Comparing machine learning screening approaches using clinical data and cytokine profiles for COVID-19 in resource-limited and resource-abundant settings.

Accurate screening of COVID-19 infection status for symptomatic patients is a critical public health task. Although molecular and antigen tests now exist for COVID-19, in resource-limited settings, screening tests are often not available. Furthermore, during the early stages of the pandemic tests were not available in any capacity. We utilized an automated machine learning (ML) approach to train and evaluate thousands of models on a clinical dataset consisting of commonly available clinical and laboratory data, along with cytokine profiles for patients (n = 150). These models were then further tested for generalizability on an out-of-sample secondary dataset (n = 120). We were able to develop a ML model for rapid and reliable screening of patients as COVID-19 positive or negative using three approaches: commonly available clinical and laboratory data, a cytokine profile, and a combination of the common data and cytokine profile. Of the tens of thousands of models automatically tested for the three approaches, all three approaches demonstrated > 92% sensitivity and > 88 specificity while our highest performing model achieved 95.6% sensitivity and 98.1% specificity. These models represent a potential effective deployable solution for COVID-19 status classification for symptomatic patients in resource-limited settings and provide proof-of-concept for rapid development of screening tools for novel emerging infectious diseases.

Assessing the role of eosinophil-mediated immune response markers in detecting hookworm infection: A case-control study in Kintampo, Ghana.

Background and Aim: Human hookworm disease caused by Ancylostoma duodenale and Necator americanus is a serious public health problem. Hookworm infection activates eosinophil-mediated tissue inflammatory responses, involving the production of the eosinophil-specific chemokine (eotaxin), recruitment of eosinophils, secretion of the cationic protein, and production of antiparasite immunoglobulin E (IgE). We investigated eosinophil-mediated immune response as markers (CCL11, eosinophil cationic protein [ECP], and IgE) for detecting hookworm infection. Methods: This case-control study was carried out in hookworm endemic areas within the Kintampo North Municipality.Forty hookworm-positive subjects and 36 apparently healthy individuals were recruited as cases and controls, respectively. Stool samples were collected for hookworm detection by the Kato-Katz technique and speciation by polymerase chain reaction. Approximately, 5 ml of intravenous blood was used to obtain plasma for the immunological assays. Results: Of eosinophil-mediated immune response markers studied, ECP and CCL11 were significantly higher among hookworm patients compared to controls. Increasing CCL11 (ß = -0.81, p = 0.015) was associated with a significant decrease hookworm intensity. However, increasing eosinophil count (ß = 0.62, p = 0.027) was associated with significant increase in hookworm intensity. In receiver operator characteristics analysis, ECP could significantly detect hookworm infection with a very high area under the curve (AUC) (AUC = 0.97, p < 0.0001). At a cutoff of 39.05, ECP was the best eosinophil-mediated immune response marker for detecting hookworm infection with a sensitivity of 97.2%, specificity of 87.8%, a positive predictive value of 89.7%, and a negative predictive value of 96.6%. Conclusion: ECP best predicts eosinophil-mediated immune response for detecting hookworm infection, while CCL11 and eosinophil count better predict the intensity of hookworm. Moreover, the ECP level is a good indicator of hookworm infection and intensity and may require additional investigations to augment current hookworm diagnostic techniques.

Cytokines and chemokines multiplex analysis in patients with low disease activity rheumatoid arthritis.

Rheumatoid arthritis is a severe chronic autoimmune disorder that results from pathological activation of immune cells and altered cytokine/chemokine network. The aim of our study was to evaluate concentrations of chosen cytokines and chemokines in blood sera and synovial fluid samples isolated from low disease activity rheumatoid arthritis (RA) patients and osteoarthritis (OA) sufferers. Blood sera and synovial fluid samples have been obtained from 24 OA and 14 RA patients. Cytokines/chemokines levels have been determined using a Milliplex® Map 38-plex human cytokine/chemokine magnetic bead-based panel (Merck Millipore, Germany) and Luminex® MAGPIX® platform (Luminex USA). Low disease activity RA patients showed altered concentration of numerous cytokine/chemokine when compared to OA controls-they were characterized by, inter alia, increased: eotaxin/CCL11 (p = 0.037), GRO/CXCL1 (p = 0.037), IL-2 (p = 0.013), IL-4 (p = 0.017), IL-7 (p = 0.003), IL-8 (p = 0.0007) and GM-CSF (p = 0.037) serum levels, whilst MDC/CCL22 concentration was decreased in this group (p = 0.034). Eotaxin/CCL11 (p = 0.001), GRO/CXCL1 (p = 0.041), IL-10 (p = 0.003), GM-CSF (p = 0.01), IL-1RA (p = 0.0005) and VEGF (p = 0.01) concentrations in synovial fluid of RA females were also increased. Even with low disease activity score, RA patients exhibited increased concentrations of cytokines with pro- and anti-inflammatory activities, as well as numerous chemokines, growth factors and regulators of angiogenesis. Surprisingly, RA subjects also shown decreased concentration of CCL22 chemokine. The attempt to restore cytokine balance and tolerogenic environment is ineffective in RA sufferers even with good disease management. Distinguished factors could serve as possible indicators of disease progression even in low disease activity patients.

