Research letters

3 Recovery from pneumonia recovering from CAP. In patients with the separation of uningested neutrophils Thorax: first published as 10.1136/thoraxjnl-2016-208505 on 28 July 2016. Downloaded from CAP, clinical severity markers are asso- from by light scatter as previ- requires efferocytosis which ciated with short-term mortality risk, but ously published.6 Using R, associations is impaired in smokers and the association of these factors with effer- with efferocytosis were analysed by mul- those with low body mass ocytosis following CAP is not known. If tiple regression. The mean of efferocytosis pre-CAP clinical factors were associated experimental replicates was used for uni- index and enhanced by statins with efferocytosis, then pro-resolution variate analysis; then, a maximal linear therapy to reduce post-CAP adverse mixed effects model was constructed using ABSTRACT events would become a possibility. We efferocytosis as the response variable, cor- Background Efferocytosis (the conducted a prospective cohort study of related (p>0.05) variables from the uni- of apoptotic self cells) is a key mechanism in adults hospitalised for CAP. We hypothe- variate analysis as fixed effects and ‘subject the resolution of inflammatory processes such sised that ex vivo efferocytosis of autolo- id’ as a random effect. The random effect as community-acquired pneumonia (CAP). gous apoptotic neutrophils by alveolar enabled us to separate the stochastic vari- Efferocytosis therefore represents a modifiable macrophages would vary in association ation associated with experimental tech- target for therapy aimed at enhancing intrinsic with patient characteristics. nique from between subject variations. recovery mechanisms. It is currently not known Backwards elimination derived the which patients recovering from CAP would METHODS minimum set of explanatory variables mostly benefit from a strategy aimed at Subjects recruited from two UK Hospitals required to give a statistically acceptable fit. enhancing efferocytosis. between February 2011 and March 2013 Methods We recruited a cohort of patients had CAP (British Thoracic Society defin- RESULTS with CAP admitted to a hospital in Liverpool. ition), were aged >16 years and were Of 169 subjects recruited, efferocytosis was One month into recovery, subjects were invited recruited within 24 hours of their first analysed in 22 (figure 1 and online for research bronchoscopy and bronchoalveolar dose of in-hospital antibiotic. We supplementary tables S1 and S2). Univariate lavage. An ex vivo efferocytosis assay was excluded patients requiring invasive venti- analysis (see online supplementary table S3) performed by challenging alveolar lation, requiring renal replacement revealed a trend towards higher efferocyto- macrophages with autologous, apoptotic therapy, with cystic fibrosis (CF), non-CF sis values with improved symptomatic neutrophils. The percentage of alveolar , lung cancer, lung metasta- recovery (see online supplementary figure macrophages that had undergone efferocytosis ses, advanced cancer of any type, immu- S3), but no relationship with age. Smoking was determined by flow cytometry. We nocompromise (including systemic status, prior statin use, body mass index conducted a multivariable regression using a corticosteroids), those treated palliatively (BMI) and gender correlated with efferocy- linear mixed effects model to determine which or admitted within 14 days. At 1 month, tosis. Those variables were combined in a clinical parameters were most closely subjects were invited for bronchoscopy linear mixed effects model which left three associated with efferocytosis. with bronchoalveolar lavage (BAL) and variables with statistically significant effects: Results We observed high rates of were excluded if they were at increased smoking status, prior statin use and BMI. As comorbidity among this CAP cohort. risk of complications. BAL was performed BMI increased, so did efferocytosis. http://thorax.bmj.com/ Efferocytosis was measured in 22 subjects. We as previously published with 200 ml saline Smoking was associated with lower rates of assessed multiple combinations of clinical instilled into the right middle lobe (RML) efferocytosis. Subjects who were taking parameters for association with efferocytosis bronchus.4 BAL was filtered, pelleted, statins had higher rates of efferocytosis. and found the best-fitting model included an washed and resuspended in Iscove’s Analysis for interactions showed the model interaction between smoking status and prior Modified Dubecco’s Medium (IMDM) with the best data-fit included an interaction statin use—smoking being associated with with human AB serum and antibiotics. between smoking status and prior statin use, decreased efferocytosis and statin use with Cells were seeded into 48 well plates and with adjustment for BMI (figure 2). The increased efferocytosis. These effects were incubated for 4 hours before the medium interaction was such that the on September 28, 2021 by guest. Protected copyright. modified by an association between was replaced with antibiotic-free IMDM. statin-associated increase in efferocytosis efferocytosis and body mass index (BMI), such Ex vivo autologous apoptotic neutro- was largest in those who were active that as BMI increased so did efferocytosis. phils were derived by published protocol.5 smokers. The final model explained 42.6% Conclusions This is the first study to Prior to incubation, the neutrophils were of variation in the data, of which 90.1% measure efferocytosis in patients recovering divided into two aliquots: one stained was the difference between patients and from CAP. The results suggest that smokers green and the other unstained. The effero- 9.9% was within experimental replicates. with low BMI have impaired efferocytosis and assay was a modification of pub- may benefit from a statin to boost recovery. lished protocols.6 After overnight DISCUSSION incubation, the media was removed from This is the first study of efferocytosis INTRODUCTION the macrophages and unstained neutro- during recovery from CAP. After adjust- Among the 80% of patients who survive phils, stained neutrophils or medium ment for BMI, the strongest associations an admission with community-acquired added to each well (see online with efferocytosis were smoking status pneumonia (CAP), a proportion suffers supplementary figure S1). After 90 min and statin use prior to CAP. prolonged symptoms.1 During recovery co-culture, the medium was removed and Strengths of this study include the use from pneumonia, dead and dying neutro- the macrophages washed to remove unin- of autologous neutrophils as a patho- phils must be cleared. This phagocytosis gested neutrophils. Macrophages were physiologically appropriate apoptotic of apoptotic ‘self’ cells, called efferocyto- detached with cold phosphate-buffered target for the efferocytosis assay, the use sis, is defective in idiopathic pulmonary saline (PBS), washed, quenched and then of linear mixed effects modelling to quan- fibrosis and COPD.2 The efferocytosis of acquired on the cytometer. tify the contributions of experimental and apoptotic neutrophils has been studied in The flow cytometric gating strategy (see between-patient variations and the flow pneumonic mice but not in humans online supplementary figure S2) involved cytometric method used. Limitations

