Acute Respiratory Distress Syndrome (ARDS) is a severe lung injury with high mortality, primarily characterized by bilateral pulmonary opacities on chest radiographs and hypoxemia. In this work, we trained a convolutional neural network (CNN) model that can reliably identify bilateral opacities on routine chest X-ray images of critically ill patients. We propose this model as a tool to generate predictive alerts for possible ARDS cases, enabling early diagnosis. Our team created a unique dataset of 7800 single-view chest-X-ray images labeled for the presence of bilateral or unilateral pulmonary opacities, or ‘equivocal’ images, by three blinded clinicians. We used a novel training technique that enables the CNN to explicitly predict the ‘equivocal’ class using an uncertainty-aware label smoothing loss. We achieved an Area under the Receiver Operating Characteristic Curve (AUROC) of 0.82 (95% CI: 0.80, 0.85), a precision of 0.75 (95% CI: 0.73, 0.78), and a sensitivity of 0.76 (95% CI: 0.73, 0.78) on the internal test set while achieving an (AUROC) of 0.84 (95% CI: 0.81, 0.86), a precision of 0.73 (95% CI: 0.63, 0.69), and a sensitivity of 0.73 (95% CI: 0.70, 0.75) on an external validation set. Further, our results show that this approach improves the model calibration and diagnostic odds ratio of the hypothesized alert tool, making it ideal for clinical decision support systems.
by
Enjae Jung;
Erin E. Perrone;
Pavan Brahmamdan;
Jacquelyn S. McDonough;
Ann M. Leathersich;
Jessica A. Dominguez;
Andrew T. Clark;
Amy C. Fox;
W. Michael Dunne;
Richard S. Hotchkiss;
Craig Coopersmith
World conditions place large populations at risk from ionizing radiation (IR) from detonation of dirty bombs or nuclear devices. In a subgroup of patients, ionizing radiation exposure would be followed by a secondary infection. The effects of radiation combined injury are potentially more lethal than either insult in isolation. The purpose of this study was to determine mechanisms of mortality and possible therapeutic targets in radiation combined injury. Mice were exposed to IR with 2.5 Gray (Gy) followed four days later by intratracheal methicillin-resistant Staphylococcus aureus (MRSA). While either IR or MRSA alone yielded 100% survival, animals with radiation combined injury had 53% survival (p = 0.01). Compared to IR or MRSA alone, mice with radiation combined injury had increased gut apoptosis, local and systemic bacterial burden, decreased splenic CD4 T cells, CD8 T cells, B cells, NK cells, and dendritic cells, and increased BAL and systemic IL-6 and G-CSF. In contrast, radiation combined injury did not alter lymphocyte apoptosis, pulmonary injury, or intestinal proliferation compared to IR or MRSA alone. In light of the synergistic increase in gut apoptosis following radiation combined injury, transgenic mice that overexpress Bcl-2 in their intestine and wild type mice were subjected to IR followed by MRSA. Bcl-2 mice had decreased gut apoptosis and improved survival compared to WT mice (92% vs. 42%; p<0.01). These data demonstrate that radiation combined injury results in significantly higher mortality than could be predicted based upon either IR or MRSA infection alone, and that preventing gut apoptosis may be a potential therapeutic target.
BACKGROUND: Mortality is significantly higher in septic patients with cancer than in septic patients without a history of cancer. We have previously described a model of pancreatic cancer followed by sepsis from Pseudomonas aeruginosa pneumonia in which cancer septic mice have higher mortality than previously healthy septic mice, associated with increased gut epithelial apoptosis and decreased T cell apoptosis. The purpose of this study was to determine whether this represents a common host response by creating a new model in which both the type of cancer and the model of sepsis are altered.
METHODS: C57Bl/6 mice received an injection of 250,000 cells of the lung cancer line LLC-1 into their right thigh and were followed three weeks for development of palpable tumors. Mice with cancer and mice without cancer were then subjected to cecal ligation and puncture and sacrificed 24 hours after the onset of sepsis or followed 7 days for survival.
RESULTS: Cancer septic mice had a higher mortality than previously healthy septic mice (60% vs. 18%, p = 0.003). Cancer septic mice had decreased number and frequency of splenic CD4+ lymphocytes secondary to increased apoptosis without changes in splenic CD8+ numbers. Intestinal proliferation was also decreased in cancer septic mice. Cancer septic mice had a higher bacterial burden in the peritoneal cavity, but this was not associated with alterations in local cytokine, neutrophil or dendritic cell responses. Cancer septic mice had biochemical evidence of worsened renal function, but there was no histologic evidence of renal injury.
CONCLUSIONS: Animals with cancer have a significantly higher mortality than previously healthy animals following sepsis. The potential mechanisms associated with this elevated mortality differ significantly based upon the model of cancer and sepsis utilized. While lymphocyte apoptosis and intestinal integrity are both altered by the combination of cancer and sepsis, the patterns of these alterations vary greatly depending on the models used.
