Prediction of cardiovascular risk factors from retinal fundus photographs: Validation of a deep learning algorithm in a prospective non-interventional study in Kenya.

AIM: Hypertension and diabetes mellitus (DM) are major causes of morbidity and mortality, with growing burdens in low-income countries where they are underdiagnosed and undertreated. Advances in machine learning may provide opportunities to enhance diagnostics in settings with limited medical infrastructure. MATERIALS AND METHODS: A non-interventional study was conducted to develop and validate a machine learning algorithm to estimate cardiovascular clinical and laboratory parameters. At two sites in Kenya, digital retinal fundus photographs were collected alongside blood pressure (BP), laboratory measures and medical history. The performance of machine learning models, originally trained using data from the UK Biobank, were evaluated for their ability to estimate BP, glycated haemoglobin, estimated glomerular filtration rate and diagnoses from fundus images. RESULTS: In total, 301 participants were enrolled. Compared with the UK Biobank population used for algorithm development, participants from Kenya were younger and would probably report Black/African ethnicity, with a higher body mass index and prevalence of DM and hypertension. The mean absolute error was comparable or slightly greater for systolic BP, diastolic BP, glycated haemoglobin and estimated glomerular filtration rate. The model trained to identify DM had an area under the receiver operating curve of 0.762 (0.818 in the UK Biobank) and the hypertension model had an area under the receiver operating curve of 0.765 (0.738 in the UK Biobank). CONCLUSIONS: In a Kenyan population, machine learning models estimated cardiovascular parameters with comparable or slightly lower accuracy than in the population where they were trained, suggesting model recalibration may be appropriate. This study represents an incremental step toward leveraging machine learning to make early cardiovascular screening more accessible, particularly in resource-limited settings.

AZD7624, an Inhaled p38 Inhibitor, Demonstrates Local Lung Inhibition of LPS-Induced TNFα with Minimal Systemic Exposure.

Inhaled drugs generally aim to drive a local pharmacological effect in lung, at the same time minimizing systemic exposure, in order to obtain efficacy in lung disease without unwanted systemic effects. Here, we demonstrate that inhaled delivery of a p38 inhibitor (AZD7624) can provide superior pharmacokinetic exposure and superior pharmacodynamic lung effects. In rats, inhaled AZD7624 had a five times higher dose-adjusted lung exposure compared with intravenous dosing. In healthy volunteers, lipopolysaccharide (LPS)-induced tumor necrosis factor α (TNFα) in sputum has been shown to be significantly reduced (85%) by means of inhaled AZD7624. Here, we demonstrate that this effect is associated with a mean unbound plasma concentration, gained from in vitro and ex vivo LPS-challenge protocols, significantly below potencies obtained for AZD7624, suggesting that lung exposure is probably much higher than systemic exposure. This assessment was made for the unbound potency (pIC50u), e.g., the potency remaining after adjustment for plasma protein binding and blood plasma ratio. Hence, the unbound potency of AZD7624 to inhibit LPS-induced TNFα in human mononuclear cells, in whole blood as well as in alveolar macrophages in vitro, was 8.4, 8.7 (full inhibition), and 9.0 (partial inhibition), respectively. The pIC50u in whole blood ex vivo was 8.8, showing good in vitro/in vivo potency correlations. Thus, a mean unbound AZD7624 plasma concentration of 0.3 nmol/l, which was associated with a decrease in LPS-induced sputum TNFα by 85%, is much lower than the pIC50u. This demonstrates that AZD7624 can achieve significant local lung pharmacodynamic effects with concomitant sub-pharmacological systemic exposure.

Expression of MATE1, P-gp, OCTN1 and OCTN2, in epithelial and immune cells in the lung of COPD and healthy individuals.

BACKGROUND: Several inhaled drugs are dependent on organic cation transporters to cross cell membranes. To further evaluate their potential to impact on inhaled drug disposition, the localization of MATE1, P-gp, OCTN1 and OCTN2 were investigated in human lung. METHODS: Transporter proteins were analysed by immunohistochemistry in lung tissue from healthy subjects and COPD patients. Transporter mRNA was analysed by qPCR in lung tissue and in bronchoalveolar lavage (BAL) cells from smokers and non-smokers. RESULTS: We demonstrate for the first time MATE1 protein expression in the lung with localization to the apical side of bronchial and bronchiolar epithelial cells. Interestingly, MATE1 was strongly expressed in alveolar macrophages as demonstrated both in lung tissue and in BAL cells, and in inflammatory cells including CD3 positive T cells. P-gp, OCTN1 and OCTN2 were also expressed in the alveolar epithelial cells and in inflammatory cells including alveolar macrophages. In BAL cells from smokers, MATE1 and P-gp mRNA expression was significantly lower compared to cells from non-smokers whereas no difference was observed between COPD patients and healthy subjects. THP-1 cells were evaluated as a model for alveolar macrophages but did not reflect the transporter expression observed in BAL cells. CONCLUSIONS: We conclude that MATE1, P-gp, OCTN1 and OCTN2 are expressed in pulmonary lung epithelium, in alveolar macrophages and in other inflammatory cells. This is important to consider in the development of drugs treating pulmonary disease as the transporters may impact drug disposition in the lung and consequently affect pharmacological efficacy and toxicity.

