Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Cell. Author manuscript; available in PMC 2017 April 21.
Published in final edited form as:
PMCID: PMC4842125

Innate lymphocyte/Ly6Chi monocyte crosstalk promotes Klebsiella pneumoniae clearance


Increasing antibiotic resistance among bacterial pathogens has rendered some infections untreatable with available antibiotics. Klebsiella pneumoniae, a bacterial pathogen that has acquired high-level antibiotic resistance, is a common cause of pulmonary infections. Optimal clearance of K. pneumoniae from the host lung requires TNF and IL-17A. Herein we demonstrate that inflammatory monocytes are rapidly recruited to the lungs of K. pneumoniae infected mice, and produce TNF, which markedly increases the frequency of IL-17-producing innate lymphoid cells. While pulmonary clearance of K. pneumoniae is preserved in neutrophil-depleted mice, monocyte depletion or TNF deficiency impairs IL-17A–dependent resolution of pneumonia. Monocyte-mediated bacterial uptake and killing is enhanced by ILC production of IL-17A, indicating that innate lymphocytes engage in a positive feedback loop with monocytes that promotes clearance of pneumonia. Innate immune defense against a highly antibiotic-resistant bacterial pathogen depends on crosstalk between inflammatory monocytes and innate lymphocytes that is mediated by TNF and IL-17A.


Klebsiella pneumoniae is a Gram-negative bacterium that normally resides in the lower gastrointestinal tract. In patients with compromised immune defenses or impaired pulmonary clearance mechanisms, K. pneumoniae can cause severe infections of the lower respiratory tract. The emergence of high-level antibiotic resistance in some strains of K. pneumoniae is limiting treatment options and, in some cases, rendering infections untreatable with available antibiotics. In the absence of effective antibiotics, optimizing or enhancing host immune defenses against K. pneumoniae represents a potential therapeutic option that might improve clinical outcomes.

As a bacterial species, Klebsiella pneumoniae is diverse and composed of a wide array of strains that differ in capsule composition, antibiotic resistance, mucoid phenotype and virulence. Extensive studies with the rodent-adapted, antibiotic-sensitive strain 43816 of K. pneumoniae have demonstrated that TNF, neutrophils, IL17A, MyD88 and C-type lectins contribute to bacterial clearance from the mouse lung (Cai et al., 2009; Laichalk et al., 1996; Moore et al., 2005; Sharma et al., 2014; Steichen et al., 2013; Ye et al., 2001). Which cell types produce defense-associated cytokines and which mediate bacterial clearance is incompletely defined, however. Further complicating the picture is the recent discovery that antibiotic-resistant strains of K. pneumoniae isolated from patients differ in their dependence on neutrophils or inflammatory monocytes for pulmonary clearance from infected mice (Xiong et al., 2015). While clearance of 5 distinct strains of K. pneumoniae was consistently dependent on inflammatory monocytes, the contribution of neutrophils to clearance was more variable.

Inflammatory monocytes (IMs) are bone marrow residing leukocytes that, in the setting of infection, rapidly traffic into the bloodstream and circulate to inflamed or infected tissues. IMs are pluripotent and, depending on the inflammatory environment they infiltrate, can acquire phenotypes that extend from pro-inflammatory TNF and iNOS production to immunosuppressive IL10-production (Biswas and Mantovani, 2010; Gordon and Taylor, 2005; Serbina et al., 2003). Although IMs enhance clearance of a wide range of pathogens, the exact mechanisms by which they promote microbial clearance remain unclear. Because TNF is essential for defense against intracellular pathogens such as Listeria monocytogenes, and IMs can be prodigious producers of TNF, it has seemed likely that activated IMs mediate at least some level of bacterial clearance by producing TNF. However, the mechanism by which TNF enhances bacterial clearance remains largely undefined.

Pulmonary infection of mice with K. pneumoniae was one of the first infectious disease models to demonstrate the importance of IL-17A for host defense against extracellular bacterial infections (Ye et al., 2001). Indeed, IL-17A–deficient mice rapidly succumbed to infection with K. pneumoniae strain 43816 and neutrophils were essential for bacterial clearance. While Th17 CD4 T lymphocytes and γδ T cells produce IL-17A (Harrington et al., 2005; Ivanov et al., 2009; Liang et al., 2006; Sutton et al., 2009), innate lymphocytes, at least in the gut, appear to be responsible for the major IL-17A–mediated antimicrobial effects (Artis and Spits, 2015). Innate lymphocytes are essential for defense against intestinal infections with Citrobacter rodentium and Clostridium difficile (Abt et al., 2015; Satoh-Takayama et al., 2008; Sonnenberg et al., 2011; Zheng et al., 2008), however their role in defense against pulmonary infection is less clear. In fact, ILC3s in the respiratory system have not been identified until recently (Kim et al., 2014; Van Maele et al., 2014). Furthermore, whether inflammatory monocytes and IL-17A–producing cells facilitate each other’s effecter functions in response to infection is unknown.

To begin to address the roles of IMs and innate lymphocytes in defense against bacterial pulmonary infection, we investigated mice infected with a highly antibiotic resistant strain of Klebsiella pneumonia that is cleared by neutrophil-independent but monocyte-dependent mechanisms (Xiong et al., 2015). We demonstrate that IMs, upon recruitment to the infected lung, produce TNF, which increases innate lymphocyte frequencies and IL-17A production. We show that IL-17A promotes monocyte mediated uptake and killing of K. pneumoniae. Our experiments demonstrate that rapid recruitment of IMs to the lung markedly enhances innate lymphocyte responses, and that innate lymphocytes engage in crosstalk with IMs by producing IL-17A to enhance bacterial killing.