Recent seminal studies have revealed that laboratory mice differ from adult humans with regard to the frequency, number, and distribution of memory T cells. Because our data show that memory T cells are more susceptible to sepsis-induced death than naive T cells, in this study we developed a model in which mice possess a memory T cell compartment more similar to that of adult humans, to better study immune responses during sepsis in the more physiologically relevant context of high frequencies of memory T cells. Using this model, we found that CD44hi memory T cells significantly upregulated the coinhibitory molecule 2B4 during sepsis, and 2B4+ memory T cells coexpressed markers of both activation and exhaustion. Genetic deficiency in 2B4 resulted in decreased mortality during sepsis. Mechanistically, this decreased mortality was associated with reduced caspase-3/7+ apoptotic T cells in 2B4-/- relative to WT, septic hosts. These results were corroborated by analysis of PBMCs isolated from human patients with sepsis, which showed increased frequencies of caspase-3/7+ apoptotic cells among 2B4+ relative to 2B4- T cells. Thus, 2B4 plays a critical role in sepsis-induced apoptosis in both murine memory T cells and those isolated from human patients with sepsis.
Background
Mice with conditional, intestine-specific deletion of microsomal triglyceride transfer protein (Mttp-IKO) exhibit a complete block in chylomicron assembly together with lipid malabsorption. Young (8–10 week) Mttp-IKO mice have improved survival when subjected to a murine model of Pseudomonas aeruginosa-induced sepsis. However, 80% of deaths in sepsis occur in patients over age 65. The purpose of this study was to determine whether age impacts outcome in Mttp-IKO mice subjected to sepsis.
Methods
Aged (20–24 months) Mttp-IKO mice and WT mice underwent intratracheal injection with P. aeruginosa. Mice were either sacrificed 24 hours post-operatively for mechanistic studies or followed seven days for survival.
Results
In contrast to young septic Mttp-IKO mice, aged septic Mttp-IKO mice had a significantly higher mortality than aged septic WT mice (80% vs. 39%, p = 0.005). Aged septic Mttp-IKO mice exhibited increased gut epithelial apoptosis, increased jejunal Bax/Bcl-2 and Bax/Bcl-XL ratios yet simultaneously demonstrated increased crypt proliferation and villus length. Aged septic Mttp-IKO mice also manifested increased pulmonary myeloperoxidase levels, suggesting increased neutrophil infiltration, as well as decreased systemic TNFα compared to aged septic WT mice.
Conclusions
Blocking intestinal chylomicron secretion alters mortality following sepsis in an age-dependent manner. Increases in gut apoptosis and pulmonary neutrophil infiltration, and decreased systemic TNFα represent potential mechanisms for why intestine-specific Mttp deletion is beneficial in young septic mice but harmful in aged mice as each of these parameters are altered differently in young and aged septic WT and Mttp-IKO mice.
While much of cancer immunology research has focused on anti-tumor immunity both systemically and within the tumor microenvironment, little is known about the impact of pre-existing malignancy on pathogen-specific immune responses. Here, we sought to characterize the antigen-specific CD8+ T cell response following a bacterial infection in the setting of pre-existing pancreatic adenocarcinoma. Mice with established subcutaneous pancreatic adenocarcinomas were infected with Listeria monocytogenes, and antigen-specific CD8+ T cell responses were compared to those in control mice without cancer. While the kinetics and magnitude of antigen-specific CD8+ T cell expansion and accumulation was comparable between the cancer and non-cancer groups, bacterial antigen-specific CD8+ T cells and total CD4+ and CD8+ T cells in cancer mice exhibited increased expression of the coinhibitory receptors BTLA, PD-1, and 2B4. Furthermore, increased inhibitory receptor expression was associated with reduced IFN-γ and increased IL-2 production by bacterial antigen-specific CD8+ T cells in the cancer group. Taken together, these data suggest that cancer's immune suppressive effects are not limited to the tumor microenvironment, but that pre-existing malignancy induces phenotypic exhaustion in T cells by increasing expression of coinhibitory receptors and may impair pathogen-specific CD8+ T cell functionality and differentiation.
Background
Patients admitted to the intensive care unit with alcohol use disorders have increased morbidity and mortality. The purpose of this study was to determine how chronic alcohol ingestion alters the host response to sepsis in mice.
Methods
Mice were randomized to receive either alcohol or water for 12 weeks and then subjected to cecal ligation and puncture. Mice were sacrificed 24 hours post-operatively or followed seven days for survival.
Results
Septic alcohol-fed mice had a significantly higher mortality than septic water-fed mice (74% vs. 41%, p = 0.01). This was associated with worsened gut integrity in alcohol-fed mice with elevated intestinal epithelial apoptosis, decreased crypt proliferation and shortened villus length. Further, alcohol-fed mice had higher intestinal permeability with decreased ZO-1 and occludin protein expression in the intestinal tight junction. The frequency of splenic and bone marrow CD4+ T cells was similar between groups; however, splenic CD4+ T cells in septic alcohol-fed mice had a marked increase in both TNF and IFN-γ production following ex vivo stimulation. Neither the frequency nor function of CD8+ T cells differed between alcohol-fed and water-fed septic mice. NK cells were decreased in both the spleen and bone marrow of alcohol-fed septic mice. Pulmonary myeloperoxidase levels and BAL levels of G-CSF and TFG-β were higher in alcohol-fed mice. Pancreatic metabolomics demonstrated increased acetate, adenosine, xanthine, acetoacetate, 3-hydroxybutyrate and betaine in alcohol-fed mice and decreased cytidine, uracil, fumarate, creatine phosphate, creatine, and choline. Serum and peritoneal cytokines were generally similar between alcohol-fed and water-fed mice, and there were no differences in bacteremia, lung wet to dry weight, or pulmonary, liver or splenic histology.