Estimation of Equipotent Doses for Anti-Inflammatory Effects of Prednisolone and AZD9567, an Oral Selective Nonsteroidal Glucocorticoid Receptor Modulator.

AZD9567 is a potent and selective nonsteroidal oral glucocorticoid receptor modulator. It is developed as an anti-inflammatory drug with improved safety profile compared with steroids like prednisolone. Throughout the clinical development of AZD9567, dose selection and data interpretation require a method for determining doses with the same anti-inflammatory effect as prednisolone. Equipotent doses of AZD9567 and prednisolone were defined by the same average inhibition of TNFα release, a biomarker of anti-inflammatory effect, measured in a lipopolysaccharide-stimulated whole blood ex vivo assay. Based on pharmacokinetic-pharmacodynamic models, TNFα dose-response relationships for AZD9567 and prednisolone were established. A comparison of the dose-response curves enabled estimation of an equipotency relationship. Specifically, 20 mg prednisolone was estimated to be equipotent to 40 mg AZD9567 (95% confidence interval: 29-54 mg). Static concentration-response analyses showed that the relative potencies for inhibition of TNFα release of AZD9567 and prednisolone were well aligned with several other pro-inflammatory cytokines.

Effects of a selective glucocorticoid receptor modulator (AZD9567) versus prednisolone in healthy volunteers: two phase 1, single-blind, randomised controlled trials.

BACKGROUND: Glucocorticoids are highly effective and widely used anti-inflammatory drugs, but their use is limited by serious side-effects, including glucocorticoid-induced hyperglycaemia and diabetes. AZD9567 is a non-steroidal, selective glucocorticoid receptor modulator that aims to reduce side-effects. We aimed to assess the safety, tolerability, and pharmacokinetics of AZD9567 in healthy volunteers. METHODS: Two phase 1 clinical studies were done. First, a randomised, placebo-controlled, single-blind, single-ascending dose study was done in healthy men who received single oral doses of AZD9567 2 mg, 10 mg, 20 mg, 40 mg, 80 mg, 100 mg, 125 mg, or 155 mg, or prednisolone 60 mg (n=8 per dose group, randomly assigned [6:2] to receive active drug or placebo). Second, a randomised, active-controlled, single-blind, multiple-ascending dose study was done, in which men and women received oral AZD9567 or prednisolone once daily for 5 days. One cohort of volunteers with prediabetes received AZD9567 10 mg (n=7) or prednisolone 20 mg (n=2). All other cohorts comprised healthy volunteers, receiving AZD9567 20 mg, 40 mg, 80 mg, or 125 mg (n=7 per dose group), or prednisolone 5 mg (n=13), 20 mg (n=16), or 40 mg (n=13). Participants and study centre staff were masked to treatment assignment for each cohort, although data were unmasked for safety review between cohorts. The primary outcome of the single-ascending dose study was the safety, tolerability, and pharmacokinetics of single ascending doses of AZD9567; for the multiple-ascending dose study it was the safety and tolerability of AZD9567 following multiple ascending doses. As a secondary outcome, effects on glycaemic control were ascertained with oral glucose tolerance tests (OGTTs) done at baseline and on day 1 of the single-ascending dose study, and at baseline and on day 4 of the multiple-ascending dose study. These trials are registered at ClinicalTrials.gov, NCT02512575 and NCT02760316. FINDINGS: In the single-ascending dose study, between Nov 18, 2015, and Sept 26, 2016, 72 healthy white men were enrolled, and all completed the study. In the multiple-ascending dose study, between May 2, 2016, and Sept 13, 2017, 77 predominantly white male volunteers (including nine individuals with prediabetes and eight women) were enrolled and 75 completed the study. All doses of AZD9567 and prednisolone were well tolerated, with no serious adverse events or events suggesting adrenal insufficiency. In the single-ascending dose study, nine adverse events of mild intensity were reported (five with AZD9567 and four with placebo); no adverse event was reported by more than one person. In the multiple-ascending dose study, 44 adverse events of mild or moderate intensity were reported (18 with AZD9567 and 26 with prednisolone). The most common were headache and micturition. Apparent clearance, volume of distribution, and half-life of AZD9567 were consistent across doses and for single versus repeated dosing. In the multiple-ascending dose study, OGTTs showed no significant difference with AZD9567 doses up to 80 mg compared with prednisolone 5 mg in glucose area under the curve from 0 h to 4 h post-OGTT (AUC0-4h) from baseline to day 4; the increase in glucose AUC0-4h from baseline to day 4 was significantly lower with all AZD9567 doses versus prednisolone 20 mg (AZD9567 20 mg p<0·0001, 40 mg p=0·0001, 80 mg p=0·0001, and 125 mg p=0·0237). INTERPRETATION: AZD9567 appears to be safe and well tolerated in healthy, predominantly white male volunteers and shows promising initial evidence for improved post-prandial glucose control. Studies of longer duration, with a greater proportion of women and other ethnic groups, and in patients requiring anti-inflammatory treatment are needed to characterise the clinical efficacy and safety profile of AZD9567. FUNDING: AstraZeneca.