Inflammatory monocytes provide defense against Klebsiella pneumoniae

Neutrophils and monocytes are rapidly recruited to sites of bacterial infection. To determine the kinetics of inflammatory cell recruitment to the lungs, we inoculated mice intratracheally with a sublethal dose of K. pneumoniae and quantified inflammatory monocytes (Ly6Chi) and neutrophils (Ly6G+) by flow cytometry. Both cell populations increased in frequency within 3 hours of inoculation, with cell numbers peaking at 24 hours and then decreasing (Fig. 1A). To determine the contribution of neutrophils and monocytes to bacterial clearance from the lungs, we depleted both populations with the anti-Gr1 RB6–8C5 monoclonal antibody (αGr1) or depleted only neutrophils by administering the Ly6G–specific 1A8 monoclonal antibody (αLy6G), or only CCR2+ monocytes by administering diphtheria toxin (DT) to CCR2-DTR mice. Monocyte depletion in K. pneumoniae infected mice resulted in increased mortality, weight loss and hypothermia in comparison to undepleted mice (Fig. 1B and D). Confirming our previous results (Xiong et al., 2015), we found that monocyte depletion resulted in a roughly 10-fold increase in bacterial CFUs in the lung while neutrophil depletion modestly reduced bacterial CFUs (Fig. 1C). Bacterial dissemination to mediastinal lymph nodes and spleen was also increased in monocyte-depleted mice and the rate of bacterial clearance was reduced (Fig. S1A). Monocyte-depleted CCR2-DTR mice also had higher levels of neutrophil infiltration to the bronchoalveolar lavage fluid (BALF) despite compromised bacterial clearance (Fig. S1B).

Figure 1
Deficiency of K. pneumoniae clearance in monocyte-depleted mice

Although predominantly expressed on inflammatory monocytes, CCR2 is also expressed on some NK cells, monocyte/dendritic cell progenitors and hematopoietic stem cells. Therefore, in order to determine whether monocytes are mediating protection against K. pneumoniae, we adoptively transferred highly purified CD45.2+ monocytes into monocyte-depleted CD45.1+ CCR2-DTR depleter mice 30 minutes after challenge with K. pneumoniae (Fig. 2A). IMs were purified from CCR2-GFP mice based on GFP expression and the lack of expression of FLT3, C-kit, NK1.1, CD19 or CD3. This purification approach circumvented the need for antibody staining of surface markers on IMs (such as CCR2, CD11b and Ly6C), which might interfere with IM trafficking, while eliminating contaminating cell populations and yielded a purity of greater than 99% Ly6Chi IMs (Fig. S2A). Transferred CD45.2+ monocytes were detected in the lungs of K. pneumoniae-infected recipient mice as CD11c+ CD11b+ CD103neg (Fig. 2B, 2D and S2C), and achieved numbers that approached those seen in equivalently infected wild type mice (Fig. 2C). Transferred monocytes isolated from infected lungs had down-regulated Ly6C and up-regulated MHC class II, CD11b and CD11c (Fig. 2D and S2D) while transferred monocytes isolated from spleen or bone marrow remained Ly6Chi (Fig. S2D). A single transfer of IMs reduced bacterial CFUs in the lungs of infected, monocyte depleted CCR2-DTR mice (Fig. 2E and S2B), supporting a role of monocytes in the clearance of K. pneumoniae from the lung.

Figure 2
Adoptively transferred IMs promote bacterial clearance

Activated monocytes are predominant producers of TNF

TNF is required for optimal clearance of K. pneumoniae from the lung and IMs, at least during some bacterial infections, are major producers of TNF (Laichalk et al., 1996; Serbina et al., 2003). We found that CD11b+CD103neg activated monocytes constitute the predominant TNF-producing cell population in K. pneumoniae-infected lungs (Fig. 3A–B and S3A). While the most frequent TNF-expressing cells were CD11b+CD103neg, a smaller subset of CD11bnegCD103neg cells also produced TNF during K. pneumoniae infection (Fig. 3B). Treatment of CCR2-DTR mice with DT eliminated the CD11b+CD103neg monocyte population in the lungs of K. pneumoniae-infected mice and markedly reduced overall TNF production (Fig. 3A). We found that when challenged with a higher dose of K. pneumoniae, wild type mice varied with respect to the frequency of CD11b+CD103neg TNF-producing cells in the lung, enabling us to stratify them into high and low TNF producers. Although it might be expected that high levels of TNF would correlate with more severe infection, we found that mice with higher numbers of TNF producing CD11b+CD103neg cells had reduced CFUs in the lungs (Fig. S3B), suggesting that TNF contributes to bacterial clearance. To confirm the function of TNF on K. pneumoniae clearance, we infected TNF−/− and TNFR1−/− mice with K. pneumoniae and observed impaired bacterial clearance with higher bacterial CFUs in the lungs of both groups (Fig. 3C), consistent with previous studies demonstrating the role of TNF in defense against this pathogen. To determine whether TNF produced by IMs is critical for optimal clearance of K. pneumoniae, we transplanted wild type recipient mice with a 50/50 mixture of bone marrow from TNF−/− and CCR2-DTR mice or WT and CCR2-DTR mice and treated mice with DT prior to infection. This mixed bone marrow chimera approach demonstrated that depletion of monocytes from TNF−/−/CCR2-DTR mice markedly impaired clearance of K. pneumoniae from the lung, while bacterial clearance following monocyte depletion from WT/CCR2-DTR mice was normal (Fig. S3C). The cell population responding to TNF via TNFR1 was not depleted by DT treatment in TNFR1−/−/CCR2-DTR mixed bone marrow chimeric mice (Fig. S3C), suggesting that TNF-responsive cells mediating K. pneumoniae clearance do not express CCR2 and thus are not IMs. In our previous work we described that MyD88 signaling was required for egress of inflammatory monocytes from the bone marrow upon challenge with TLR ligands (Shi et al., 2011). To better understand the role of MyD88 signaling during K. pneumoniae infection model, we infected WT and MyD88−/− mice with Kp-MH258, and measured clearance from the lung. MyD88−/− mice had approximately 10 fold more bacteria in the lung than WT mice, indicating that MyD88 signaling is critical for defense (Fig. S3D). MyD88−/− mice had significantly reduced TNF production (Fig. S3E). These results indicate that MyD88-signaling plays an important role in inducing TNF production in response to K. pneumoniae invasion.

