Abstract
Sepsis disproportionately affects the very old and the very young. IL-1 signaling is important in innate host defense but may also play a deleterious role in acute inflammatory conditions (including sepsis) by promulgating life-threatening inflammation. IL-1 signaling is mediated by two distinct ligands: IL-1α and IL-1β, both acting on a common receptor (IL-1R1). IL-1R1 targeting has not reduced adult human sepsis mortality despite biologic plausibility. Because the specific role of IL-1α or IL-1β in sepsis survival is unknown in any age group and the role of IL-1 signaling remains unknown in neonates, we studied the role of IL-1 signaling, including the impact of IL-1α and IL-1β, on neonatal murine sepsis survival. IL-1 signaling augments the late plasma inflammatory response to sepsis. IL-1α and not IL-1β is the critical mediator of sepsis mortality, likely because of paracrine actions within the tissue. These data do not support targeting IL-1 signaling in neonates.
Introduction
Sepsis has its greatest impact at the extremes of age: preterm newborns and elderly adults. The risk of acquiring infection and the progression of infection to sepsis in neonates are attributed to distinct immune function as compared with healthy adults. We and others have shown that in contrast to adults, neonates rely on the innate immune system for host defense (1). The majority of innate immune cells are activated by engagement of IL-1R1 (2). Although the impact of IL-1 signaling has been well characterized in septic adult humans and animals, the role of IL-1 in neonatal sepsis remains obscure.
IL-1 signaling is mediated by two distinct ligands: IL-1α and IL-1β, both acting on a common receptor (IL-1R1) (3). IL-1β is critically important for innate immune cell activation but may also be detrimental if produced in excess (4, 5). In contrast to IL-1β, relatively less is known about IL-1α in neonatal inflammatory pathology. IL-1α is a nearly ubiquitous alarmin that also exists in a proform similar to IL-1β (6). In contrast to the biologically inactive proform of IL-1β, pro–IL-1α demonstrates biological activity similar to mature IL-1α and thus can promulgate the initial inflammatory response when released by injured tissue during sepsis (7). However, the specific contributions of these two IL-1 ligands to polymicrobial sepsis survival have not been reported in adults or neonates. Therefore, we investigated the role of IL-1 signaling mediated by both IL-1α and IL-1β in a murine model of neonatal sepsis. Unexpectedly, we discovered that IL-1β is unnecessary for lethal neonatal sepsis; rather, IL-1α is the predominant IL-1R1–dependent determinant of mortality.
Materials and Methods
Mice
The Institutional Animal Care and Use Committee at the University of Florida approved all studies prior to their initiation. Specific pathogen–free male and female C57BL/6 mice (wild-type [WT] and genetically modified on a B6 background) were purchased from The Jackson Laboratory (Bar Harbor, ME) between 6 and 8 wk of age and allowed at least 7 d to equilibrate before any further use as breeders to generate pups. Pups aged 5–7 d were considered neonates and used for all experiments (8). IL-1β null mice on a B6 background were previously described (8). Mixed-sex litters were used.
Human subjects and sample processing
All studies were institutional review board–approved prior to initiation. A description of the patient cohort including methods for gene expression profiling and plasma protein analyses was previously described (9).
Cecal slurry–induced peritonitis
Mice were made septic using the cecal slurry model (10). Briefly, a 6-wk old, nonpregnant, female WT (C57BL/6) mouse was euthanized <2 wk after arrival from the vendor, and the cecum was isolated. Cecal contents were expressed, weighed, suspended in 5% dextrose at a concentration of 80 mg/ml, and administered via i.p. injection to neonatal mice at the desired lethal dose. Mice were monitored after injection as described (1).