1052 Thorax November 2016 Vol 71 No 11 Research letters Thorax: first published as 10.1136/thoraxjnl-2016-208505 on 28 July 2016. Downloaded from

Figure 1 Bronchoscopy flow chart. Approximately half of the subjects assessed were deemed ineligible for research bronchoscopy—the majority of those being excluded on safety grounds due to comorbidity. CAP, community-acquired pneumonia include the small study size, lower median patients were recovering from RML CAP. Previous studies have found that cigar- age and lower range of severity than that It is possible that recent local inflamma- ette smoke affects molecular pathways in clinical practice. It is not known if CAP tion by the RML CAP may have affected that lead to the activation and in humans has an effect on efferocytosis rates of local efferocytosis; however, in localisation of the enzyme Rac which or whether any such effect would be local our univariate analysis, there was no stat- facilitates the cytoskeletal rearrangements or generalised, but we consistently istically significant difference in efferocy- needed for efferocytosis.7 Statins cause lavaged the RML, and three of the tosis associated with RML involvement. the enzymes Rac and RhoA to sequester http://thorax.bmj.com/ on September 28, 2021 by guest. Protected copyright.

Figure 2 Interactions between covariates in the linear mixed effects model. The result of the linear regression was a model that included the interaction of smoking and statin with adjustment for body mass index (BMI). The efferocytosis values fitted by the model (including all replicates) are on the Y-axis, with BMI on the X-axis. Points are ‘jittered’ to avoid overlap. The modelled interactions of each level of smoking and statin are represented by six straight lines. These lines are displayed in colour pairs based on smoking status. The y-intercepts of each line are derived from the model, as is the BMI-dependent slope of 0.6. From these lines, it can be seen that without statins the differences between smoking status are large, but with statins the differences are smaller. The magnitude of the statin effect appears greatest in smokers. There was no association between efferocytosis and sampling from the right middle lobe (RML).