Figure 3
TNF production by activated IMs

TNF potentiates IL-17A production by innate lymphocytes

Consistent with previous studies demonstrating that IL-17A plays a key role in pulmonary clearance of K. pneumoniae infection, we found that in vivo administration of an IL-17A blocking monoclonal antibody (αIL-17A) increased mortality in K. pneumoniae infected mice (Fig. 4A). IL-17A blockade or infection of IL-17A−/− mice resulted in higher CFUs in lungs following K. pneumoniae infection (Fig. 4B). IL-17A blockade or deficiency did not reduce recruitment of neutrophils to infected lungs (Fig. S4B) following K. pneumoniae infection, suggesting that the protective effect of IL-17 does not result from enhanced neutrophil recruitment. Because deficiency of monocytes, TNF or IL-17A similarly increased susceptibility to K. pneumoniae infection, we next investigated IL-17A production in K. pneumoniae-infected TNFR1−/− and DT-treated CCR2-DTR mice. We found significantly reduced IL-17A production in infected TNFR1−/−and DT-treated CCR2-DTR mice when compared to infected WT mice, suggesting that TNF-signaling and CCR2+ cells are required for optimal IL-17A expression (Fig. 4C). Although IL-17A blockade enhanced infection in WT mice, treatment of TNF−/− or monocyte-depleted CCR2-DTR mice with IL-17A blocking antibody did not worsen K. pneumoniae infection (Fig. 4D), suggesting that TNF and IL-17A function in series within the same antimicrobial defense pathway. Kinetic analysis of TNF and IL-17A production and monocyte recruitment during the first 3 hours following pulmonary K. pneumoniae infection suggested that monocyte recruitment and TNF-production preceded induction of IL-17A production (Fig. 4E).

Figure 4
TNF potentiates IL-17 production

IL-17A can be produced by Th17 CD4 T cells and by type 3 innate lymphocytes (ILC3). Induction of IL-17A expression during the first two days of pulmonary K. pneumoniae infection suggested that ILC3 cells could be the major IL-17A producers. Clearance of K. pneumoniae from lungs was similar in Rag2−/− and WT mice (Fig. 5A) but markedly reduced in Rag2/Common Gamma Chain double knockout (Rag2/cγc−/−) mice, which lack all ILC subsets (Fig. 5B). Mortality following K. pneumoniae infection of Rag2/cγc−/− mice was markedly increased when compared to either WT or Rag2−/− mice (Fig. 5C). Depletion of ILCs in Rag2−/− mice with a depleting monoclonal antibody specific for CD90 resulted in markedly increased pulmonary K. pneumoniae CFUs following infection (Fig. 5D) and enhanced weight loss and mortality (Fig. S5A–B), indicating that ILCs enhance bacterial clearance. Although ILC3 cells can produce IL-17A, they are also known to produce IL-22, a cytokine that contributes to antibacterial defenses. However, IL-22 does not detectably contribute to early defense against K. pneumoniae infection, since IL-22−/− and WT mice cleared infection similarly (Fig. S4C).

Figure 5
Innate lymphoid cells protect against K. pneumoniae infection

ILC3-mediated clearance of K. pneumoniae is IL-17A–dependent

Depletion of ILCs from Rag2−/− mice reduced clearance of K. pneumoniae from the lung (Fig. 5D) and blockade of IL-17A also rendered Rag2−/− mice highly susceptible to infection (Fig. 6A). As expected, IL-17A expressing cells from the lungs of Rag2−/− mice expressed CD90, RORγt and CCR6 (Fig. 6B and Fig. S4A), and were otherwise lineage-negative (Fig. S4A), which is characteristic of ILC3 cells. No other major immune cell types produced IL-17A (Fig. S6A). Exogenously administered, recombinant murine IL-17 (rmIL-17) corrected the clearance defect in K. pneumoniae-infected Rag2/cγc−/− mice (Fig. 6C) and reduced mortality (Fig. 6D). Adoptive transfer of ILCs from Rag2−/−, but not IL-17A/Rag2−/− mice, reduced the bacterial burden in the lungs of recipient Rag2/cγc−/− mice (Fig. S4D). Since ILC3s are the only ILC subset that produces IL-17, this result demonstrates that protection is attributable to IL-17-producing ILC3s. To determine whether CCR2+ monocytes and IL-17 producing ILC3 cells contribute to the same bacterial clearance pathway in series, as noted previously for monocytes, TNF and IL-17 (Fig. 4D), we generated Rag2/cγc−/− CCR2-DTR mice and challenged them with K. pneumoniae. Fig. 6E demonstrates that monocyte depletion in Rag2/cγc−/−CCR2-DTR did not result in increased CFUs compared to Rag2/cγc−/− mice, suggesting that CCR2+ monocytes and ILC3 cells mediate K. pneumoniae clearance along the same functional axis (Fig. 6E). Depletion of IMs or TNF blockade reduced the recruitment of ILCs to the lung (Fig. 6F). The increased frequency of ILCs following infection was not the result of increased proliferation as BrDU incorporation was similarly low in ILCs from uninfected and K. pneumoniae infected mice (Fig. S6B). Transcription of CCL20, the ligand for CCR6, was reduced in the lungs of IM-depleted or TNF-blocked animals (Fig. 6G), suggesting that IM-derived TNF enhances CCL20 expression, thereby facilitating recruitment of CCR6+ ILC3s. A single transfer of purified IMs into K. pneumoniae infected, monocyte-depleted mice increased the frequency of total ILCs and, in particular, the frequency of RORγt+ ILC3s, but not GATA3+ ILC2s in the lung, providing further evidence for the role of IMs in the recruitment of ILC3s to the site of infection (Fig. 6H, 6I).

Figure 6
ILC3-mediated defenses through IL-17

Kp-MH258 is Carbapenem-resistant and requires IMs but not neutrophils for clearance (Xiong et al., 2015). We wanted to examine whether IM-ILC mediated defense mechanisms and IL-17 enhancement apply to the other clinical strains. To this end, we tested a different K. pneumoniae isolate, Kp-MH1867, which is highly antibiotic-resistant but sensitive to Carbapenems, and which depends on both IMs and neutrophils for clearance (Xiong et al., 2015). Both IL-17 deficiency and IM depletion rendered mice more susceptible to Kp-MH1867 (Fig. S5C), and the IL-17 response to infection was dampened by IM depletion in CCR2-DTR mice (Fig. S5D). Furthermore, IL-17A+ cells in the infected lung were LineageCD45+ CD90+, indicating that they were ILCs (Fig. S5E), and depletion of ILCs by anti-CD90 depleting antibody led to a higher bacterial burden (Fig. S5F). These data indicate that inflammatory monocytes and ILCs contribute to defense against Carbapenem-resistant and sensitive strains.