Bacteremia and plasma cytokine analyses
Whole blood was obtained by intracardiac puncture and evaluated for bacterial colonization as described previously (8). Murine blood plasma cytokine concentrations were measured using a magnetic bead–based multiplex assay for 32 analytes [G-CSF, GM-CSF, M-CSF, VEGF, TNF-α, IFN-γ, IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-9, IL-10, IL-12(p40), IL-12(p70), IL-13, IL-15, IL-17A, KC(CXCL1), MIP2(CXCL2), LIX(CXCL5), MIG(CXCL9), IP10(CXCL10), LIF, MCP1(CCL2), MIP1α(CCL3), MIP1β(CCL4), RANTES(CCL5), and Eotaxin(CCL11)] (MilliporeSigma) on a magnetic bead–based platform (Luminex). For all values that were less than or in excess of the endpoints of linearity for the assay, the limits of detection (minimum and maximum values) were substituted to avoid a type I error based on extrapolation, thus enabling the performance of statistical comparisons.
Western blot
Lung and ileum were homogenized in radioimmunoprecipitation assay buffer with a protease and phosphatase inhibitor mixture (Sigma-Aldrich, St. Louis, MO), and total protein was quantified using Bio-Rad DC Protein Assay reagents (Bio-Rad, Hercules, CA) as per manufacturer’s instructions. Equal amounts of extracted protein from each sample were separated on a NuPAGE polyacrylamide gel (Invitrogen, Waltham, MA) and transferred to a nitrocellulose membrane (Invitrogen). After blocking with Odyssey Blocking Buffer (LI-COR, Lincoln, NE), membranes were incubated with primary and secondary Abs. Blots were detected using the Odyssey CLx imaging system (LI-COR).
Immunohistochemistry
Neonatal lung and ileum were harvested and were frozen in embedding medium (Tissue Tek, Sakura Finetek) using isopentane on dry ice. Frozen tissues were sectioned (10 μm), fixed with 10% formalin (09122; Neogen Vet, Lexington, KY), permeabilized with 0.1% Triton X-100 (9410; MilliporeSigma, Burlington, MA), blocked with Background Buster (NB306; Innovex Biosciences, Richmond, CA), and incubated overnight with rabbit anti–IL-1α (ab7632; Abcam, Cambridge, MA) followed by AF555 donkey anti-rabbit IgG (A31572; Life Technologies, Carlsbad, CA). Sections were mounted with ProLong Gold with DAPI (P36935; Invitrogen, Carlsbad, CA) and visualized by confocal microscopy using a Carl Zeiss LSM 710 confocal microscope.
Reagents
Statistics
Survival was compared using the Fisher exact test. Values were considered significant if the two-tailed confidence level was p < 0.05. Depending on whether the descriptive analyses passed normality and equal variance, a Student t test, Mann–Whitney, or Wilcoxon signed-rank test was used to compare results from two groups. Values were considered significant if p < 0.05. Analyses were performed using Prism 7 (GraphPad, La Jolla, CA).
Results
Sepsis survival is enhanced in IL-1R1–deficient neonatal mice
To determine the role of IL-1 signaling in neonatal sepsis, we examined polymicrobial sepsis survival in IL-1R1 null (−/−) mice compared with WT mice (Fig. 1A). Mice lacking IL-1R1 demonstrated a substantial sepsis survival advantage compared with WT mice (63% versus 14%, p < 0.005). IL-1R1 deletion did not significantly modify bacteremia at 6 and 12 h, suggesting that the survival advantage seen in IL-1R1 null compared with WT mice was not a result of enhanced bacterial clearance or antimicrobial activity but rather could result from a decreased systemic inflammatory response (Fig. 1B).
(A) Polymicrobial sepsis survival in IL-1R1 null (−/−) mice compared with WT mice (63% versus 14%). *p < 0.005 by Fisher exact test. (B) Bacteremia at 6 and 12 h among IL-1R1−/− and WT mice (n ≥ 5 for all groups at all time points). Each symbol represents an individual mouse. Survival experiments were independently performed a minimum of twice, with results combined and presented.