Thorax November 2016 Vol 71 No 11 1053 Research letters

in the cytosol, resulting in increased effer- Ethics approval This work was approved by the UK deacetylase/Rac/CD9 pathways. Int Immunol 2013;25 Thorax: first published as 10.1136/thoraxjnl-2016-208505 on 28 July 2016. Downloaded from ocytosis of apoptotic neutrophils.8 These NHS Research Ethics Committee (NHS Research Ethics (Suppl 11):643–50. 8 Morimoto K, Vandivier RW, Janssen WJ, et al. studies suggest smoking and statins have Committee (REC) Number 10/WNo03/40), was sponsored by Aintree University Hospital NHS Lovastatin enhances clearance of apoptotic cells antagonistic effects on Rac1 and RhoA, Foundation Trust and was listed on the NIHR Clinical (efferocytosis) with implications for chronic obstructive and as a consequence opposite effects on Research Network portfolio. All subjects provided pulmonary disease. J Immunol 2006;176 (Suppl efferocytosis, providing a possible mech- informed consent to join the study. A consultee 12):7657–65. anistic explanation for our finding of a provided assent on behalf of those who lacked capacity 9 Kahlon S, Eurich DT, Padwal RS, et al. Obesity as a consequence of CAP-related delirium—with and outcomes in patients hospitalized with negative association between efferocytosis consent being retrospectively obtained upon recovery of pneumonia. Clin Microbiol Infect 2013;19 and smoking, a positive association with capacity. In addition, at follow-up, all volunteers (Suppl 8):709–16. statins and a statistical interaction between provided informed consent for bronchoscopy and BAL. 10 Phung DT, Wang Z, Rutherford S, et al. Body mass index and risk of pneumonia: a systematic review smoking and statin use. Provenance and peer review Not commissioned; and meta-analysis. Obes Rev 2013;14 (Suppl externally peer reviewed. We also showed a novel association 10):839–57. between BMI and efferocytosis. Previous ▸ Additional material is published online only. To view studies have shown that reduced CAP please visit the journal online (http://dx.doi.org/ mortality is associated with high BMI9 10.1136/thoraxjnl-2016-208505). and that low BMI is associated with increased risk of developing CAP.10 Differential rates of efferocytosis may explain these correlations. Our study suggests that smokers with CAP and low BMI may benefit most from augmented efferocytosis; a statin would be Open Access This is an Open Access article an appropriate candidate for such a trial. distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which Daniel G Wootton,1,2 Peter J Diggle,3 permits others to distribute, remix, adapt and build Joanne Court,4 Odiri Eneje,4 Lynne Keogan,2 upon this work, for commercial use, provided the Laura Macfarlane,4 Sarah Wilks,4 original work is properly cited. See: http:// Mark Woodhead,5,6 Stephen B Gordon3 creativecommons.org/licenses/by/4.0/ 1Institute of Infection and Global Health, University of Liverpool, Liverpool, UK 2Department of Respiratory Research, Aintree University Hospital NHS Foundation Trust, Liverpool, UK 3CHICAS, Lancaster University Medical School, Lancaster University, Lancaster, UK To cite Wootton DG, Diggle PJ, Court J, et al. Thorax 4Department of Clinical Sciences, Liverpool School of 2016;71:1052–1054. Tropical Medicine, Liverpool, UK

Received 19 February 2016 http://thorax.bmj.com/ 5 Department of Respiratory Medicine, Central Revised 31 May 2016 Manchester University Hospitals NHS Foundation Trust, Accepted 12 June 2016 Manchester, UK Published Online First 28 July 2016 6Manchester Academic Health Science Centre and – Faculty of Medical and Human Sciences, University of Thorax 2016;71:1052 1054. Manchester, Manchester, UK doi:10.1136/thoraxjnl-2016-208505 Correspondence to Dr Daniel G Wootton, Department of Clinical Infection, Microbiology and REFERENCES Immunology, Institute of Infection and Global Health, 1 El Moussaoui R, Opmeer BC, de Borgie CA, et al. University of Liverpool, 8 West Derby Street, Liverpool Long-term symptom recovery and health-related on September 28, 2021 by guest. Protected copyright. L69 7BE, UK; [email protected] quality of life in patients with mild-to-moderate-severe community-acquired Acknowledgements The authors thank the patients pneumonia. Chest 2006;130(Suppl 4):1165–72. who volunteered for the study, the NIHR for funding 2 McCubbrey AL, Curtis JL. Efferocytosis and lung the work and the NIHR Clinical Local Research Network disease. Chest 2013;143(Suppl 6):1750–7. (CLRN) for providing strategic support for recruitment. 3 Morimoto K, Amano H, Sonoda F, et al. Alveolar Contributors DGW conceived the study, wrote the macrophages that phagocytose apoptotic neutrophils grant to obtain the funding, defended the ethics, produce during bacterial participated in recruitment, followed up the subjects, pneumonia in mice. Am J Respir Mol Biol – performed all bronchoscopies, performed all lab assays, 2001;24(Suppl 5):608 15. analysed the results and wrote the manuscript. JC, LK, 4 Collins AM, Rylance J, Wootton DG, et al. LM and SW helped recruit the patients, collate the data Bronchoalveolar lavage (BAL) for research; obtaining and reviewed the manuscript. In addition LK assisted adequate sample yield. J Vis Exp 2014;(85):e4345. with the BALs, recovery post bronchoscopy and patient doi:10.3791/4345 fi follow-up. PJD and SBG helped design the study, write 5 Taylor EL, Rossi AG, Drans eld I, et al. Analysis of the grant, and along with MW supervised the neutrophil . Methods Mol Biol – fellowship project including discussion of recruitment, 2007;412:177 200. fi data analysis and the writing of this manuscript. PJD 6 Michlewska S, Drans eld I, Megson IL, et al. supervised the statistical plan and analysis. phagocytosis of apoptotic neutrophils is critically regulated by the opposing actions of Funding The study was funded by the National pro-inflammatory and anti-inflammatory agents: key Institute of Health Research (NIHR) Doctoral Research role for TNF-α. FASEB J 2009;23 (Suppl 3):844–54. Fellowship awarded to DG Wootton with support from 7 Noda N, Matsumoto K, Fukuyama S, et al. Cigarette the North West CLRN. smoke impairs phagocytosis of apoptotic neutrophils Competing interests None declared. by alveolar macrophages via inhibition of the histone

1054 Thorax November 2016 Vol 71 No 11