IL-17A enhances microbicidal activity of inflammatory monocytes

Although IL-17A–mediated signaling during bacterial infection has been shown to enhance neutrophil recruitment and activation, we found that neutrophil depletion during K. pneumoniae infection did not reduce clearance of Kp-MH258 (Fig. 1B and C), suggesting that IL-17A production is activating neutrophil-independent clearance mechanisms. Staining of cell populations isolated from K. pneumoniae-infected lungs demonstrated that IL-17 Receptor A (IL-17RA) is highly expressed on pulmonary epithelial cells but that IMs also express high levels of IL-17RA, while neutrophils and CD103+CD11bneg DCs express lower levels (Fig. 7A). To determine whether IMs harbor live K. pneumoniae during pulmonary infection, we purified IMs from infected WT and IL-17A−/− mice and quantified CFUs (Fig. 7B and S7B). IMs isolated from WT mice harbored more CFUs than IMs isolated from IL-17A−/−mice, suggesting that IL-17A is either enhancing IM-uptake of live bacteria or, alternatively, reducing killing of bacteria (Fig. 7B). To determine whether IL-17RA signaling in IMs impacts their ability to clear K. pneumoniae, we purified K. pneumoniae-infected IMs (IMs/Kp) from infected WT or IL-17A−/− mice and transferred them into uninfected recipient mice. Mice that received IMs/Kp from IL-17A−/− mice were also treated with blocking αIL-17A antibody in order to prevent IL-17RA signaling following adoptive transfer (Fig. S7A). Clearance of IM-associated bacteria that were transferred into recipient mice was measured after 2 hours and it revealed increased survival and expansion of IM-associated K. pneumoniae in the absence compared to the presence of IL-17A (Fig. 7C). Short-term IL-17A blockade in recipient mice did not impact in vivo bacterial CFUs following intratracheal inoculation of free K. pneumoniae bacteria (free Kp, Fig. S7A and and7C),7C), indicating that IL-17A enhances IM-mediated bacterial inhibition. Our results indicate that IL-17A–mediated signaling in IMs enhances uptake of K. pneumoniae, potentially by increasing phagocytosis, and facilitates bacterial killing. In order to investigate the IL-17-mediated bactericidal mechanism of IMs, we developed an in-vitro bacterial killing assay by co-incubating purified IMs and K. pneumoniae. Reduction of CFUs was observed in the presence of IMs, suggesting the capability of IMs to kill the pathogen in-vitro (Fig. 7D). Importantly, addition of IL-17 to the system further reduced CFUs and increased production of reactive oxygen species (ROS) by IMs (Fig. 7E). These results suggest that IL-17 directly enhances the microbicidal activity of IMs by potentiating ROS production. In addition, TNF upregulation in response to K. pneumoniae infection was reduced in IL-17A−/−compared to WT animals (Fig. 7F), indicating that IL-17, in addition to reinforcing the bactericidal activity of IMs, also contributes to enhancing TNF production by IMs. Addition of TNF to IM-K. pneumoniae co-cultures didn’t reduce CFUs (Fig. S7C), suggesting that, unlike IL-17, TNF does not directly enhance the bactericidal capability of IMs. This is consistent with the TNFR1−/−/CCR2-DTR bone marrow chimera experiments showing that the TNF-responsive cells are not CCR2+ (Fig. S3C).

Figure 7
Enhanced antimicrobial activities of IMs by IL-17


While TNF stands as one of the most impactful cytokines in defense against a wide range of infections, it also plays a central role in the pathogenesis of an array of inflammatory diseases. Indeed, TNF blockade represents one of the most effective treatments for autoimmune diseases such as rheumatoid arthritis while increasing the risk of some infectious diseases. Despite extensive investigation for over four decades, the mechanism by which TNF provides resistance against infection remains unclear and has been attributed to many effects, such as enhancing expression of chemokines, cytokines, adhesion molecules, bactericidal factors, inflammatory cell trafficking and activating phagocyte oxidase (Sedgwick et al., 2000; Strieter et al., 2002). The breadth of TNF’s effects during infection has made it difficult to identify the key cellular and molecular pathways that facilitate pathogen inactivation and clearance. Herein we demonstrate that during pulmonary infection with K. pneumonia, recruitment of IL17-producing ILCs to the infected lung depends on TNF production by recruited CCR2+ IMs.

TNF activates a variety of effector cells by binding to TNFR1 and promoting inflammatory responses. Although previous studies demonstrated that TNF deficient animals are more susceptible to infection with a rodent adapted strain of K. pneumoniae (Laichalk et al., 1996), the TNF-producing cells in the lung and the induced effector functions have remained unclear. Our study identified monocytes as the principal TNF producers and recruitment and activation of IL-17 producing ILCs as the major downstream impact of TNF expression. Our results suggest that TNF enhances ILC recruitment to the lung by up-regulating the chemokine CCL20, which has been shown to be expressed by pulmonary epithelial cells and its expression can be boosted by TNF and IL-17 (Conti et al., 2009; Hirota et al., 2007; Kao et al., 2005; Starner et al., 2003; Sugita et al., 2002). A recent paper described the major ILC subsets in the intestine and ILC2s in the lung as tissue resident cells that maintain the population through self-renewal. ILC3s were not detected in the lungs during steady state or after helminth infection (Gasteiger et al., 2015). Our work identifies the ILC3 population in the lung after acute bacterial infection, and characterizes their important roles as the innate source of IL-17. Since the intestine harbors a large number of self-renewing ILC3s whereas the lung has almost undetectable numbers of ILC3s under homeostatic conditions, we speculate that ILC3s can potentially be recruited to the infected pulmonary mucosa in a CCR6-CCL20 dependent manner from other sites. How ILC3s are recruited will require further studies.

ILCs play a crucial role in defense against a wide range of infections at mucosal sites. ILC3s that produce IL-17 and IL-22, in particular, have drawn attention in recent years due to their important role in intestinal homeostasis and disease pathogenesis. In contrast to many innate immune cells, ILCs depend on other cells to respond to infection for recruitment and activation. Indeed, it has been shown that in the intestine, CD103negCD11B+ DCs produce TNF, thereby enhancing IL-17 production by ILCs (Powell et al., 2012), and CX3CR1+ mononuclear phagocytes (MNPs), which develop from CCR2+ monocyte progenitors, produce IL-1b and TL1A to stimulate or enhance ILC secretion of GM-CSF and IL-22, respectively (Longman et al., 2014; Mortha et al., 2014; Seo et al., 2015). GM-CSF, in return, promotes MNP effector functions (Mortha et al., 2014). The situations in the lung have been less well characterized: although Th17 cells have been demonstrated as the major protective IL-17-producing cell type in the memory response against K. pneumoniae (Chen et al., 2011), the cell types that generate IL-17 in the primary response remain unclear. Furthermore, ILC3s in the respiratory system were not identified until recently (Kim et al., 2014). Our findings indicate that in the pulmonary infections, there is a similar innate source of IL-17 produced by ILC3s, and importantly, there is a similar crosstalk and feedback loop between inflammatory monocytes and ILC3s that modulate ILCs’ functions and in return, enhances the monocytes’ activity.