IL-1R1 deletion is associated with a significant reduction in the late plasma inflammatory response to sepsis
IL-1 signaling is important for generation and amplification of the inflammatory response to sepsis (11). Therefore, we next examined the effect of IL-1R1 deletion on plasma inflammatory mediator production at 4, 6, and 12 h after neonatal sepsis. The majority of plasma mediators examined were not significantly different in WT versus IL-1R1−/− plasma at early time points (4 and 6 h, Fig. 2) suggesting that IL-1R1 signaling was not required to generate the early plasma inflammatory response. Notable exceptions included increased CXCL2 at 4 h, increased CXCL10 at 6 h, and reduced IL-17A in IL-1R1−/− versus WT mice. In contrast, multiple plasma inflammatory mediators were reduced in IL-1R1−/− mice compared with WT mice at 12 h after sepsis, including TNF-α, IL-1β, IL-6, IL-12p70, IL-13, IL-17A, CCL3, CCL4, CCL11, CXCL1, CXCL2, and CXCL10 (all p < 0.05). Of note, plasma IL-1α concentrations did not differ between IL-1R1−/− and WT mice at any of the time points examined. Together with our studies of sepsis survival and bacteremia in IL-1R1−/− and WT neonatal mice, these data suggest that IL-1 signaling augments neonatal sepsis by sustaining the plasma inflammatory response.
Plasma mediator concentrations (median with quartiles) from septic WT and IL-1R1−/− neonates at 4, 6, and 12 h after cecal slurry. n ≥ 5 for all groups at all time points. Each symbol represents an individual mouse. *p < 0.05.
Plasma IL-1α and IL-1β are increased early and sustained in neonatal sepsis
To understand the kinetics of IL-1 production after polymicrobial sepsis induction in neonatal mice, we measured plasma IL-1α and IL-1β at rest and at 2, 4, 6, 8, and 16 h after sepsis in WT mice (Fig. 3). All plasma samples were drawn prior to initial mortality to avoid selection bias. Plasma IL-1α and IL-1β increased 25–40-fold at 2 h after cecal slurry and remained comparably elevated at all subsequent time points. These data confirm that both IL-1α and IL-1β are not only strongly induced and present in circulation early postinfection but also remain elevated and thereby can amplify other aspects of the host inflammatory response.
Plasma IL-1α and IL-1β from WT neonates at rest and time points shown after sepsis. n ≥ 5 for all groups at all time points. Data represent medians with 95% confidence intervals.
IL-1β does not play a significant role in sepsis survival
To better understand the specific contribution of IL-1β to sepsis mortality, we used genetic (IL-1β−/− and caspase-1/11−/−) and pharmacologic (prophylactic treatment i.p. with 100 μg/mouse of anti–IL-1β immediately prior to slurry injection) targeting of IL-1β. None of these strategies modified sepsis mortality (Figs. 4A, 4B, 5B) as compared with WT mice treated with or without an isotype control Ab. Although excessive IL-1β production can be detrimental to the host, IL-1β can also enhance innate immune function (12, 13). To determine whether the IL-1β response was inadequate in neonates and could be enhanced, concurrent treatment with rIL-1β (10 ng/g i.p.) (12–14) was performed but also did not modify subsequent sepsis survival (26%, n = 34) compared with septic WT mice treated with PBS (28%, n = 36) (Fig. 4C). Taken together, these data do not support a significant role for IL-1β in sepsis survival using this model.
Impact of genetic and pharmacologic interventions that target or enhance IL-1β. (A) IL-1β−/− versus WT. (B) Sepsis survival among WT neonates (33%) versus caspase-1/11−/− (38%). (C) Sepsis survival among WT neonates that received 10 ng/g body weight rIL-1β concurrently with cecal slurry (26%) or PBS (28%). Survival experiments were independently performed a minimum of twice, with results combined and presented.
(A) Sepsis survival in IL-1α−/− mice versus WT mice. (B) Sepsis survival among IL-1α−/− and WT mice pretreated with anti–IL-1β or isotype Ab. (C) Sepsis survival among IL-1β−/− and WT mice pretreated with anti–IL-1β or isotype Ab. Survival experiments were independently performed a minimum of twice, with results combined and presented.