Discovery of a Novel Oral Glucocorticoid Receptor Modulator (AZD9567) with Improved Side Effect Profile.

Synthetic glucocorticoids (GC) are essential for the treatment of a broad range of inflammatory diseases. However, their use is limited by target related adverse effects on, e.g., glucose homeostasis and bone metabolism. Starting from a nonsteroidal GR ligand (4) that is a full agonist in reporter gene assays, we exploited key functional triggers within the receptor, generating a range of structurally diverse partial agonists. Of these, only a narrow subset exhibited full anti-inflammatory efficacy and a significantly reduced impact on adverse effect markers in human cell assays compared to prednisolone. This led to the discovery of AZD9567 (15) with excellent in vivo efficacy when dosed orally in a rat model of joint inflammation. Compound 15 is currently being evaluated in clinical trials comparing the efficacy and side effect markers with those of prednisolone.

The development of AZD7624 for prevention of exacerbations in COPD: a randomized controlled trial.

Background: p38 mitogen-activated protein kinase (MAPK) plays a central role in the regulation and activation of pro-inflammatory mediators. COPD patients have increased levels of activated p38 MAPK, which correlate with increased lung function impairment, alveolar wall inflammation, and COPD exacerbations. Objectives: These studies aimed to assess the effect of p38 inhibition with AZD7624 in healthy volunteers and patients with COPD. The principal hypothesis was that decreasing lung inflammation via inhibition of p38α would reduce exacerbations and improve quality of life for COPD patients at high risk for acute exacerbations. Methods: The p38 isoform most relevant to lung inflammation was assessed using an in situ proximity ligation assay in severe COPD patients and donor controls. Volunteers aged 18-55 years were randomized into the lipopolysaccharide (LPS) challenge study, which investigated the effect of a single dose of AZD7624 vs placebo on inflammatory biomarkers. The Proof of Principle study randomized patients aged 40-85 years with a diagnosis of COPD for >1 year to AZD7624 or placebo to assess the effect of p38 inhibition in decreasing the rate of exacerbations. Results: The p38 isoform most relevant to lung inflammation was p38α, and AZD7624 specifically inhibited p38α and p38ß isoforms in human alveolar macrophages. Thirty volunteers were randomized in the LPS challenge study. AZD7624 reduced the increase from baseline in sputum neutrophils and TNF-α by 56.6% and 85.4%, respectively (p<0.001). In the 213 patients randomized into the Proof of Principle study, there was no statistically significant difference between AZD7624 and placebo when comparing the number of days to the first moderate or severe exacerbation or early dropout. Conclusion: Although p38α is upregulated in the lungs of COPD patients, AZD7624, an isoform-specific inhaled p38 MAPK inhibitor, failed to show any benefit in patients with COPD.

Gene expression analysis of membrane transporters and drug-metabolizing enzymes in the lung of healthy and COPD subjects.

This study describes for the first time the expression levels of genes encoding membrane transporters and drug-metabolizing enzymes in the lungs of ex-smoking patients with chronic obstructive pulmonary disease (COPD). Membrane transporters and drug-metabolizing enzymes are key determinants of drug uptake, metabolism, and elimination for systemically administered as well as inhaled drugs, with consequent influence on clinical efficacy and patient safety. In this study, while no difference in gene expression was found between healthy and COPD subjects, we identified a significant regional difference in mRNA expression of both membrane transporters and drug-metabolizing enzymes between central and peripheral tissue in both healthy and COPD subjects. The majority of the differentially expressed genes were higher expressed in the central airways such as the transporters SLC2A1 (GLUT1), SLC28A3 (CNT3), and SLC22A4 (OCTN1) and the drug-metabolizing enzymes GSTZ1, GSTO2, and CYP2F1. Together, this increased knowledge of local pharmacokinetics in diseased and normal lung may improve modeling of clinical outcomes of new chemical entities intended for inhalation therapy delivered to COPD patients. In addition, based on the similarities between COPD and healthy subjects regarding gene expression of membrane transporters and drug-metabolizing enzymes, our results suggest that clinical pharmacological studies in healthy volunteers could be a valid model of COPD patients regarding drug disposition of inhaled drugs in terms of drug metabolism and drug transporters.