Increased peripheral levels of TARC/CCL17 in first episode psychosis patients.

BACKGROUND: Evidence for a link between the pathophysiology of schizophrenia and the immune system is mounting. Altered levels of chemokines in plasma have previously been reported in patients with schizophrenia under antipsychotic medication. Here we aimed to study both peripheral and central chemokine levels in drug-naïve or short-time medicated first episode psychosis (FEP) patients. METHOD: We analyzed nine chemokines in plasma and CSF from 41 FEP patients and 22 healthy controls using electrochemiluminescence assay. RESULTS: In plasma four chemokines; TARC/CCL17, eotaxin/CCL11, MDC/CCL22, IP-10/CXCL10 and in CSF one chemokine; IP-10/CXCL10 showed reliable detection in >50% of the cases. FEP patients displayed increased levels of TARC/CCL17 in plasma compared to healthy controls, 89.6 (IQR 66.2-125.8) pg/mL compared to 48.6 (IQR 28.0-71.7) pg/mL (p = 0.001). The difference was not attributed to confounding factors. Plasma TARC/CCL17 was not associated with PANSS, CGI or GAF scores, neither with cognitive functions. The chemokines eotaxin/CCL11, MDC/CCL22, IP-10/CXCL10 in plasma and IP-10/CXCL10 in CSF did not differ between FEP patients and controls. CONCLUSION: In line with a previous study showing that chronic patients with schizophrenia display increased plasma TARC/CCL17 levels, we here found an elevation in FEP patients suggesting a role of TARC/CCL17 in early stages of schizophrenia. The exact mechanism of this involvement is still unknown and future longitudinal studies as well as studies of central and peripheral chemokine levels would be of great interest.

Increased serum levels of eotaxin/CCL11 in late-stage patients with bipolar disorder: An accelerated aging biomarker?

BACKGROUND: Bipolar disorder (BD) is commonly comorbid with many medical disorders including atopy, and appears characterized by progressive social, neurobiological, and functional impairment associated with increasing number of episodes and illness duration. Early and late stages of BD may present different biological features and may therefore require different treatment strategies. Consequently, the aim of this study was to evaluate serum levels of eotaxin/CCL11, eotaxin-2/CCL24, IL-2, IL-4, IL-6, IL-10, IL-17, TNF-α, IFNγ, BDNF, TBARS, carbonyl, and GPx in a sample of euthymic patients with BD at early and late stages compared to controls. METHODS: Early-stage BD patients, 12 late-stage patients, and 25 controls matched for sex and age were selected. 10mL of peripheral blood was drawn from all subjects by venipuncture. Serum levels of BDNF, TBARS, carbonyl content, glutathione-peroxidase activity (GPx), cytokines (IL-2, IL-4, IL-6, IL-10, IL-17, TNF-α and IFNγ), and chemokines (eotaxin/CCL11 and eotaxin-2/CCL24) were measured. RESULTS: There were no demographic differences between patients and controls. No significant differences were found for any of the biomarkers, except chemokine eotaxin/CCL11, whose serum levels were higher in late-stage patients with BD when compared to controls (p=0.022; Mann-Whitney U test). LIMITATIONS: Small number of subjects and use of medication may have influenced in our results. CONCLUSION: The present study suggests a link between biomarkers of atopy and eosinophil function and bipolar disorder. These findings are also in line with progressive biological changes partially mediated by inflammatory imbalance, a process referred to as neuroprogression.

Basophils regulate the recruitment of eosinophils in a murine model of irritant contact dermatitis.

BACKGROUND: Although eosinophils have been detected in several human skin diseases in the vicinity of basophils, how eosinophils infiltrate the skin and the role of eosinophils in the development of skin inflammation have yet to be examined. OBJECTIVE: Using murine irritant contact dermatitis (ICD) as a model, we sought to clarify the roles of eosinophils in ICD and the underlying mechanism of eosinophil infiltration of the skin. METHODS: We induced croton oil-induced ICD in eosinophil-deficient ΔdblGATA mice with or without a reactive oxygen species (ROS) inhibitor. We performed cocultivation with fibroblasts and bone marrow-derived basophils and evaluated eosinophil migration using a chemotaxis assay. RESULTS: ICD responses were significantly attenuated in the absence of eosinophils or by treatment with the ROS inhibitor. ROS was produced abundantly by eosinophils, and both basophils and eosinophils were detected in human and murine ICD skin lesions. In coculture experiments, basophils attracted eosinophils, especially in the presence of fibroblasts. Moreover, basophils produced IL-4 and TNF-α in contact with fibroblasts and promoted the expression of eotaxin/CCL11 from fibroblasts in vitro. CONCLUSION: Eosinophils mediated the development of murine ICD, possibly through ROS production. Recruitment of eosinophils into the skin was induced by basophils in cooperation with fibroblasts. Our findings introduce the novel concept that basophils promote the recruitment of eosinophils into the skin through fibroblasts in the development of skin inflammation.