Conventional and unconventional T cell responses contribute to the prediction of clinical outcome and causative bacterial pathogen in sepsis patients

Sepsis is characterised by a dysfunctional host response to infection culminating in life-threatening organ failure that requires complex patient management and rapid intervention. Timely diagnosis of the underlying cause of sepsis is crucial, and identifying those at risk of complications and death is imperative for triaging treatment and resource allocation. Here, we explored the potential of explainable machine learning models to predict mortality and causative pathogen in sepsis patients. By using a modelling pipeline employing multiple feature selection algorithms, we demonstrate the feasibility to identify integrative patterns from clinical parameters, plasma biomarkers and extensive phenotyping of blood immune cells. Whilst no single variable had sufficient predictive power, models that combined five and more features showed a macro area under the curve (AUC) of 0.85 to predict 90 day mortality after sepsis diagnosis, and a macro AUC of 0.86 to discriminate between Gram-positive and Gram-negative bacterial infections. Parameters associated with the cellular immune response contributed the most to models predictive of 90 day mortality, most notably, the proportion of T cells among PBMCs, together with expression of CXCR3 by CD4+ T cells and CD25 by mucosal-associated invariant T (MAIT) cells. Frequencies of Vδ2+ γδ T cells had the most profound impact on the prediction of Gram-negative infections, alongside other T cell-related variables and total neutrophil count. Overall, our findings highlight the added value of measuring the proportion and activation patterns of conventional and unconventional T cells in the blood of sepsis patients in combination with other immunological, biochemical and clinical parameters.

Inflammatory biomarker signatures in post-surgical drain fluid may detect anastomotic leaks within 48 hours of colorectal resection

BACKGROUND

The optimal treatment of colorectal cancer is surgical resection and primary anastomosis. Anastomotic leak can affect up to 20% of patients and creates significant morbidity and mortality. Current diagnosis of a leak is based on clinical suspicion and subsequent radiology. Peritoneal biomarkers have shown diagnostic utility in other conditions and could be useful in providing earlier diagnosis. This pilot study was designed to assess the practical utility of peritoneal biomarkers after abdominal surgery utilising an automated immunoassay system in routine use for quantifying cytokines.

METHODS

Patients undergoing an anterior resection for a rectal cancer diagnosis were recruited at University Hospital of Wales, Cardiff between June 2019 and June 2021. A peritoneal drain was placed in the proximity of the anastomosis during surgery, and peritoneal fluid was collected at days 1 to 3 post-operatively, and analysed using the Siemens IMMULITE platform for interleukin (IL)-1β, IL-6, IL-10, CXCL8, tumour necrosis factor alpha (TNFα) and C-reactive protein (CRP).

RESULTS

A total of 42 patients were recruited (22M:20F, median age 65). Anastomotic leak was detected in four patients and a further five patients had other intra-abdominal complications. The IMMULITE platform was able to provide robust and reliable results from the analysis of the peritoneal fluid. A metric based on the combination of peritoneal IL-6 and CRP levels was able to accurately diagnose three anastomotic leaks, whilst correctly classifying all negative control patients including those with other complications.

CONCLUSIONS

This pilot study demonstrates that a simple immune signature in surgical drain fluid could accurately diagnose an anastomotic leak at 48 h postoperatively using instrumentation that is already widely available in hospital clinical laboratories.

GeoWaVe: geometric median clustering with weighted voting for ensemble clustering of cytometry data

MOTIVATION

Clustering is an unsupervised method for identifying structure in unlabelled data. In the context of cytometry, it is typically used to categorize cells into subpopulations of similar phenotypes. However, clustering is greatly dependent on hyperparameters and the data to which it is applied as each algorithm makes different assumptions and generates a different 'view' of the dataset. As such, the choice of clustering algorithm can significantly influence results, and there is often not one preferred method but different insights to be obtained from different methods. To overcome these limitations, consensus approaches are needed that directly address the effect of competing algorithms. To the best of our knowledge, consensus clustering algorithms designed specifically for the analysis of cytometry data are lacking.