Pro–IL-1α is present in healthy ileum and lung
The failure to attenuate sepsis mortality by genetic and pharmacologic targeting of IL-1β suggests that deleterious IL-1α signaling mediates the IL-1R1–dependent sepsis pathophysiology in this model. Tissue injury and release of alarmins, including IL-1α, are well described with sepsis (15). As expected (6), neither pro–IL-1β nor mature IL-1β protein was present in healthy neonatal lung and ileum (Fig. 6A). In contrast, pro–IL-1α protein was detectable in both lung and ileum from healthy neonatal mice (Fig. 6B, 6C).
(A) Representative immunoblots of murine rIL-1β as well as lung and ileum from healthy neonatal mice. (B) Representative immunofluorescent image (original magnification ×200) of total IL-1α in lung and ileum from healthy neonatal mice. (C) Representative immunoblots for IL-1α in lung and ileum from healthy neonatal mice. Each experiment was independently performed a minimum of twice on a minimum of three mice.
IL-1α plays a significant role in sepsis survival
In contrast to our results targeting IL-1β, sepsis survival was dramatically increased in IL-1α−/− mice as compared with WT mice (52% versus 17%, p < 0.02, Fig. 5A). Prophylactic treatment with anti–IL-1β (100 μg/mouse of anti–IL-1β i.p. immediately prior to slurry injection) to neutralize circulating IL-1β further enhanced sepsis survival in IL-1α−/− mice but had no effect on WT mice over isotype control treatment (Fig. 5B). Although prophylactic treatment of WT and IL-1β null mice with anti–IL-1α (100 μg/mouse i.p. immediately prior to slurry injection) to neutralize circulating IL-1α resulted in a small reduction in sepsis mortality, this effect was not statistically significant and likely reflects the local production and paracrine action of IL-1α, which is difficult to neutralize with systemic treatments (Fig. 5C).
Discussion
There has been a longstanding interest in the deleterious role of IL-1R1 signaling in sepsis (3, 16–20). To our knowledge, this report is the first study in which the roles of both IL-1α and IL-1β are examined in a comparative manner using an established and clinically relevant model of sepsis rather than endotoxemia. We show IL-1 signaling through IL-1R1 is detrimental primarily because it sustains the late plasma inflammatory response. Intriguingly, a deleterious role on neonatal sepsis mortality for IL-1β by itself was not supported by genetic or pharmacologic attenuation approaches in this study, and exogenous IL-1β did not increase sepsis mortality. These studies strongly argue that it is IL-1α, not IL-1β, that mediates the detrimental effects of IL-1R1-signaling on neonatal sepsis survival.
Prior studies have suggested a deleterious role for IL-1β in adults using either knockout mice or Ab-mediated neutralization (21). Importantly, these studies were not done in accepted polymicrobial sepsis models, such as cecal ligation and puncture, but rather focused on LPS to mimic septic shock (21). Although nearly ubiquitous, less emphasis has been placed on the role of IL-1α in sepsis. Relevant to neonates specifically, IL-1α is the top upregulated gene in LPS-stimulated, enriched neutrophils isolated from human neonatal cord blood (22). Further, IL-1α replicates many of the effects of endotoxemia in primates, including hypotension (23).
Attenuation of IL-1R1 signaling via IL-1ra has been well studied but produced only minor reductions in mortality in adult mice after cecal ligation and puncture and in adult humans with sepsis (24, 25). Similarly, we have found that a single prophylactic administration of IL-1ra (100 μg/mouse i.p. immediately prior to slurry injection) did not modify sepsis survival in neonatal WT mice (Supplemental Fig. 1). Because pro–IL-1α is present in tissues at rest, does not require synthesis, and signals rapidly and potently, we also attempted to pharmacologically block IL-1α–IL-1R1 signaling early by administering an anti–IL-1α Ab prior to the slurry injection. However, pharmacologic targeting of IL-1α (prophylaxis immediately prior to slurry injection) was also not as effective as genetic deletion of IL-1α on sepsis survival. We have previously used the i.p. route to successfully administer Abs to murine neonates (26). The anti–IL-1α Ab used in these experiments has been previously used in vivo (27), and the dose used for pharmacological neutralization (100 μg/mouse) was sufficient (13); thus, technical limitations do not readily explain the discrepancy between knockout mice and pharmacologically treated WT mice. Although neutralizing Abs and receptor antagonist treatments can act on circulating ligands and easily accessible receptors in well-vascularized tissues, they may not effectively penetrate all tissues to inhibit local IL-1α–mediated IL-1R1 signaling. Thus, it is likely that paracrine (tissue-level) signaling mediated by IL-1α (pro- or mature) released early as a result of tissue necrosis secondary to shifts in perfusion may be the critical determinant of sepsis mortality (28). Local IL-1α release and rapid IL-1R1 signaling following tissue injury with sepsis may also explain why IL-1ra treatment has failed to reduce mortality in humans.