Kp-MH258 belongs to the Sequence Type 258 (ST258), which as the dominant strain prevailing in the US that is carbapenem-resistant, has likely undergone clonal expansion, and has been disseminated to other countries (Garcia-Fernandez et al., 2012; Kitchel et al., 2009). The clearance of Kp-MH258 requires inflammatory monocytes but not neutrophils, whereas the clearance of other antibiotic-resistant, but carbapenem-sensitive clinical isolates of K. pneumoniae requires both IMs and neutrophils (Xiong et al., 2015). Our results indicate K. pneumoniae strains differ in their sensitivity to innate immune defenses mediated by monocytes and neutrophils. Although strains of K. pneumoniae share the same genus and species names, genetic analysis indicate substantial diversity that extends beyond antibiotic sensitivity and capsule types. Given our findings, it is likely that, some of the genetic diversity between strains extends to defense mechanism that tangles with neutrophil versus monocyte-mediated bactericidal mechanisms. Future studies will likely correlate specific genetic loci within K. pneumoniae that lead to resistance against neutrophil-mediated killing.

IL-17A can bind to the IL-17RA/RC heterodimeric receptor complex (Hu et al., 2011; Kuestner et al., 2007), which is expressed on many cell types (Kolls and Linden, 2004; Yao et al., 1995). While one of the major functions of IL-17 is facilitating neutrophil recruitment (Kolls and Linden, 2004; Ye et al., 2001), its impact on IM recruitment and activation has been less clear. Previous studies demonstrated IL-17R expression on mouse and human monocytes, and that IL-17 signaling can regulate their differentiation and migration (Ge et al., 2014; Shahrara et al., 2009). We found high IL-17R expression by lung IMs, and demonstrate that IL-17 signaling enhances bacterial phagocytosis and killing by IMs. This finding provides new insights into the functional plasticity of inflammatory monocytes as both modulator and effector cells. Among the cytokines produced by RORγt-expressing ILCs, IL-22 has been implicated in defense against the rodent-adapted K. pneumoniae strain and Streptococcus pneumoniae, and is suggested to be produced by NK cells and ILC3s (Van Maele et al., 2014; Xu et al., 2014). In contrast to these studies, clearance of the antibiotic-resistant clinical K. pneumoniae strain did not depend on IL-22 production. Given the differences in the mechanisms of cellular clearance of distinct K. pneumoniae strains, it is likely that the cytokines that are essential for optimal bacterial clearance will also differ.

K. pneumoniae has received much attention for its high-level antibiotic resistance, but it is also notable for the wide range of infections it can cause. In recent years, specific strains of K. pneumoniae have been discovered to cause systemic infection with liver abscess formation (Chuang et al., 2006; Fung et al., 2002). While K. pneumoniae also causes infections of the urinary tract, bloodstream and lungs, the associations between specific strains and infections in distinct anatomic sites are incompletely defined. Ultimately, correlating K. pneumoniae genotypes with the type of innate immune clearance mechanisms in different anatomic sites will identify factors that contribute to pathogenesis but also approaches to enhance resistance that bypass the need for additional antibiotic therapy.



C57BL/6 wild type (WT), IL17-GFP, Rag2−/− and Rag2/Common gamma chain Double Knockout (Rag2/cγc−/−) mice were purchased from the Jackson Laboratory. The generation of CCR2-DTR and CCR2-GFP mice were previously described (Hohl et al., 2009). For depletion experiments, CCR2-DTR mice were injected i.p. with 25 ng/g body weight diphtheria toxin (DT) every other day starting from 5 days before infection. The administration of αLy6G or αGr1 to WT animals were previously described (Xiong et al., 2015). All mice were bred and maintained under specific pathogen-free conditions at the Memorial Sloan Kettering Research Animal Resource Center. Sex-, age- and weight- matched controls were used in all experiments according to institutional guidelines for animal care. All animal procedures were approved by the Institutional Animal Care and Use Committee of the Memorial Sloan-Kettering Cancer Center. Endpoints were met when animals were moribund or lost a significant percentage (>20%) of bodyweight.

For antibody-mediated depletion, antibodies were purchased from BioXCell. In order to achieve a complete depletion in the lung, anti-IL-17A antibody (Clone: 17F3; catalog#: BE0173) was administered to mice through i.p. injection (400µg), i.v. injection (200µg) and intratracheal instillation (200µg) on a daily base starting from one day before infection. Anti-CD90 antibody (Clone: T24/31. Catalog#: BE0212) was delivered through i.p. injection (250µg), i.v. injection (100µg) and intratracheal instillation (100µg) on day −4, −2, 0, 2 and 4 of infection. Anti-TNF antibody (Clone: TN3–19.12. Catalog#: BE0091) was administered through i.p. injection (400µg) and intratracheal instillation (100µg) on a daily base starting from day −1 of infection.

Regarding the bone marrow chimera, recipient animals received 1000 r whole body lethal irradiation and were grafted with 5×106 T cell-depleted bone marrow cells on the next day. Mice were maintained on antibiotic water for 2 weeks after engraftment.

Bacterial strains

Clinical K. pneumoniae isolate Kp-MH258 that belongs to sequencing type 258 (ST258) was used in this study. The bacteria were grown in LB media to log phase. O.D.600 was measured and the cultures were diluted accordingly in PBS to 1 × 104 CFU/ µl for intratracheal instillation.

Intratracheal instillation

Animals were anesthetized by Ketamine through i.p. injection. 40µl of bacterial inoculum, IM suspension or antibody was applied to the trachea through the mouth via a curved, blunt-end needle. Mice were held vertically for 1 min after the inoculation.