RESULTS

We present a novel ensemble clustering methodology based on geometric median clustering with weighted voting (GeoWaVe). Compared to graph ensemble clustering methods that have gained popularity in single-cell RNA sequencing analysis, GeoWaVe performed favourably on different sets of high-dimensional mass and flow cytometry data. Our findings provide proof of concept for the power of consensus methods to make the analysis, visualization and interpretation of cytometry data more robust and reproducible. The wide availability of ensemble clustering methods is likely to have a profound impact on our understanding of cellular responses, clinical conditions and therapeutic and diagnostic options.

AVAILABILITY AND IMPLEMENTATION

GeoWaVe is available as part of the CytoCluster package https://github.com/burtonrj/CytoCluster and published on the Python Package Index https://pypi.org/project/cytocluster. Benchmarking data described are available from https://doi.org/10.5281/zenodo.7134723.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

CytoPy: An autonomous cytometry analysis framework

Cytometry analysis has seen a considerable expansion in recent years in the maximum number of parameters that can be acquired in a single experiment. In response to this technological advance there has been an increased effort to develop new computational methodologies for handling high-dimensional single cell data acquired by flow or mass cytometry. Despite the success of numerous algorithms and published packages to replicate and outperform traditional manual analysis, widespread adoption of these techniques has yet to be realised in the field of immunology. Here we present CytoPy, a Python framework for automated analysis of cytometry data that integrates a document-based database for a data-centric and iterative analytical environment. In addition, our algorithm-agnostic design provides a platform for open-source cytometry bioinformatics in the Python ecosystem. We demonstrate the ability of CytoPy to phenotype T cell subsets in whole blood samples even in the presence of significant batch effects due to technical and user variation. The complete analytical pipeline was then used to immunophenotype the local inflammatory infiltrate in individuals with and without acute bacterial infection. CytoPy is open-source and licensed under the MIT license. CytoPy is available at https://github.com/burtonrj/CytoPy, with notebooks accompanying this manuscript (https://github.com/burtonrj/CytoPyManuscript) and software documentation at https://cytopy.readthedocs.io/.

Toll-like receptor linked cytokine profiles in cerebrospinal fluid discriminate neurological infection from sterile inflammation

Rapid determination of an infective aetiology causing neurological inflammation in the cerebrospinal fluid can be challenging in clinical practice. Post-surgical nosocomial infection is difficult to diagnose accurately, as it occurs on a background of altered cerebrospinal fluid composition due to the underlying pathologies and surgical procedures involved. There is additional diagnostic difficulty after external ventricular drain or ventriculoperitoneal shunt surgery, as infection is often caused by pathogens growing as biofilms, which may fail to elicit a significant inflammatory response and are challenging to identify by microbiological culture. Despite much research effort, a single sensitive and specific cerebrospinal fluid biomarker has yet to be defined which reliably distinguishes infective from non-infective inflammation. As a result, many patients with suspected infection are treated empirically with broad-spectrum antibiotics in the absence of definitive diagnostic criteria. To begin to address these issues, we examined cerebrospinal fluid taken at the point of clinical equipoise to diagnose cerebrospinal fluid infection in 14 consecutive neurosurgical patients showing signs of inflammatory complications. Using the guidelines of the Infectious Diseases Society of America, six cases were subsequently characterized as infected and eight as sterile inflammation. Twenty-four contemporaneous patients with idiopathic intracranial hypertension or normal pressure hydrocephalus were included as non-inflamed controls. We measured 182 immune and neurological biomarkers in each sample and used pathway analysis to elucidate the biological underpinnings of any biomarker changes. Increased levels of the inflammatory cytokine interleukin-6 and interleukin-6-related mediators such as oncostatin M were excellent indicators of inflammation. However, interleukin-6 levels alone could not distinguish between bacterially infected and uninfected patients. Within the patient cohort with neurological inflammation, a pattern of raised interleukin-17, interleukin-12p40/p70 and interleukin-23 levels delineated nosocomial bacteriological infection from background neuroinflammation. Pathway analysis showed that the observed immune signatures could be explained through a common generic inflammatory response marked by interleukin-6 in both nosocomial and non-infectious inflammation, overlaid with a toll-like receptor-associated and bacterial peptidoglycan-triggered interleukin-17 pathway response that occurred exclusively during infection. This is the first demonstration of a pathway dependent cerebrospinal fluid biomarker differentiation distinguishing nosocomial infection from background neuroinflammation. It is especially relevant to the commonly encountered pathologies in clinical practice, such as subarachnoid haemorrhage and post-cranial neurosurgery. While requiring confirmation in a larger cohort, the current data indicate the potential utility of cerebrospinal fluid biomarker strategies to identify differential initiation of a common downstream interleukin-6 pathway to diagnose nosocomial infection in this challenging clinical cohort.