Our findings in mice point to a prominent deleterious role for IL-1α–mediated IL-1R1 signaling. Because IL-1R2 reduces IL-1α–induced inflammation that occurs with necrosis (29), and IL-1α was the most upregulated gene in LPS-stimulated human neonatal neutrophils (22), we re-examined whole blood gene expression profiles from septic human neonates, with specific attention to IL-1R2 and IL-1–related gene expression as well as IL-1–related plasma concentrations (9). Compared with uninfected neonates, IL-1R2 was among the top five upregulated genes in the blood from septic neonates (8.2-fold over uninfected, Supplemental Fig. 2). Similarly, sIL-1RII was increased by 50% in the plasma of septic neonates over uninfected neonates. In contrast, plasma IL-1α and IL-1β were not different when compared with uninfected neonates. Strong negative regulation of IL-1 signaling in the blood in the absence of robust increases in plasma IL-1 proteins in these human neonates strongly suggests IL-1 signaling may occur in tissue compartments, as we saw in our murine neonatal studies.
The identification of early paracrine IL-1α–mediated, IL-1R1–dependent signaling that leads to sepsis mortality but is not amenable to IL-1–targeting therapy underscores the need to pursue targeted sepsis therapeutics downstream of IL-1, especially in the setting of significant and evolving tissue injury. We recently described a novel IL-18/IL-1/IL-17a axis that foments sepsis and demonstrated a significant reduction in sepsis mortality with transient IL-17R blockade (8). Similar benefits in sepsis survival were provided by IL-17R blockade following cecal ligation and puncture in adult mice (30), and a deleterious role for IL-17A also occurs in necrotizing enterocolitis (31). These data provide a strong rationale for the selective use of IL-17A or IL-17R mAbs in septic patients with the highest risk of mortality that renews hope for effective interventions in this vulnerable population.
In conclusion, we show IL-1α, not IL-1β, is the IL-1R1 ligand portending mortality in sepsis. Because the biologically active proform of IL-1α is released locally within tissue compartments and acts immediately with the tissue injury inherent to sepsis, IL-1R1 blocking or IL-1α–specific attenuation strategies we employed were not as effective as results seen with IL-1R1 or IL-1α null mice. Further investigation into targets downstream of IL-1R1 signaling, such as IL-17A, may hold more promise to attenuate pathologic inflammation in the setting of sepsis.
Disclosures
The authors have no financial conflicts of interest.
Acknowledgments
We thank Prof. Linc Moldawer for comments on the manuscript.
Footnotes
This work supported by the National Institutes of Health National Heart, Lung, and Blood Institute (K08HL133484 to J.T.B.), the National Institute of Allergy and Infectious Diseases (R21AI121549 to D.J.M.), the National Institute of General Medical Sciences (R01GM128452 to J.L.W.), and the Eunice Kennedy Shriver National Institute of Child Health and Human Development (R01HD089939 to J.L.W.) as well as the JDRF Career Development Award (to D.J.M.).
The online version of this article contains supplemental material.
Abbreviations used in this article:
- IL-1ra
- IL-1R1 antagonist
- WT
- wild-type.
- Received August 10, 2018.
- Accepted September 12, 2018.
- Copyright © 2018 by The American Association of Immunologists, Inc.