Tissue homogenization

The whole lung was perfused before harvest, homogenized and digested with 5% FCS, 5000U/ml collagenase type IV (Worthington), and 20 U/ml DNase (Roche) and were incubated at 37°C for 40 min to obtain single cell suspension for staining or plating.

Flow cytometry

Single cell suspensions were stained and analyzed on a BD LSR II cytometer. Antibodies were purchased from BD bioscience unless otherwise listed. For monocyte, neutrophil, DC, macrophage and T cell stainings, the following antibodies were used: anti-Ly6C (clone AL-21), Ly6G (1A8), CD11b (M1/70), CD45 (30F-11), CD11c (HL3), CD103 (2E7), NK1.1(PK136), CD49b(PanNK)(DX5), CD3 (145-2C11), CD4 (L3T4), γδ TCR (GL3), TNF (MP6-XT22), IL-17RA (PAJ-17R). For ILC staining, the following antibodies were used: Lineage 1(Lin1, in FITC): CD3, CD4, CD8 (53–6.7), CD11c; Lineage2 (Lin2, in PE-CY7): CD11b, NK1.1, CD19 (1D3), Ter119 (TER119). CD45, CD45.1 (A20), CD45.2 (104), CD90 (53–2.1), CD127 (A7R34) were used for the subsequent positive gating.

For intracellular TNF staining, BD Cytofix/Cytoperm Plus Kit (with BD GolgiPlug) (cat# 555028) with Protein Transport Inhibitor (Containing Brefeldin A) (cat#555029) was used. Brefeldin A was added in the digestion and staining buffers. For intracellular IL-17A and IL-22 staining, the antibodies were purchased from ebioscience (Clone: eBio17B7, cat# 17–7177-81 for anti IL-17 and Clone: IL22JOP cat# 17–7222-82 for anti IL-22). Intracellular Fixation & Permeabilization Buffer (plus Brefeldin A) (cat# 88–8823-88) was used. Cells were stimulated in vitro with ionomycin (750ng/ml), PMA (50ng/ml) and IL-23 (40ng/ml) for 3 hours before IL-17A/IL-22 staining.

BrdU staining

BrdU Flow Kit from BD (#559619) was used. In brief, 2.5 mg BrdU solution was injected i.p. into mice 30 minutes before infection. Animals were sacrificed on day 1 following infection and cells were permeabilized, fixed and stained according to the Kit protocol.

MACs beads sorting

Anti Ly6G Microbead Kit (130-092–332) and anti CD11b beads (130-049–601) were purchased from Miltenyi Biotec. LS column was used for the purification of the bead-bound cells. First, anti-Ly6G kit was used to pull down neutrophils, then, anti-CD11b beads were applied to the Ly6G fraction to purify IMs. Isolated cells were stained and analyzed using flow cytometry to check the purity.

Adoptive inflammatory monocyte transfer

Inflammatory monocytes were harvested from the bone marrow of CCR2-GFP mice. The cells were stained with PE antibodies as shown in Fig. S2A. PE GFP+ cells were sorted by the Memorial Sloan-Kettering Cancer Center flow cytometry core facility. The cells were re-suspended in PBS and injected intravenously to the recipient mice. This strategy did not require antibody staining for trafficking molecules (CCR2 and CD11b) and eliminated NK cells, T cells, dendritic cell progenitors and hematopoietic stem cells, yielding a highly purified (>99%) population of IMs.

Bacterial killing assay

IMs were purified from the bone marrow of CCR2-GFP mice as Fig. S2 shows. 1×105 IMs were mixed with 100 CFUs of K. pneumoniae using antibiotic-free cRPMI medium in a 96-well plate. PBS or 100ng/ml recombinant IL-17 (rmIL-17) was added to the culture. The plate was briefly centrifuged and incubated at 37°C, and supernatant was collected and plated at designated time points for bacterial counts. CellROX staining kit (Invitrogen) was added to the culture to measure the ROS production by IMs.

Real-time PCR

Cells were collected in Trizol. Chloroform and 2-propanol were used to extract and precipitate RNA, respectively. Reverse transcript was performed using QuantiTect Reverse Transcription Kit (QIAGEN). Primers were ordered from Taqman website. The real time PCR reaction was run with Thermo Scientific DyNAmo Flash Probe qPCR Kit for the amplification and quantification of cDNA. The expression level was normalized to that of β-actin.

Statistical analysis

All data are analyzed using Graphpad Prism software and are presented as the arithmetic mean ± SEM. Statistical validation was done with the Student’s t test. *p < 0.05, **p < 0.01, ***p < 0.001. p < 0.05 was considered statistically significant.

Supplementary Material




We would like to thank members of the Pamer laboratory for helpful discussions. We thank the Memorial Sloan Kettering Flow Cytometry and Cell Sorting Core at Memorial Sloan Kettering Cancer Center (Core grant P30 CA008748) for their services and advice. This study was supported by grants to E.G.P. from the NIH (R37AI039031) and P01 (A023766)


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


H.X. and E.G.P. conceptualized, designed the research and prepared manuscript; H.X., J.W.K. and D.W.S performed experiments; H.X. and E.G.P. analyzed and interpreted data. R.A.C. and I.M.L. contributed to experimental design and data analysis.