Control of neutrophil influx during peritonitis by transcriptional cross-regulation of chemokine CXCL1 by IL-17 and IFN-γ

Neutrophil infiltration is a hallmark of peritoneal inflammation, but mechanisms regulating neutrophil recruitment in patients with peritoneal dialysis (PD)-related peritonitis are not fully defined. We examined 104 samples of PD effluent collected during acute peritonitis for correspondence between a broad range of soluble parameters and neutrophil counts. We observed an association between peritoneal IL-17 and neutrophil levels. This relationship was evident in effluent samples with low but not high IFN-γ levels, suggesting a differential effect of IFN-γ concentration on neutrophil infiltration. Surprisingly, there was no association of neutrophil numbers with the level of CXCL1, a key IL-17-induced neutrophil chemoattractant. We investigated therefore the production of CXCL1 by human peritoneal mesothelial cells (HPMCs) under in vitro conditions mimicking clinical peritonitis. Stimulation of HPMCs with IL-17 increased CXCL1 production through induction of transcription factor SP1 and activation of the SP1-binding region of the CXCL1 promoter. These effects were amplified by TNFα. In contrast, IFN-γ dose-dependently suppressed IL-17-induced SP1 activation and CXCL1 production through a transcriptional mechanism involving STAT1. The SP1-mediated induction of CXCL1 was also observed in HPMCs exposed to PD effluent collected during peritonitis and containing IL-17 and TNFα, but not IFN-γ. Supplementation of the effluent with IFN-γ led to a dose-dependent activation of STAT1 and a resultant inhibition of SP1-induced CXCL1 expression. Transmesothelial migration of neutrophils in vitro increased upon stimulation of HPMCs with IL-17 and was reduced by IFN-γ. In addition, HPMCs were capable of binding CXCL1 at their apical cell surface. These observations indicate that changes in relative peritoneal concentrations of IL-17 and IFN-γ can differently engage SP1-STAT1, impacting on mesothelial cell transcription of CXCL1, whose release and binding to HPMC surface may determine optimal neutrophil recruitment and retention during peritonitis. © 2020 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland.

Using artificial intelligence to reduce diagnostic workload without compromising detection of urinary tract infections

BACKGROUND

A substantial proportion of microbiological screening in diagnostic laboratories is due to suspected urinary tract infections (UTIs), yet approximately two thirds of urine samples typically yield negative culture results. By reducing the number of query samples to be cultured and enabling diagnostic services to concentrate on those in which there are true microbial infections, a significant improvement in efficiency of the service is possible.

METHODOLOGY

Screening process for urine samples prior to culture was modelled in a single clinical microbiology laboratory covering three hospitals and community services across Bristol and Bath, UK. Retrospective analysis of all urine microscopy, culture, and sensitivity reports over one year was used to compare two methods of classification: a heuristic model using a combination of white blood cell count and bacterial count, and a machine learning approach testing three algorithms (Random Forest, Neural Network, Extreme Gradient Boosting) whilst factoring in independent variables including demographics, historical urine culture results, and clinical details provided with the specimen.

RESULTS

A total of 212,554 urine reports were analysed. Initial findings demonstrated the potential for using machine learning algorithms, which outperformed the heuristic model in terms of relative workload reduction achieved at a classification sensitivity > 95%. Upon further analysis of classification sensitivity of subpopulations, we concluded that samples from pregnant patients and children (age 11 or younger) require independent evaluation. First the removal of pregnant patients and children from the classification process was investigated but this diminished the workload reduction achieved. The optimal solution was found to be three Extreme Gradient Boosting algorithms, trained independently for the classification of pregnant patients, children, and then all other patients. When combined, this system granted a relative workload reduction of 41% and a sensitivity of 95% for each of the stratified patient groups.