Conclusions
When subjected to the same septic insult, mice with chronic alcohol ingestion have increased mortality. Alterations in intestinal integrity, the host immune response, and pancreatic metabolomics may help explain this differential response.
Sepsis is a leading cause of morbidity and mortality associated with significant impairment in memory T cells. These changes include the upregulation of co-inhibitory markers, a decrease in functionality, and an increase in apoptosis. Due to recent studies describing IL-27 regulation of TIGIT and PD-1, we assessed whether IL-27 impacts these co-inhibitory molecules in sepsis. Based on these data, we hypothesized that IL-27 was responsible for T cell dysfunction during sepsis. Using the cecal ligation and puncture (CLP) sepsis model, we found that IL-27Rα was associated with the upregulation of TIGIT on memory CD4+ T cells following CLP. However, IL-27 was not associated with sepsis mortality.
by
Lindsay M. Margoles;
Rohit Mittal;
Nathan J. Klingensmith;
John D. Lyons;
Zhe Liang;
Mara A. Serbanescu;
Maylene E. Wagener;
Craig Coopersmith;
Mandy Ford
Sepsis is the leading cause of death in intensive care units in the US, and it is known that chronic alcohol use is associated with higher incidence of sepsis, longer ICU stays, and higher mortality from sepsis. Both sepsis and chronic alcohol use are associated with immune deficits such as decreased lymphocyte numbers, impaired innate immunity, delayed-type hypersensitivity reactions, and susceptibility to infections; however, understanding of specific pathways of interaction or synergy between these two states of immune dysregulation is lacking. This study therefore sought to elucidate mechanisms underlying the immune dysregulation observed during sepsis in the setting of chronic alcohol exposure. Using a murine model of chronic ethanol ingestion followed by sepsis induction via cecal ligation and puncture, we determined that while CD4+ and CD8+ T cells isolated from alcohol fed mice eventually expressed the same cellular activation markers (CD44, CD69, and CD43) and effector molecules (IFN-ã, TNF) as their water fed counterparts, there was an overall delay in the acquisition of these phenotypes. This early lag in T cell activation was associated with significantly reduced IL-2 production at a later timepoint in both the CD4+ and CD8+ T cell compartments in alcohol sepsis, as well as with a reduced accumulation of CD8dim activated effectors. Taken together, these data suggest that delayed T cell activation may result in qualitative differences in the immune response to sepsis in the setting of chronic alcohol ingestion.
by
Marco Colonna;
Bijoy K. Ghosh;
J. Perren Cobb;
Jonathan E. McDunn;
Kareem D. Husain;
Ashoka D. Polpitiya;
Anton Burykin;
Qing Li;
William Schierding;
Nan Lin;
David Dixon;
Weixiong Zhang;
Craig Coopersmith;
W. Michael Dunne
Background: Diagnosis of acute infection in the critically ill remains a challenge. We hypothesized that circulating leukocyte transcriptional profiles can be used to monitor the host response to and recovery from infection complicating critical illness. Methodology/Principal Findings: A translational research approach was employed. Fifteen mice underwent intratracheal injections of live P. aeruginosa, P. aeruginosa endotoxin, live S. pneumoniae, or normal saline. At 24 hours after injury, GeneChip microarray analysis of circulating buffy coat RNA identified 219 genes that distinguished between the pulmonary insults and differences in 7-day mortality. Similarly, buffy coat microarray expression profiles were generated from 27 mechanically ventilated patients every two days for up to three weeks. Significant heterogeneity of VAP microarray profiles was observed secondary to patient ethnicity, age, and gender, yet 85 genes were identified with consistent changes in abundance during the seven days bracketing the diagnosis of VAP. Principal components analysis of these 85 genes appeared to differentiate between the responses of subjects who did versus those who did not develop VAP, as defined by a general trajectory (riboleukogram) for the onset and resolution of VAP. As patients recovered from critical illness complicated by acute infection, the riboleukograms converged, consistent with an immune attractor. Conclusions/Significance: Here we present the culmination of a mouse pneumonia study, demonstrating for the first time that disease trajectories derived from microarray expression profiles can be used to quantitatively track the clinical course of acute disease and identify a state of immune recovery. These data suggest that the onset of an infection-specific transcriptional program may precede the clinical diagnosis of pneumonia in patients. Moreover, riboleukograms may help explain variance in the host response due to differences in ethnic background, gender, and pathogen. Prospective clinical trials are indicated to validate our results and test the clinical utility of riboleukograms.