  • Abt MC, Lewis BB, Caballero S, Xiong H, Carter RA, Susac B, Ling L, Leiner I, Pamer EG. Innate Immune Defenses Mediated by Two ILC Subsets Are Critical for Protection against Acute Clostridium difficile Infection. Cell Host Microbe. 2015;18:27–37. [PMC free article] [PubMed]
  • Artis D, Spits H. The biology of innate lymphoid cells. Nature. 2015;517:293–301. [PubMed]
  • Biswas SK, Mantovani A. Macrophage plasticity and interaction with lymphocyte subsets: cancer as a paradigm. Nat Immunol. 2010;11:889–896. [PubMed]
  • Cai S, Batra S, Shen L, Wakamatsu N, Jeyaseelan S. Both TRIF- and MyD88-dependent signaling contribute to host defense against pulmonary Klebsiella infection. J Immunol. 2009;183:6629–6638. [PMC free article] [PubMed]
  • Chen K, McAleer JP, Lin Y, Paterson DL, Zheng M, Alcorn JF, Weaver CT, Kolls JK. Th17 cells mediate clade-specific, serotype-independent mucosal immunity. Immunity. 2011;35:997–1009. [PMC free article] [PubMed]
  • Chuang YP, Fang CT, Lai SY, Chang SC, Wang JT. Genetic determinants of capsular serotype K1 of Klebsiella pneumoniae causing primary pyogenic liver abscess. J Infect Dis. 2006;193:645–654. [PubMed]
  • Conti HR, Shen F, Nayyar N, Stocum E, Sun JN, Lindemann MJ, Ho AW, Hai JH, Yu JJ, Jung JW, et al. Th17 cells and IL-17 receptor signaling are essential for mucosal host defense against oral candidiasis. J Exp Med. 2009;206:299–311. [PMC free article] [PubMed]
  • Fung CP, Chang FY, Lee SC, Hu BS, Kuo BI, Liu CY, Ho M, Siu LK. A global emerging disease of Klebsiella pneumoniae liver abscess: is serotype K1 an important factor for complicated endophthalmitis? Gut. 2002;50:420–424. [PMC free article] [PubMed]
  • Garcia-Fernandez A, Villa L, Carta C, Venditti C, Giordano A, Venditti M, Mancini C, Carattoli A. Klebsiella pneumoniae ST258 producing KPC-3 identified in italy carries novel plasmids and OmpK36/OmpK35 porin variants. Antimicrobial agents and chemotherapy. 2012;56:2143–2145. [PMC free article] [PubMed]
  • Gasteiger G, Fan X, Dikiy S, Lee SY, Rudensky AY. Tissue residency of innate lymphoid cells in lymphoid and nonlymphoid organs. Science. 2015;350:981–985. [PMC free article] [PubMed]
  • Ge S, Hertel B, Susnik N, Rong S, Dittrich AM, Schmitt R, Haller H, von Vietinghoff S. Interleukin 17 receptor A modulates monocyte subsets and macrophage generation in vivo. PLoS One. 2014;9:e85461. [PMC free article] [PubMed]
  • Gordon S, Taylor PR. Monocyte and macrophage heterogeneity. Nat Rev Immunol. 2005;5:953–964. [PubMed]
  • Harrington LE, Hatton RD, Mangan PR, Turner H, Murphy TL, Murphy KM, Weaver CT. Interleukin 17-producing CD4+ effector T cells develop via a lineage distinct from the T helper type 1 and 2 lineages. Nat Immunol. 2005;6:1123–1132. [PubMed]
  • Hirota K, Yoshitomi H, Hashimoto M, Maeda S, Teradaira S, Sugimoto N, Yamaguchi T, Nomura T, Ito H, Nakamura T, et al. Preferential recruitment of CCR6-expressing Th17 cells to inflamed joints via CCL20 in rheumatoid arthritis and its animal model. J Exp Med. 2007;204:2803–2812. [PMC free article] [PubMed]
  • Hohl TM, Rivera A, Lipuma L, Gallegos A, Shi C, Mack M, Pamer EG. Inflammatory monocytes facilitate adaptive CD4 T cell responses during respiratory fungal infection. Cell Host Microbe. 2009;6:470–481. [PMC free article] [PubMed]
  • Hu Y, Shen F, Crellin NK, Ouyang W. The IL-17 pathway as a major therapeutic target in autoimmune diseases. Ann N Y Acad Sci. 2011;1217:60–76. [PubMed]
  • Ivanov II, Atarashi K, Manel N, Brodie EL, Shima T, Karaoz U, Wei D, Goldfarb KC, Santee CA, Lynch SV, et al. Induction of intestinal Th17 cells by segmented filamentous bacteria. Cell. 2009;139:485–498. [PMC free article] [PubMed]
  • Kao CY, Huang F, Chen Y, Thai P, Wachi S, Kim C, Tam L, Wu R. Up-regulation of CC chemokine ligand 20 expression in human airway epithelium by IL-17 through a JAK-independent but MEK/NF-kappaB-dependent signaling pathway. J Immunol. 2005;175:6676–6685. [PubMed]
  • Kim HY, Lee HJ, Chang YJ, Pichavant M, Shore SA, Fitzgerald KA, Iwakura Y, Israel E, Bolger K, Faul J, et al. Interleukin-17-producing innate lymphoid cells and the NLRP3 inflammasome facilitate obesity-associated airway hyperreactivity. Nat Med. 2014;20:54–61. [PMC free article] [PubMed]
  • Kitchel B, Rasheed JK, Patel JB, Srinivasan A, Navon-Venezia S, Carmeli Y, Brolund A, Giske CG. Molecular epidemiology of KPC-producing Klebsiella pneumoniae isolates in the United States: clonal expansion of multilocus sequence type 258. Antimicrobial agents and chemotherapy. 2009;53:3365–3370. [PMC free article] [PubMed]
  • Kolls JK, Linden A. Interleukin-17 family members and inflammation. Immunity. 2004;21:467–476. [PubMed]
  • Kuestner RE, Taft DW, Haran A, Brandt CS, Brender T, Lum K, Harder B, Okada S, Ostrander CD, Kreindler JL, et al. Identification of the IL-17 receptor related molecule IL-17RC as the receptor for IL-17F. J Immunol. 2007;179:5462–5473. [PMC free article] [PubMed]
  • Laichalk LL, Kunkel SL, Strieter RM, Danforth JM, Bailie MB, Standiford TJ. Tumor necrosis factor mediates lung antibacterial host defense in murine Klebsiella pneumonia. Infect Immun. 1996;64:5211–5218. [PMC free article] [PubMed]
  • Liang SC, Tan XY, Luxenberg DP, Karim R, Dunussi-Joannopoulos K, Collins M, Fouser LA. Interleukin (IL)-22 and IL-17 are coexpressed by Th17 cells and cooperatively enhance expression of antimicrobial peptides. J Exp Med. 2006;203:2271–2279. [PMC free article] [PubMed]
