al. delineate a novel molecular mechanism by which an ongoing immune response affects a subsequent one1. More studies like this will be needed to figure out how this sometimes undesired interference
works, not least to find better prophylactic and therapeutic interventions to stop the bacterial executioners in severe coinfection. COMPETING FINANCIAL INTERESTS The author declares no competing financial interests. 1. Schliehe C. et al. Nat. Immunol. 16, 67–74 (2015). 2. Cruveilhier, L. Ann. Inst. Pasteur 33, 448–461 (1919) 3. Trinchieri, G. J. Exp. Med. 207, 2053–2063 (2010). 4. Morens, D.M. et al. J. Infect. Dis. 198, 962–970 (2008). 5. Shrestha, S. et al. Sci. Transl. Med. 5, 191ra184 (2013).
6. McCullers, J.A. Nat. Rev. Microbiol. 12, 252–262 (2014). 7. Kudva, A. et al. J. Immunol. 186, 1666–1674 (2011). 8. Li, W. et al. J. Virol. 86, 12304–12312 (2012). 9. Navarini, A.A. et al. Proc. Natl. Acad. Sci. USA 103, 15535–15539 (2006). 10. Nakamura, S. et al. J. Clin. Invest. 121, 3657–3665 (2011). 11. Shahangian, A. et al. J. Clin. Invest. 119, 1910–1920 (2009). 12. Goulding, J. et al. Proc. Am. Thorac. Soc. 4, 618–625 (2007). 13. Crotta, S. et al. PLoS Pathog. 9, e1003773 (2013). 14. Sun, K. & Metzger, D.W. Nat. Med. 14, 558–564 (2008).
A surprising role for TLR7
npg
© 2015 Nature America, Inc. All rights reserved.
Michael M Lederman Ligation of the Toll-like receptor TLR7 in human CD4+ T cells elicits an anergic state that may contribute to CD4+ T cell hyporesponsiveness after infection with human immunodeficiency virus type 1 and may also enhance propagation of this virus.
T
LR7 is one of several Toll-like receptors (TLR3, TLR7, TLR8 and TLR9) that recognize microbial nucleic acid sequences. TLR7 and TLR8, which recognize singlestranded RNA, are distributed broadly among myeloid and other cells and are characteristically expressed in endosomal compartments, where their engagement with microbial sequences is thought to take place. In this issue of Nature Immunology, DominguezVillar et al. demonstrate a surprising role for TLR7 in host-virus relationships1. In a paper fraught with novelty, the authors provide evidence that engagement of TLR7 expressed in primary human CD4+ T cells can induce anergy. They also find that activation of a TLR7-dependent mechanism in CD4+ T cells promotes the replication of human immunodeficiency virus type 1 (HIV-1) within these cells—a hitherto unexpected role for this microbial sensor. The authors first report that incubation of human CD4+ T cells with various TLR7 agonists such as imiquimod impairs their proliferation and the expression of various cytokines in response to costimulation via the invariant signaling protein CD3 and coreceptor CD28. They do not obtain a similar effect with TLR8 agonists, and this effect of TLR7 stimulation is abrogated by silencing of TLR7-encoding RNA; therefore, the proanergic effect seems to be a specific property of TLR7 ligation. Michael M. Lederman is with Case Western Reserve University, University Hospitals Case Medical Center, Cleveland, Ohio, USA. e-mail: [email protected]
8
There is an emerging literature suggesting that non-synchronized activation of the T cell antigen receptor and costimulation results in calcium flux and induction of anergy through activation of the transcription factor NFAT1 (ref. 2). Dominguez-Villar et al. now show that engagement of TLR7 in CD4+ T cells results in induction of calcium flux and activation of the transcription factor NFATc2 via dephosphorylation that then drives the expression of a host of anergy-related genes1. This effect is abrogated by silencing of the gene encoding NFATc2. The engagement of TLR7 with imiquimod also results in phosphorylation of the transcription factor NF-κB and the signaling molecules IRAK4 and p38 but lower basal expression of the signaling molecule Jnk and less phosphorylation of Jnk and its target Jun (a component of the transcription factor AP-1) induced by the phorbol ester PMA plus ionomycin. Collectively these results offer two potential means by which stimulation via TLR7 diminishes the immunological responsiveness of CD4+ T cells: one by the induction of various anergy-related genes, and the other by interference with AP-1-dependent signals. To try to place their findings into a more physiological context, the authors seek to understand whether the hyporesponsive state of CD4+ T cells seen in HIV infection could be explained by that TLR7 effect. The authors confirm earlier reports that infection of CD4+ T cells with HIV-1 in vitro diminishes their ability to produce the cytokines IL-2 and IFN-γ after restimulation. Productive infection is apparently not
necessary for this anergic effect, as cells that have not supported expression of an HIV-1 reporter gene seem to be as compromised in these assays as are their neighboring productively infected cells. It might be a stretch to contend that this phenomenon drives the dysfunction of CD4+ T cells in untreated HIV infection, as an even smaller minority of CD4+ T cells seem to be HIV-1 infected in vivo. However, in the setting of unrestrained viral infection that is linked to both clinical evidence and laboratory evidence of CD4+ T cell dysfunction, it may be possible that many CD4+ T cells are in fact exposed to virus, viral proteins or viral sequences that do not reflect completion of the viral life cycle but still may engage the coreceptor CD4, TLR7 or other innate sensors3. The authors then report an unexpected role for engagement of TLR7 that facilitates the infection of CD4+ T cells with HIV-1, as silencing of the gene encoding TLR7 or blockade of this innate receptor with the inhibitor IRS661 attenuates infection. Here too a role for NFATc2 is suggested, as silencing of the gene encoding NFATc2 or blockade of NFATc2 itself also attenuates HIV-1 replication. The authors conclude that HIV-1 uses the anergic state induced by ligation of TLR7 to support its own propagation. As best as can be seen, their systems use broad activation of T cells to render target cells susceptible to in vitro infection, but still it is fascinating to propose that HIV has ‘learned’ to utilize for its own propagation in CD4+ T cells an innate sensor that is typically used by dendritic cells, monocytes and
volume 16 number 1 january 2015 nature immunology