CONCLUSION

Based on the considerable time and cost savings achieved, without compromising the diagnostic performance, the heuristic model was successfully implemented in routine clinical practice in the diagnostic laboratory at Severn Pathology, Bristol. Our work shows the potential application of supervised machine learning models in improving service efficiency at a time when demand often surpasses resources of public healthcare providers.

Neutrophil-derived miR-223 as local biomarker of bacterial peritonitis

Infection remains a major cause of morbidity, mortality and technique failure in patients with end stage kidney failure who receive peritoneal dialysis (PD). Recent research suggests that the early inflammatory response at the site of infection carries diagnostically relevant information, suggesting that organ and pathogen-specific "immune fingerprints" may guide targeted treatment decisions and allow patient stratification and risk prediction at the point of care. Here, we recorded microRNA profiles in the PD effluent of patients presenting with symptoms of acute peritonitis and show that elevated peritoneal miR-223 and reduced miR-31 levels were useful predictors of bacterial infection. Cell culture experiments indicated that miR-223 was predominantly produced by infiltrating immune cells (neutrophils, monocytes), while miR-31 was mainly derived from the local tissue (mesothelial cells, fibroblasts). miR-223 was found to be functionally stabilised in PD effluent from peritonitis patients, with a proportion likely to be incorporated into neutrophil-derived exosomes. Our study demonstrates that microRNAs are useful biomarkers of bacterial infection in PD-related peritonitis and have the potential to contribute to disease-specific immune fingerprints. Exosome-encapsulated microRNAs may have a functional role in intercellular communication between immune cells responding to the infection and the local tissue, to help clear the infection, resolve the inflammation and restore homeostasis.

The death receptor 3/TL1A pathway is essential for efficient development of antiviral CD4⁺ and CD8⁺ T-cell immunity

Death receptor 3 (DR3, TNFRSF25), the closest family relative to tumor necrosis factor receptor 1, promotes CD4(+) T-cell-driven inflammatory disease. We investigated the in vivo role of DR3 and its ligand TL1A in viral infection, by challenging DR3-deficient (DR3(KO)) mice and their DR3(WT) littermates with the β-herpesvirus murine cytomegalovirus or the poxvirus vaccinia virus. The phenotype and function of splenic T-cells were analyzed using flow cytometry and molecular biological techniques. We report surface expression of DR3 by naive CD8(+) T cells, with TCR activation increasing its levels 4-fold and altering the ratio of DR3 splice variants. T-cell responses were reduced up to 90% in DR3(KO) mice during acute infection. Adoptive transfer experiments indicated this was dependent on T-cell-restricted expression of DR3. DR3-dependent CD8(+) T-cell expansion was NK and CD4 independent and due to proliferation, not decreased cell death. Notably, impaired immunity in DR3(KO) hosts on a C57BL/6 background was associated with 4- to 7-fold increases in viral loads during the acute phase of infection, and in mice with suboptimal NK responses was essential for survival (37.5%). This is the first description of DR3 regulating virus-specific T-cell function in vivo and uncovers a critical role for DR3 in mediating antiviral immunity.

The role of tumor necrosis factor receptor superfamily members in mammalian brain development, function and homeostasis

Tumor necrosis factor receptor superfamily (TNFRSF) members were initially identified as immunological mediators, and are still commonly perceived as immunological molecules. However, our understanding of the diversity of TNFRSF members' roles in mammalian physiology has grown significantly since the first discovery of TNFRp55 (TNFRSF1) in 1975. In particular, the last decade has provided evidence for important roles in brain development, function and the emergent field of neuronal homeostasis. Recent evidence suggests that TNFRSF members are expressed in an overlapping regulated pattern during neuronal development, participating in the regulation of neuronal expansion, growth, differentiation and regional pattern development. This review examines evidence for non-immunological roles of TNFRSF members in brain development, function and maintenance under normal physiological conditions. In addition, several aspects of brain function during inflammation will also be described, when illuminating and relevant to the non-immunological role of TNFRSF members. Finally, key questions in the field will be outlined.