  • Longman RS, Diehl GE, Victorio DA, Huh JR, Galan C, Miraldi ER, Swaminath A, Bonneau R, Scherl EJ, Littman DR. CX(3)CR1(+) mononuclear phagocytes support colitis-associated innate lymphoid cell production of IL-22. J Exp Med. 2014;211:1571–1583. [PMC free article] [PubMed]
  • Moore TA, Lau HY, Cogen AL, Standiford TJ. Defective innate antibacterial host responses during murine Klebsiella pneumoniae bacteremia: tumor necrosis factor (TNF) receptor 1 deficiency versus therapy with anti-TNF-alpha. Clin Infect Dis. 2005;3(41 Suppl):S213–S217. [PubMed]
  • Mortha A, Chudnovskiy A, Hashimoto D, Bogunovic M, Spencer SP, Belkaid Y, Merad M. Microbiota-dependent crosstalk between macrophages and ILC3 promotes intestinal homeostasis. Science. 2014;343:1249288. [PMC free article] [PubMed]
  • Powell N, Walker AW, Stolarczyk E, Canavan JB, Gokmen MR, Marks E, Jackson I, Hashim A, Curtis MA, Jenner RG, et al. The transcription factor T-bet regulates intestinal inflammation mediated by interleukin-7 receptor+ innate lymphoid cells. Immunity. 2012;37:674–684. [PMC free article] [PubMed]
  • Satoh-Takayama N, Vosshenrich CA, Lesjean-Pottier S, Sawa S, Lochner M, Rattis F, Mention JJ, Thiam K, Cerf-Bensussan N, Mandelboim O, et al. Microbial flora drives interleukin 22 production in intestinal NKp46+ cells that provide innate mucosal immune defense. Immunity. 2008;29:958–970. [PubMed]
  • Sedgwick JD, Riminton DS, Cyster JG, Korner H. Tumor necrosis factor: a master-regulator of leukocyte movement. Immunol Today. 2000;21:110–113. [PubMed]
  • Seo SU, Kuffa P, Kitamoto S, Nagao-Kitamoto H, Rousseau J, Kim YG, Nunez G, Kamada N. Intestinal macrophages arising from CCR2(+) monocytes control pathogen infection by activating innate lymphoid cells. Nat Commun. 2015;6:8010. [PMC free article] [PubMed]
  • Serbina NV, Salazar-Mather TP, Biron CA, Kuziel WA, Pamer EG. TNF/iNOS-producing dendritic cells mediate innate immune defense against bacterial infection. Immunity. 2003;19:59–70. [PubMed]
  • Shahrara S, Pickens SR, Dorfleutner A, Pope RM. IL-17 induces monocyte migration in rheumatoid arthritis. J Immunol. 2009;182:3884–3891. [PMC free article] [PubMed]
  • Sharma A, Steichen AL, Jondle CN, Mishra BB, Sharma J. Protective role of Mincle in bacterial pneumonia by regulation of neutrophil mediated phagocytosis and extracellular trap formation. J Infect Dis. 2014;209:1837–1846. [PMC free article] [PubMed]
  • Shi C, Jia T, Mendez-Ferrer S, Hohl TM, Serbina NV, Lipuma L, Leiner I, Li MO, Frenette PS, Pamer EG. Bone marrow mesenchymal stem and progenitor cells induce monocyte emigration in response to circulating toll-like receptor ligands. Immunity. 2011;34:590–601. [PMC free article] [PubMed]
  • Sonnenberg GF, Monticelli LA, Elloso MM, Fouser LA, Artis D. CD4(+) lymphoid tissue-inducer cells promote innate immunity in the gut. Immunity. 2011;34:122–134. [PMC free article] [PubMed]
  • Starner TD, Barker CK, Jia HP, Kang Y, McCray PB., Jr CCL20 is an inducible product of human airway epithelia with innate immune properties. Am J Respir Cell Mol Biol. 2003;29:627–633. [PubMed]
  • Steichen AL, Binstock BJ, Mishra BB, Sharma J. C-type lectin receptor Clec4d plays a protective role in resolution of Gram-negative pneumonia. J Leukoc Biol. 2013;94:393–398. [PubMed]
  • Strieter RM, Belperio JA, Keane MP. Cytokines in innate host defense in the lung. J Clin Invest. 2002;109:699–705. [PMC free article] [PubMed]
  • Sugita S, Kohno T, Yamamoto K, Imaizumi Y, Nakajima H, Ishimaru T, Matsuyama T. Induction of macrophage-inflammatory protein-3alpha gene expression by TNF-dependent NF-kappaB activation. J Immunol. 2002;168:5621–5628. [PubMed]
  • Sutton CE, Lalor SJ, Sweeney CM, Brereton CF, Lavelle EC, Mills KH. Interleukin-1 and IL-23 induce innate IL-17 production from gammadelta T cells, amplifying Th17 responses and autoimmunity. Immunity. 2009;31:331–341. [PubMed]
  • Van Maele L, Carnoy C, Cayet D, Ivanov S, Porte R, Deruy E, Chabalgoity JA, Renauld JC, Eberl G, Benecke AG, et al. Activation of Type 3 Innate Lymphoid Cells and Interleukin 22 Secretion in the Lungs During Streptococcus pneumoniae Infection. J Infect Dis. 2014 [PubMed]
  • Xiong H, Carter RA, Leiner IM, Tang YW, Chen L, Kreiswirth BN, Pamer EG. Distinct contributions of neutrophils and CCR2+ monocytes to pulmonary clearance of different Klebsiella pneumoniae strains. Infect Immun. 2015 [PMC free article] [PubMed]
  • Xu X, Weiss ID, Zhang HH, Singh SP, Wynn TA, Wilson MS, Farber JM. Conventional NK cells can produce IL-22 and promote host defense in Klebsiella pneumoniae pneumonia. J Immunol. 2014;192:1778–1786. [PMC free article] [PubMed]
  • Yao Z, Fanslow WC, Seldin MF, Rousseau AM, Painter SL, Comeau MR, Cohen JI, Spriggs MK. Herpesvirus Saimiri encodes a new cytokine, IL-17, which binds to a novel cytokine receptor. Immunity. 1995;3:811–821. [PubMed]
  • Ye P, Rodriguez FH, Kanaly S, Stocking KL, Schurr J, Schwarzenberger P, Oliver P, Huang W, Zhang P, Zhang J, et al. Requirement of interleukin 17 receptor signaling for lung CXC chemokine and granulocyte colony-stimulating factor expression, neutrophil recruitment, and host defense. J Exp Med. 2001;194:519–527. [PMC free article] [PubMed]
  • Zheng Y, Valdez PA, Danilenko DM, Hu Y, Sa SM, Gong Q, Abbas AR, Modrusan Z, Ghilardi N, de Sauvage FJ, et al. Interleukin-22 mediates early host defense against attaching and effacing bacterial pathogens. Nat Med. 2008;14:282–289. [PubMed]